In general, I thought that this was an excellent paper, with some very useful ideas on how to improve the rigor of the analytical process of determining whether dumping and injury are likely to continue or recur if antidumping duties are removed in the context of a sunset review. In this brief comment, I note some of my concerns with the entire process of sunset reviews – could there ever be an analytically satisfactory way of assessing the likelihood of recurring dumping and injury? In addition, I include a few comments on the overall approach of the WTO Appellate Body in the sunset review cases it has decided, to date, on the usefulness and proper interpretation of the so-called mandatory–discretionary distinction in ‘as such’ challenges to antidumping measures, and on the Appellate Body's treatment of the Panel's analysis of the US Sunset Policy Bulletin.
The paper essentially argues for a greater use of econometric models in the determination of whether dumping and injury are likely to recur. As noted, I think that this introduces more, desirable rigor into a process that would benefit therefrom. As such, it is a useful proposal. However, I think that there are very basic problems with the whole idea of the sunset review process. While improving the rigor of that process is laudable, it may be ultimately unsatisfying because the process itself is bankrupt. Essentially, the sunset review process tries to predict the likelihood that dumping and injury will recur if antidumping duties imposed five years earlier are removed. I think that is virtually unknowable, and any attempt at such a prediction is pure speculation. Consider the key factor in making such a prediction: How will the foreign producers/exporters behave if the antidumping duties are removed? While it would be interesting to know why they dumped in the first instance, the market conditions and other factors that led them to do so all occurred at least six years prior to the sunset review. Since there is no particular reason to believe, in this era of rapid change and globalization, that the market structure and the other relevant factors will be the same, any attempt to base a prediction of future behavior on the behavior that occurred six years ago has to be speculative. Consider the following: The market structure has likely changed. The advent of a globalized economy has meant that ownership structures are more fluid. Even industries like steel – the major users of antidumping actions historically – now have much different ownership patterns than they did only a few years ago, and, in particular, there has been a significant increase in cross-border holdings.Footnote 1 In many industries, technological innovation means that product cycles last less than six years. Even where ownership patterns and products have not changed, it is likely that there will be significantly different trade flows than there were six years earlier. This makes any attempt to base predictions on the past, pre-dumping duty period simply not realistic. Moreover, the prior behavior of the dumpers occurred at a time when the possibility of an antidumping investigation being initiated was unclear, a situation that no longer obtains. The mere fact of the sunset review indicates that domestic industry has an interest in preserving the antidumping duties and may well be expected to start a case if dumping behavior recurs. Thus, while it may be interesting to know what market conditions were present when the dumping initially occurred, I think that it can give only very limited insight into what might be expected in a post antidumping duty market six or seven years later. The period that the dumping duties were in effect also offers little guidance as to how the market will evolve if the antidumping duties are removed. It seems clear that different producer/exporters respond to antidumping duties in different ways – some abandon the market; others may behave strategically and attempt to use the existence of the antidumping duties to manipulate their prices so as to maximize their profits and/or market share. Thus, behavior in this period is also of little use in predicting the future. The fact is that the sunset-review process asks a question that is impossible to answer with confidence – how will businesses behave if antidumping duties are removed in a market that has never existed before?
As a policy matter, I think that five years of antidumping duty protection (not counting the market effects of the year-long investigation period prior to the imposition of duties) is sufficient for any injured industry. It is longer than the protection typically afforded to seriously injured industries in safeguards proceedings.Footnote 2 Accordingly, in my view the rule should be that antidumping duties terminate after five years. If dumping recurs, a new investigation can be initiated, and if dumping is found, perhaps the sunset policy should be suspended for that product. But the basic rule should be automatic sunset. While this may sound utopian, I would note that Australia – one of the traditional major users of antidumping measures – once had a three-year sunset rule.
