c) Rebuttal Arguments Made by Korea

1.431 Korea in its oral statement at the first meeting of the Panel with the Parties, further argued as follows:

1.432 The United States improperly established the facts; and evaluated the facts in a biased and non-objective manner, in violation of the standards set forth in Article 17.6. The United States violated its obligation under Article 6.6 by failing to satisfy itself as to the accuracy of data supplied by Petitioner. Indeed, the United States uncritically accepted and relied on Petitioners' data without taking any action to confirm that it was accurate. The United States disregarded cost data prepared by Respondents which were in accordance with generally accepted accounting principles of Korea and accurately reflected costs, thereby violating its obligation under Article 2.2.1.1. Finally, to support its decision regarding the likelihood of dumping in the future, the United States disregarded a valid econometric study on DRAM cost and pricing and Respondent-specific cost and pricing data and relied instead on irrelevant spot market pricing and speculative assumptions of future costs supplied by Petitioner. Also, the DOC disregarded Respondents' data collection proposal, apparently assuming (incorrectly) that it had no bearing on the likelihood issue.

d) Rebuttal Arguments Made by the United States

1.433 The United States makes the following rebuttal arguments:

(i) The United States Appropriately Satisfied Itself as to the Accuracy of the Data Relied Upon

1.434 Korea's generalized allegation that the DOC relied upon data submitted by Micron without satisfying itself as to the accuracy of the data is baseless. The DOC appropriately evaluated all of the information gathered in the underlying administrative proceeding, and relied upon information submitted by both Respondents and by Micron in reaching its determination concerning the likelihood of future dumping.

1.435 First, the DOC conducted on-site verifications of the cost and sales data submitted by LG Semicon and Hyundai in their questionnaire responses for the May 1995 to April 1996 third review period (including home market cost and price data through June 1996).³³³ As discussed in the first US submission, the DOC relied on this verified cost and price data in its determination not to revoke, when it concluded: (i) that Respondents had made a substantial number of sales below cost in Korea in May and June 1996, as the rapid and continuing decline in DRAM prices drove their prices below cost; and, (ii) that Respondents' verified cost data for 1995 and the first half of 1996, extrapolated forward to the second half of 1996, indicated that Respondents may have already resumed dumping in the latter half of 1996.³³⁴

1.436 Second, the DOC satisfied itself as to the accuracy of the information it relied upon concerning developments following the end of the third-review period. This information, submitted by the Respondents as well as by Micron and other interested parties, included independent market analysts' reports from such reputable brokerage houses as Goldman Sachs, Merrill Lynch, Lehman Brothers, and ABN Amro Hoare Govett; business and market news reporting by well-known news organizations such as the Wall Street Journal, New York Times, Financial Times, Reuters, Korea Herald, and Nikkei; and reports from various trade journals.³³⁵ As the DOC noted in the Final Results Third Review, the Respondents and their customers submitted data on average US prices reported by Dataquest and the American IC Exchange, studies by independent analysts and numerous newspaper and magazine articles.³³⁶

1.437 The DOC did not accept such reports uncritically, but instead appropriately evaluated the factual assertions made by the interested parties, and satisfied itself as to the accuracy of the information on which it relied. For example, as discussed in the first US submission, the Korean Respondents contended that the prevailing excess of DRAM supply over demand would be ameliorated in the future by production cutbacks the Korean producers had announced on 30 January 1997. However, while their own production figures were obviously available to them, the Respondents did not provide any data to substantiate their assertion that they had, in fact, cut back production. Instead, in their 18 April 1997 case briefs before the DOC, they relied exclusively on news reports concerning their earlier announcements.³³⁷ The DOC appropriately discounted the news reports of production cutbacks which were not substantiated by data readily available to Respondents.

