What's New?
 - Sitemap - Calendar
Trade Agreements - FTAA Process - Trade Issues 

espa�ol - fran�ais - portugu�s
Search

World Trade
Organization

WT/DS99/R
29 January 1999
(99-0256)
Original: English

United States - Anti-Dumping Duty on Dynamic Random Access Memory Semiconductors (DRAMS) of one Megabit or Above from Korea

Report of the Panel

(Continued)


8. Respondents Met the Criteria for Revocation

a) Submission by Korea

1.335 Korea submits to the Panel that the record evidence supported revocation, even applying the US scheme (which it deems improper), and that in support for its failure to revoke the United States mischaracterized the facts on the administrative record. Korea presents the following arguments in support of this submission:

(i) The Record Evidence Supported Revocation, Even Applying the US Scheme.

1.336 Even assuming for the sake of argument, that the US revocation scheme complies with the WTO agreements. Korea will demonstrate once again that the Panel should find that the DOC's establishment of the facts was improper and that its evaluation of the facts was biased and not objective.

1.337 In this regard paragraph 56 of the US first submission states:

56. In its rebuttal submission, Korea will undoubtedly argue that one of the ways the market could have "changed" at this time was for the better. It will cite passages in the administrative record where various investment bankers and industry experts predict higher prices and better times for the DRAM industry in 1997. What it will ignore, is roughly the same number of experts who were not sure which way the market was headed and who openly expressed concerns about its future.218

1.338 In this passage, the United States concedes that the import of the record evidence regarding the market was at least 50-50 in support of revocation.219 In other words, at least half the record evidence supported the conclusion that the market was getting healthier and, in the words of the United States, "roughly the same number of experts . . . were not sure which way the market was headed and . . . openly expressed concern about its future." So, actually, 50 percent of the experts said the market was getting healthier and 50 percent said they were not sure which way the market would go. The United States therefore concedes that the split is not 50-50, but that 50 percent predicted a recovery and 50 percent were not sure and, thus, the evidence is strongly in favor of recovery and revocation. Thus, even if these forecasts were the only evidence on the record, it could not support a "no likelihood/not likely" finding even under the DOC's own regulations.

1.339 But this was not the only evidence on the record. During the 18-19 June 1998 Panel meeting, the United States stated that three consecutive reviews of no dumping (the first criterion) and certification not to dump and to submit to reinstatement of the order (the third criterion) also bore on the "not likely" issue (the second criterion).220

1.340 So, the United States admits that the record evidence in favor of a finding of "not likely" and revocation was, at a minimum:

1. three consecutive reviews (three-and-one-half years) of no dumping by Respondents;

2. a certification by Respondents that they would not dump in the future and would submit to reinstatement if they were to do so; and

3. the considered opinion of at least 50 percent of the experts on the record that the DRAM market would continue to strengthen or, in the words of the United States, that the future would bring "higher prices and better times for the DRAM industry."221

And, in favor of rejecting revocation, the United States points to:

1. a number of experts "who were not sure which way the market was headed and who openly expressed concern"; and

2. a fiction that the companies had dumped during the "last downturn" (in the original investigation) and, thus, would do so again.222

1.341 This summary of the record evidence and US admissions demonstrate that there was not a 50-50 split in the record evidence; to the contrary, the record evidence was and is strongly in Respondents' favor. Thus, even accepting all of the United States' assertions, the United States' basic argument must fall.

1.342 The AD Agreement does not permit the United States to maintain a definitive duty where only a minimal portion of the evidence, at most, supports doing so. In this case, the Korean producers overwhelmingly demonstrated, through their past conduct, through their certification and through expert economic opinion and analysis, that they would not resume dumping if the DOC were to revoke the order. By refusing to revoke, the United States has abused the limited right to impose and maintain anti-dumping remedies granted by the AD Agreement.