In their analysis of the Appellate Body's decision in this case, the authors express concern that its decisions have made sunset review an ‘empty box’, suggesting that the provisions on sunset review in the Antidumping Agreement have been interpreted so as to make them largely meaningless. In this regard, I would note that the authors may assume that a greater effect was intended for these provisions than may have been in the minds of at least some negotiators. The negotiations of the Antidumping Agreement were hotly contested, and the reality is that (i) different words were used in this provision than in the general provisions on dumping (‘review’ instead of ‘investigation’) and (ii) a sunset review certainly involves issues distinct from an original investigation. As a result, it is not so clear that the rules on investigations should be imported indiscriminately into the section on reviews. The suggestion that ‘common sense’ should prevail over the ‘text’ is not convincing, especially given the textualist approach generally applied by the Appellate Body. That said, I would concede that the Appellate Body seems to take a much more restrained view of the agreement as a control on antidumping authorities in the context of reviews than it has in other areas, most notably ‘zeroing’. In its ‘zeroing’ cases, it has paid little heed to the text at times, while it seems fixated on the text in its sunset review cases. Why? Perhaps it is a matter of timing and politics. The first limiting view of the sunset rulesFootnote 3 came shortly after the Byrd Amendment report,Footnote 4 which was quite severely criticized in the United States.Footnote 5 Perhaps the Appellate Body reacted by ‘going slow’ in the sunset review context. That would be consistent with the authors' view that the Appellate Body has given recent signs of tightening its control over national authorities' discretion in sunset reviews by suggesting there are (as of yet not clearly defined) requirements for such reviews embodied in such terms as ‘review’ and ‘determination’ and the need for a ‘nexus’ between continuing dumping and injury, even if not a need for a ‘causal’ relation.
The so-called ‘mandatory–discretionary’ distinction recognized in GATT/WTO dispute settlement allows mandatory government laws or regulations (‘measures’) to be challenged on an ‘as such’ basis. That is to say, if a measure, which the government is required to apply, necessitates a WTO violation, the measure itself may be challenged and not just its specific application. Initially, the distinction was used by Panels to reject challenges to measures that did not mandate a GATT violation, particularly where the government had discretion that it had not yet exercised.Footnote 6 This made sense, as it was a sort of ‘ripeness’ requirement.Footnote 7 In recent times, the distinction has been perverted into a suggestion that no measure can be challenged on an ‘as such’ basis if the government has the discretion not to violate WTO/GATT rules. Thus, in the ‘zeroing’ cases,Footnote 8 the US argued that its practice of zeroing – a practice that it applied in every antidumping investigation where the issue arose (and it arose in most cases) – could not be challenged on an ‘as such’ basis since it was not required by US law to violate WTO rules, notwithstanding the fact that it always exercised its discretion so as to violate them. That makes no sense. Thus, while there is an open issue as how to define what sort of practice constitutes a challengeable measure, the mandatory–discretionary distinction should not be applied to prevent ‘as such’ challenges to consistent patterns of WTO violations.
In the instant case, the question was whether the US applied certain presumptions in a way that they were not rebuttable, such that they should be challengeable ‘as such’. I tend to think that the Appellate Body may have been justified in finding that the Panel's examination of the US practice in the area had failed to establish such an irrebuttable presumption. While one can differ on this issue, it appears to me that there were some problems with the Panel's approach. Mexico had submitted evidence on 306 sunset reviews, but only 21 – or 7% – were considered to be relevant by the Panel in that they were contested. In seven of those 21 cases, no additional evidence was offered to the authorities; in five of them, no attempt was made to show that there was ‘good cause’ to consider additional factors. While one can wonder about the appropriateness of a ‘good cause’ requirement in this situation, it was not challenged, with the result that there were only nine cases on which the Panel could have based a finding that the presumption was irrebuttable. And even in respect of that handful of cases, for the most part the conclusion of the Panel was that the authorities' treatment of arguments that good cause existed was ‘troubling’.Footnote 9 In any event, I think that it is arguable that the Panel did not give sufficient details of its examination of the individual cases to allow one to be satisfied that these nine cases showed that the presumption was irrebutable. In the end, the key fact for the Panel seemed to be that even though the US authorities sometimes said they were open to looking at other factors, in these few cases they did not. Thus, I was not surprised that the Appellate Body found that the Panel had not made the required ‘qualitative’ analysis. In any event, this issue seems quite minor. Even if the US had been found to have violated the WTO rules on an ‘as such’ basis, implementation would have involved either (i) a simple re-writing of the sunset policy bulletin to eschew firmly any irrebuttable presumptions or (ii) the elimination of the offending language and the adoption of more opaque procedures, common in almost all other WTO members that impose antidumping duties.