1.438 Similarly, the DOC carefully considered each of the parties' contentions concerning both the reliability of the submitted information and the appropriate conclusions to be drawn from it.³³⁸ For example, the DOC specifically addressed LG Semicon's argument that the DRAM pricing data series compiled by Lehman Brothers (a well-known US brokerage house) and submitted by Micron was unreliable, noting that [t]hese data were similar to other pricing data submitted on the record, including the pricing data obtained from the American Integrated Chip Exchange (AICE) and Dataquest.³³⁹ In each case, the DOC applied its considerable experience in market analysis and considered the source of the information, its internal logic, and its consistency with other information in determining its accuracy and usefulness. In short, there simply is no basis for Korea's claim that the DOC did not satisfy itself as to the accuracy of the data that it relied upon.

(ii) The United States Properly Evaluated All Cost Data Submitted by the Korean Respondents

1.439 Korea alleges that the United States violated Article 2.2.1.1 of the AD Agreement by disregarding cost data prepared by Respondents, which, it claims, was prepared in accordance with Korean GAAP and accurately reflected costs. The record is clear, however, that the DOC verified the cost data submitted by both Respondents as it pertained to cost of production for the third-review period (covering quarterly costs from the first quarter of 1995 through the second quarter of 1996). The DOC relied on this verified cost data in considering whether Respondents may have resumed dumping in the months following the end of the third-review period.

1.440 In addition to the cost data for the third review submitted by both Respondents, late in the proceeding, LG Semicon also submitted information which purportedly represented its cost of production for two DRAM models in the second half of 1996. These data were not submitted in the same format as the data in the third review questionnaire response, with underlying detail and copies of the company's financial statements for the period, nor were they provided in the form of hard figures of the kind that could be verified by reconciliation to financial and cost records. Instead, the claimed cost of production information was submitted in the form of data points on a graph, with an express disclaimer that [t]he reported figures were calculated based on best information currently available.³⁴⁰

1.441 In addition, numerous news reports and market analysts' reports were submitted which indicated that LG Semicon had made several changes in accounting practices, including changes in the accounting for depreciation and for foreign exchange losses.³⁴¹ These reports indicated that LG Semicon had a loss of 40 billion won in the second half of 1996, and would have reported a loss for the year if it continued to use the accounting methods it had used in the prior year.³⁴² The DOC noted these changes, and further noted the fact that LG Semicon failed to identify these adjustments to its costs significantly reduces the reliability of the information. We are uncertain whether LG Semicon made other adjustments to its reported costs. ³⁴³

1.442 Despite LG Semicon's claim that its late-submitted 1996 cost data were prepared in conformity with its financial statements, LG Semicon never submitted its 1996 financial statements to allow the DOC to confirm LG Semicon's assertion, or to determine whether any other accounting changes had been made. Under these circumstances, the DOC properly concluded that, taking into consideration LG Semicon's verified cost data for the first half of 1996, that the piecemeal data submitted for the second half of 1996 could not be accepted at face value. The DOC was under no obligation to accept the incomplete cost data in the format submitted by LG Semicon when it contained undisclosed changes in accounting methods and an unsubstantiated claim that it was prepared in accordance with its normal accounting records.

(iii) The United States Appropriately Discounted Hyundai's Economic Study Containing Unrealistic Projections of Costs and Prices and Respondents' Unsubstantiated Cost and Pricing Data

1.443 Korea also claims that the DOC improperly disregarded an economic study submitted by Respondent Hyundai that purported to show, based on a number of assumptions concerning the trend in prices and costs, that Hyundai's prices would remain above its costs in the second half of 1997 and 1998. The DOC did not ignore the Flamm study. To the contrary, in the Final Results Third Review, the DOC carefully considered Respondent's submissions, but concluded that the unrealistic assumptions on which the study was based (which assumptions were contradicted by other data submitted by Respondents and by pricing data submitted by their computer customers) rendered the study's projections unreliable.³⁴⁴ Indeed, as noted by the DOC, those optimistic projections had already proved incorrect by the pricing trends and analysts' reports in June 1997.³⁴⁵ While baldly asserting that the Flamm study was a valid econometric study, Korea does not address any of the specific flaws in its underlying assumptions that the DOC identified in its determination.

1.444 Korea also complains that the DOC disregarded the Respondent-specific cost and pricing data submitted by the companies. The DOC in fact relied upon the verified cost and pricing data for the third-review period that Respondents had submitted in their questionnaire responses. In addition, one Respondent, LG Semicon, submitted monthly averages for selected cost and pricing data. The DOC properly discounted that data (which in itself showed that LG Semicon's earlier projections were erroneous).