1.343 Moreover, the language chosen by the United States to discuss its findings indicates its lack of certainty and shows that it was speculating as to what was "likely" or "not likely" to occur. For example, in the Final Results of the Third Review, the DOC discusses the 1996 downturn and concludes that revocation should be denied, in part, because the pricing pattern it saw was "suggestive of deteriorating market conditions that often give rise to dumping."223 The DOC did not find that the pricing pattern "showed" or even "indicated" deteriorating conditions that, e.g., always give rise to dumping. This type of speculation appears throughout the Final Results and, also, the first US submission. For example, according to the United States:

The agency engaged in a painstaking analysis of voluminous data on the administrative record and only then did it determine that "dumping may have taken place during the 1996 downturn.

First, how "painstaking" could an analysis be that found only that dumping "may" have occurred? In any case, even if the analysis was painstaking, that is not the point. The DOC did not make the finding Article 11 requires. Second, each of the DOC's stated rationales for failing to revoke are speculative: "may," "suggestive," "often."

1.344 This, Korea submits, is not a permissible way to regulate dumping. The United States clearly has violated its WTO obligations under Article 11 of the AD Agreement and Article VI of the General Agreement.

(ii) The United States Has Mischaracterized the Facts on the Administrative Record.

1.345 Korea asks the Panel to examine four core facts in detail should it conclude that the United States has not otherwise violated Article 11. The following are discussed in turn:

  • the United States continues to ignore the downturns that occurred during the First and Third Review Periods;
  • the United States ignores the fact that although the 1996 downturn was severe, prices already had begun to recover in 1997;
  • the United States selected and analyzed record data in an improper, biased and not objective fashion to avoid the implications of the 1996-1997 recovery; and
  • Respondents did, in fact, cut production, but while the United States conveniently recognized Respondents' production cutbacks for one purpose, it then ignored them when it refused to revoke.

(1) The United States Continues to Ignore the Downturns That Occurred During the First and Third Review Periods.

1.346 The United States refused to recognize that the DRAM industry suffered downturns during 1993 and during 1995. The record evidence cited by the United States in its first submission demonstrates that industry downturns occurred during the second half of 1993 and during the latter part of 1995 and into early 1996. These periods were covered by the First and Third Administrative Reviews.

1.347 Korea has never argued, as the United States implies, that the semiconductor industry did not experience net positive growth from the end of 1993 through the middle of 1995, i.e., Korea agrees that the book-to-bill ratio indicated that the market was larger and more robust in the middle of 1995 than it was at the end of 1993. However, Korea has demonstrated that a closer look at the ratios shows that downturns in the industry occurred during the second half of 1993 and from the latter part of 1995 into the second half of 1996, periods that fell within the First and Third Administrative Reviews, respectively.

1.348 The book-to-bill ratio was produced for years by the US Semiconductor Industry Association (SIA) and was universally accepted as an indicator of DRAM market conditions. It was relied upon by industry leaders and analysts studying market trends. SIA abandoned the statistic, not because it was inaccurate, but because: (i) it reflected only the conditions of the US market--the key market in this case (and the DRAM market increasingly was perceived as a global market); and, more importantly, (ii) it was thought to be having an unduly strong and negative impact on the prices of stocks in the industry.224

1.349 In the Final Results of the Third Administrative Review, the United States noted that "[t]he DRAM industry is highly cyclical in nature with periods of sharp upturn and downturn in market prices."225 But, the United States attempts to cloud the cyclicality of the DRAM market by quoting completely out of context a passage from Dr. Flamm's economic analysis regarding the alleged stability of prices during 1993, 1994 and 1995. In fact, the study directly contradicts the US position.

1.350 Dr. Flamm's study rests on a quarter-by-quarter analysis of annual pricing data for 1993 through 1996.226 Dr. Flamm concludes that from the fourth quarter of 1992 through the first quarter of 1996, prices for DRAMs declined in 6 out of 14 quarters (or nearly half of the time). Of course, during this time, the DOC found that neither Hyundai nor LG Semicon were dumping. This led Dr. Flamm to conclude that "even in an environment of falling prices (and falling by quite a lot in the first half of 1996) [Respondent] was not dumping, as verified by the DOC."227

1.351 Moreover, these quarterly pricing data fully support the downturns indicated in the book-to-bill ratios. Thus, regardless of whether the DRAM industry experienced overall growth between 1993 and 1995, that same time period also was marked by at least two clear and distinguishable downturns during which the DOC found that LG Semicon and Hyundai were not dumping.