You have
Access
In general, I thought that this was an excellent paper, with some very useful ideas on how to improve the rigor of the analytical process of determining whether dumping and injury are likely to continue or recur if antidumping duties are removed in the context of a sunset review. In this brief comment, I note some of my concerns with the entire process of sunset reviews – could there ever be an analytically satisfactory way of assessing the likelihood of recurring dumping and injury? In addition, I include a few comments on the overall approach of the WTO Appellate Body in the sunset review cases it has decided, to date, on the usefulness and proper interpretation of the so-called mandatory–discretionary distinction in ‘as such’ challenges to antidumping measures, and on the Appellate Body's treatment of the Panel's analysis of the US Sunset Policy Bulletin.
The paper essentially argues for a greater use of econometric models in the determination of whether dumping and injury are likely to recur. As noted, I think that this introduces more, desirable rigor into a process that would benefit therefrom. As such, it is a useful proposal. However, I think that there are very basic problems with the whole idea of the sunset review process. While improving the rigor of that process is laudable, it may be ultimately unsatisfying because the process itself is bankrupt. Essentially, the sunset review process tries to predict the likelihood that dumping and injury will recur if antidumping duties imposed five years earlier are removed. I think that is virtually unknowable, and any attempt at such a prediction is pure speculation. Consider the key factor in making such a prediction: How will the foreign producers/exporters behave if the antidumping duties are removed? While it would be interesting to know why they dumped in the first instance, the market conditions and other factors that led them to do so all occurred at least six years prior to the sunset review. Since there is no particular reason to believe, in this era of rapid change and globalization, that the market structure and the other relevant factors will be the same, any attempt to base a prediction of future behavior on the behavior that occurred six years ago has to be speculative. Consider the following: The market structure has likely changed. The advent of a globalized economy has meant that ownership structures are more fluid. Even industries like steel – the major users of antidumping actions historically – now have much different ownership patterns than they did only a few years ago, and, in particular, there has been a significant increase in cross-border holdings.Footnote 1 In many industries, technological innovation means that product cycles last less than six years. Even where ownership patterns and products have not changed, it is likely that there will be significantly different trade flows than there were six years earlier. This makes any attempt to base predictions on the past, pre-dumping duty period simply not realistic. Moreover, the prior behavior of the dumpers occurred at a time when the possibility of an antidumping investigation being initiated was unclear, a situation that no longer obtains. The mere fact of the sunset review indicates that domestic industry has an interest in preserving the antidumping duties and may well be expected to start a case if dumping behavior recurs. Thus, while it may be interesting to know what market conditions were present when the dumping initially occurred, I think that it can give only very limited insight into what might be expected in a post antidumping duty market six or seven years later. The period that the dumping duties were in effect also offers little guidance as to how the market will evolve if the antidumping duties are removed. It seems clear that different producer/exporters respond to antidumping duties in different ways – some abandon the market; others may behave strategically and attempt to use the existence of the antidumping duties to manipulate their prices so as to maximize their profits and/or market share. Thus, behavior in this period is also of little use in predicting the future. The fact is that the sunset-review process asks a question that is impossible to answer with confidence – how will businesses behave if antidumping duties are removed in a market that has never existed before?
As a policy matter, I think that five years of antidumping duty protection (not counting the market effects of the year-long investigation period prior to the imposition of duties) is sufficient for any injured industry. It is longer than the protection typically afforded to seriously injured industries in safeguards proceedings.Footnote 2 Accordingly, in my view the rule should be that antidumping duties terminate after five years. If dumping recurs, a new investigation can be initiated, and if dumping is found, perhaps the sunset policy should be suspended for that product. But the basic rule should be automatic sunset. While this may sound utopian, I would note that Australia – one of the traditional major users of antidumping measures – once had a three-year sunset rule.