1.445 Korea further asserts that Respondents' data (which were verified by the DOC) showed that their contract prices were above spot prices. However, Korea does not cite any specific evidence in the record to support its assertion, apparently relying generally on its earlier reference to the data submitted by Respondents in their 18 April 1997 case briefs. The DOC properly evaluated Respondents' data, in light of all the evidence on the record including actual pricing data submitted by the OEM computer customers themselves and concluded that Respondents' contract pricing generally followed the pricing trends in the spot market.³⁴⁶

1.446 LG Semicon submitted two charts in its 18 April 1997 case brief (at 49-50, Figures 5 and 6) in support of its argument that its contract prices to OEMs during the third-review period were above the average US spot-market price. Those charts actually support the DOC's conclusion that Respondents' US prices followed the declining trend of spot prices, and that the extrapolation of those trends indicated that Respondents already may have made sales below cost during 1996.

1.447 To the extent that the charts show a lag between LG Semicon's prices and the rapidly declining spot prices in the second quarter of 1996, it reflects a fundamental distortion in the way the data were presented by LG Semicon. The charts present LG Semicon's quarterly average prices and costs for 4 meg and 16 meg DRAMs from the first quarter of 1995 through the second quarter of 1996, based on US prices and cost data submitted in the third review. However, because the third review covered US sales made during the period 1 May 1995 through 30 April 1996,³⁴⁷ what is shown in the chart as the average US price for the second quarter of 1996 actually represents sales for just the first month (April 1996) in the quarter. In a rapidly declining market, using prices for just the first month of a quarter to represent a quarterly average badly skews the comparison. This error was compounded when LG Semicon uses these quarterly averages to derive trends on which it based cost and price projections for the second half of 1996.³⁴⁸

1.448 The distortion in the calculation of LG Semicon's average DRAM prices for second quarter 1996 was demonstrated by the additional pricing data submitted by LG Semicon, purporting to show actual monthly average US prices for May to December 1996.³⁴⁹ The monthly US pricing data for May and June show the gross distortion reflected by the quarterly average data for the second quarter of 1996, as well as in the projections for the second half of 1996 that are based on a trend using the distorted quarterly average.³⁵⁰

(iv) The United States Properly Relied on Publicly Available Reports Concerning Market Price Trends and Projections of Costs Based on Respondents' Verified Cost Submissions

1.449 Korea contends that the DOC improperly relied on publicly reported spot-pricing data in considering pricing trends and price levels, arguing that spot prices are irrelevant to the assessment of the market-pricing trends. Similarly, Korea asserts that the DOC improperly utilized the cost projections submitted by Micron, which were based on Respondents' own verified cost data for the third-review period. The United States reaffirms the discussion in its first written submission which showed that Korea's arguments are groundless.

1.450 The DOC carefully considered Respondents' arguments against reliance on independent analysts' spot-market pricing data in making determinations concerning the pricing levels and trends in the DRAM market. Respondents claimed that their sales to OEM customers, which Korea now asserts were made according to long-term contracts, did not precisely track spot-market prices.

1.451 First, the evidence does not support Korea's assertion that Hyundai and LG Semicon sold to US customers pursuant to long-term contracts. While Respondents may have reached periodic understandings on prices with their OEM customers, the actual prices were set in individual purchase orders, and were subject to change. As stated in their questionnaire responses, Respondents reported the invoice date (generally coincident with the date of shipment to the customer) as the date of sale.³⁵¹ Under the DOC's well-established practice, if all of the material terms of a sale are fixed in a legally binding long-term contract including, most importantly, the product, the price and quantity of the sale then the DOC would use the date of the long-term contract, not the invoice date, as the date of sale in its dumping analysis.³⁵² Neither Respondent indicated in its verified questionnaire response that it sold pursuant to such long-term contracts, and the DOC used the date of invoice as the date of sale.