1.352 The United States also relies on a Merrill Lynch report to claim that "Korea overlooks the fact that a downturn in the market may not manifest itself for many months following a low point in the book-to-bill ratio." But the United States ignores the fact that the same Merrill Lynch report uses quarterly sales data that demonstrate that downturns occurred in 1993 and 1995.

1.353 All of the relevant data--including those cited by the United States in this proceeding--demonstrate that downturns occurred during periods covered by the First and Third Administrative Reviews. The DOC found Respondents did not dump during these periods. So, contrary to the DOC's assertion in this proceeding, Respondents did demonstrate that they do not dump during downturns.

(2) The United States Ignores the Fact That Although the 1996 Downturn Was Severe, Prices Already Had Begun to Recover in 1997.

1.354 The 1996 downturn was more severe than the 1991 downturn. However, the 1996 downturn began in late 1995 (during the Third Review period) and Respondents showed that the market conditions improved in early 1997.

1.355 The United States attempts to extrapolate from the 1991 downturn to the 1996 downturn. However, the United States would have the Panel ignore, first, that although some analysts remained uncertain or hesitant about the future of the DRAM market in 1997, many analysts confidently asserted that because, unlike the 1991 downturn, the 1996 downturn did not occur during an economic recession in the United States, but, to the contrary, occurred during an unprecedented period of growth, the recovery would be swifter and well underway by 1997.

1.356 Second, the United States would have the Panel ignore that, in any event, because the 1996 downturn began at the end of 1995, it was covered by the Third Review, during which the United States found that LG Semicon and Hyundai were not dumping. The United States can not have it both ways. If it insists that the end of 1996 was characterized as a downturn, it also must acknowledge that the Third Review covered a downturn and, thus, that the DOC had found that Respondents did not dump during the most recent downturn.

(3) The United States Selected and Analyzed Record Data in an Improper, Biased and Not Objective Fashion to Avoid the Implications of the 1996-1997 Recovery.

1.357 In discussing DRAM market conditions from 1996 and projecting them into 1997, the United States continued its pattern of selectively choosing and omitting portions of the administrative record in order to create a biased and inaccurately negative market assessment. For example, the United States quotes out of context a passage from a De Dios & Associates report, which states that in regard to the recovery in the DRAM market "[w]hat we have here is a temporary situation that will change."

1.358 When read in context, however, the passage has an altogether different meaning. In particular, it does not support, but, in fact, contradicts the US position:

What we have here is a temporary situation that will change. But change does not necessarily mean a plunge in prices similar to last year's behaviour. Many other analysts are too quick to offer this as the only other alternative. How it will change requires a deeper understanding of the market forces that affect price.228

The De Dios analysis then goes on to consider the impact of various market forces, concluding that prices should continue to increase.229

1.359 In short, the United States has selected fragments from the De Dios and other analyses. A careful reading of those analyses, however, reveals that the vast majority of the evidence on the record indicated the market had strengthened and was continuing to strengthen.230

1.360 Finally, the administrative record overwhelmingly supports the fact that the predicted recovery for 1997 would not be due entirely to reports of cutbacks in production by Korean producers. Again, the United States misleadingly cites to the De Dios analysis. But, one need only review the multiple analyses on the administrative record, including the De Dios analysis, to understand that many market forces were driving the market upward. These factors included, in addition to the strong US economy and many other factors, increases in the production and memory content of PCs.231

(4) Respondents Did, in Fact, Cut Production, But While the United States Conveniently Recognized Respondents' Production Cutbacks for one Purpose, it then Ignored them when it Refused to Revoke.