In their analysis of the Appellate Body's decision in this case, the authors express concern that its decisions have made sunset review an ‘empty box’, suggesting that the provisions on sunset review in the Antidumping Agreement have been interpreted so as to make them largely meaningless. In this regard, I would note that the authors may assume that a greater effect was intended for these provisions than may have been in the minds of at least some negotiators. The negotiations of the Antidumping Agreement were hotly contested, and the reality is that (i) different words were used in this provision than in the general provisions on dumping (‘review’ instead of ‘investigation’) and (ii) a sunset review certainly involves issues distinct from an original investigation. As a result, it is not so clear that the rules on investigations should be imported indiscriminately into the section on reviews. The suggestion that ‘common sense’ should prevail over the ‘text’ is not convincing, especially given the textualist approach generally applied by the Appellate Body. That said, I would concede that the Appellate Body seems to take a much more restrained view of the agreement as a control on antidumping authorities in the context of reviews than it has in other areas, most notably ‘zeroing’. In its ‘zeroing’ cases, it has paid little heed to the text at times, while it seems fixated on the text in its sunset review cases. Why? Perhaps it is a matter of timing and politics. The first limiting view of the sunset rulesFootnote 3 came shortly after the Byrd Amendment report,Footnote 4 which was quite severely criticized in the United States.Footnote 5 Perhaps the Appellate Body reacted by ‘going slow’ in the sunset review context. That would be consistent with the authors' view that the Appellate Body has given recent signs of tightening its control over national authorities' discretion in sunset reviews by suggesting there are (as of yet not clearly defined) requirements for such reviews embodied in such terms as ‘review’ and ‘determination’ and the need for a ‘nexus’ between continuing dumping and injury, even if not a need for a ‘causal’ relation.
The so-called ‘mandatory–discretionary’ distinction recognized in GATT/WTO dispute settlement allows mandatory government laws or regulations (‘measures’) to be challenged on an ‘as such’ basis. That is to say, if a measure, which the government is required to apply, necessitates a WTO violation, the measure itself may be challenged and not just its specific application. Initially, the distinction was used by Panels to reject challenges to measures that did not mandate a GATT violation, particularly where the government had discretion that it had not yet exercised.Footnote 6 This made sense, as it was a sort of ‘ripeness’ requirement.Footnote 7 In recent times, the distinction has been perverted into a suggestion that no measure can be challenged on an ‘as such’ basis if the government has the discretion not to violate WTO/GATT rules. Thus, in the ‘zeroing’ cases,Footnote 8 the US argued that its practice of zeroing – a practice that it applied in every antidumping investigation where the issue arose (and it arose in most cases) – could not be challenged on an ‘as such’ basis since it was not required by US law to violate WTO rules, notwithstanding the fact that it always exercised its discretion so as to violate them. That makes no sense. Thus, while there is an open issue as how to define what sort of practice constitutes a challengeable measure, the mandatory–discretionary distinction should not be applied to prevent ‘as such’ challenges to consistent patterns of WTO violations.
In the instant case, the question was whether the US applied certain presumptions in a way that they were not rebuttable, such that they should be challengeable ‘as such’. I tend to think that the Appellate Body may have been justified in finding that the Panel's examination of the US practice in the area had failed to establish such an irrebuttable presumption. While one can differ on this issue, it appears to me that there were some problems with the Panel's approach. Mexico had submitted evidence on 306 sunset reviews, but only 21 – or 7% – were considered to be relevant by the Panel in that they were contested. In seven of those 21 cases, no additional evidence was offered to the authorities; in five of them, no attempt was made to show that there was ‘good cause’ to consider additional factors. While one can wonder about the appropriateness of a ‘good cause’ requirement in this situation, it was not challenged, with the result that there were only nine cases on which the Panel could have based a finding that the presumption was irrebuttable. And even in respect of that handful of cases, for the most part the conclusion of the Panel was that the authorities' treatment of arguments that good cause existed was ‘troubling’.Footnote 9 In any event, I think that it is arguable that the Panel did not give sufficient details of its examination of the individual cases to allow one to be satisfied that these nine cases showed that the presumption was irrebutable. In the end, the key fact for the Panel seemed to be that even though the US authorities sometimes said they were open to looking at other factors, in these few cases they did not. Thus, I was not surprised that the Appellate Body found that the Panel had not made the required ‘qualitative’ analysis. In any event, this issue seems quite minor. Even if the US had been found to have violated the WTO rules on an ‘as such’ basis, implementation would have involved either (i) a simple re-writing of the sunset policy bulletin to eschew firmly any irrebuttable presumptions or (ii) the elimination of the offending language and the adoption of more opaque procedures, common in almost all other WTO members that impose antidumping duties.