1.361 According to the United States, "a careful reading of the administrative record casts a large shadow of doubt over the 'production cutbacks' announced by the Korean producers."232 Actually, a "careful reading" of the record reveals that eight days before issuing its Final Results, the DOC acknowledged that Respondents had, indeed, cut production. In order to undercut Dr. Flamm's study so as to provide a rationale for rejecting its conclusions, the DOC stated in its internal decision memorandum, "in light of the announced cutbacks in DRAM production, and the recent announcements of actual factory shutdowns, it is difficult to accept the [confidential data omitted] scenario."233

1.362 Now the United States argues before this Panel, as it did in the Final Results, that Respondents did not cut production. But, if the administrative record allegedly proves that Respondents did not reduce production, why did the DOC acknowledge the opposite just a few days before issuing the Final Results? The United States apparently reached the conclusion that cutbacks had occurred to criticize and reject the results of the Flamm study, and then, eight days later, claimed that cutbacks had not occurred in order to create a negative picture of the DRAM market and to reject revocation. The Panel should not accept this transparently biased and not objective assessment of the record evidence.

To continue with Response by the United States


218 The Panel notes that this argument is set forth in Paragraph 4.412 of this report.

219 Elsewhere, the United States concedes that a recovery was underway, noting its opinion that the recovery "was, at best, bumpy."

220 The United States thus effectively concedes that it is in violation of its GATT Article X obligations, because the relevant US regulation (19 C.F.R. � 353.25(a)) does not reflect this methodology. Korea also notes that the US imposition of the certification and "no likelihood/not likely" requirements is incoherent. To test the likelihood of future dumping after obtaining a certification makes no sense--the certification is legally binding on the Respondents and, thus, is definitive, but the likelihood test is based largely on speculation.

221 There were other facts on the record supporting revocation, including the fact that the Korean producers had weathered two downturns without dumping.

222 The United States has misstated a number of facts, including the condition of the market during the First and Third Reviews and Respondents' production cutbacks, in order to try to buttress its failure to revoke.

223 See DRAMs from Korea, 62 Fed. Reg. 39809, 39817 (24 July 1994) (emphasis added by Korea) (Exh. ROK 3).

224 See "Bye-Bye B:B Oct's Rebound," Electronic Buyer's News, 18 November 1996 (Ex. ROK-86). (The Electronic Buyer's News article not only demonstrates that market analysts widely accepted the book-to-bill ratio as an accurate indicator of market conditions and shows the real reasons SIA abandoned it, but also shows that the end of 1996 coincided with an upturn in the market.)

225 See DRAMs from Korea, 62 Fed. Reg. 39809, 39810 (24 July 1994) (Ex. ROK-3).

226 See Dr. Kenneth Flamm, Economic Analysis of 16 Megabit DRAM Costs and Pricing: Projections for 1997 and 1998 (Revised and Supplemented), April 1997, Figure 3 (Ex. ROK-35, Attachment 2).

227 Id. at page 6.

228 See Ex. ROK-35, Attachment 5 at 3 (emphasis added by Korea).

229 Id. at 3-5.

230 According to the US First Submission, the DOC determined that: (i) contract prices follow the direction of spot market prices; and (ii) Respondents' contract prices fell below spot market prices in February 1997. These assertions are not supported. Moreover, though the United States implies that they are correct, for the reasons presented below, they are not. First, contract and spot prices obviously correlate. But this most certainly does not mean that contract prices fell below fair value (or fell faster than costs). Second, unrebutted evidence on the record shows that the gap between Respondent's contract and spot market prices was quite large during the period of review and remained large. Finally, the United States claims that the DOC determined that "contract and spot market prices were rapidly trending below Respondents' reported costs of production throughout the period that immediately followed the third administrative review." Again, the DOC's determination is incorrect. Even worse, the DOC's sole support for this determination was a chart that showed that Respondent's prices actually remained well above its falling costs of production.

231 See Case Brief of Hyundai, Case No. A-580-812 (21 April 1997), at Attachments 5, 12 and 13 (Ex. ROK-35).

232 The Panel notes that this argument is set forth in Paragraph 4.422 of this report.

233 See Analysis of Data, DRAMs from Korea, US DOC of Commerce, Case No. A-580-812 (16 July 1997) (Ex. ROK-87).