1.452 Second, the evidence showed that even those OEM customers that had non-binding understandings regarding pricing over longer periods abandoned such understandings as the DRAM market price continued its dramatic fall throughout 1996. A report on LG Semicon from the brokerage house ABN Amro Hoare Govett, prepared in November 1996, stated:

1.453 Consequently, while LG Semicon might have been in a position to obtain some price premium over spot-market prices at the beginning of 1996, this situation clearly evaporated by the second half of 1996, as longer-term price agreements were displaced and OEM customers had ample access to an excess supply of DRAMs at even lower prices.

To continue with Claims under Article X:1 and X:2 of GATT 1994

³³³ Final Results Third Review, 62 Fed. Reg. at 39815-17 (Ex. USA-1).

³³⁴ Final Results Third Review, 62 Fed. Reg. at 34817 (Ex. USA-1).

³³⁵ See, e.g., LG Semicon Case Br. at Appendix B (submitting 13 industry research studies) and Appendix C (submitting 18 news reports) (Ex. USA-15 & 96); Case Brief of Hyundai Ex. 5, 10, 15 (market analysts' reports) and Ex. 7-9, 11 (news reports) (Ex. USA-12); Micron Rebuttal Br. Ex. 1 (Goldman Sachs) and Ex. 2, 3, 7-9 (news reports) (Ex. USA-26 & 97).

³³⁶ Final Results Third Review, 62 Fed. Reg. at 39814 (Ex. USA-1).

³³⁷ Case Brief of Hyundai at 19-20 and Ex. 7 (Ex. USA-12 & 98); LG Semicon Case Br. at 66-69 (Ex. USA-99). In this regard, it is worth noting that in the Indonesia Autos case, the panel justified a negative ruling of market displacement or impedance under Article 6.3 of the Agreement on Subsidies and Countervailing Measures on the failure of the complainants to provide evidence that, according to the panel, was at the disposal of the companies in question. Indonesia - Certain Measures Affecting the Automobile Industry, WT/DS54/R, WT/DS55/R, WT/DS59/R, WT/DS64/R, Report of the Panel issued 2 July 1998, para. 14.234 (unadopted).

³³⁸ See Final Results Third Review, 62 Fed. Reg. at 39817 ( We have reviewed the data submitted by the petitioner as well as all arguments and information on the record regarding the veracity of the data and the underlying assumptions. . . . [M]any of the Respondents' arguments concerning the alleged distortions and inaccuracies in the petitioner's analysis lack merit. ) (Ex. USA-1).

³³⁹ Id. at 39818. See also Micron Rebuttal Br. Ex. 6 (Ex. USA-97); Case Brief of Hyundai Ex 5 (Ex. USA 12).

³⁴⁰ LG Semicon Case Br. at 57, n. 21 (Ex. USA-99).

³⁴¹ Final Results Third Review, 62 Fed. Reg. at 39818 (Ex. USA-1).

³⁴² Micron Rebuttal Br. at 8-10 and Exhibits 2 & 3 (Ex. USA-97).

³⁴³ Final Results Third Review, 62 Fed. Reg. at 39818 (Ex. USA-1).

³⁴⁴ Final Results Third Review, 62 Fed. Reg. at 39818 (Ex. USA-1). See also Micron Rebuttal Br. at Ex. 5 (Critique of Flamm Analysis) (Ex. USA-97).

³⁴⁵ Id. ( current market conditions do not bear the strong demand assumption out ).

³⁴⁶ Final Results Third Review, 62 Fed. Reg. at 39817-18 (Ex. USA-1).

³⁴⁷ By contrast, pursuant to its standard questionnaire format, the DOC had collected information on Respondents' costs and sales in the home market through June 1996, as potential comparisons for the reported US sales for the period through April 1996.

³⁴⁸ LG Semicon Case Br. at 52-54, Figures 7 and 8 (Ex. USA-99).

³⁴⁹ Id. at 58-59, Figures 9 and 10 (Ex. USA-99). These unsubstantiated data, first submitted late in the proceeding with the case briefs, were submitted only in the form of data points on a graph showing simple monthly weighted average prices, an averaging methodology that is not consistent with the DOC's practice.

³⁵⁰ Compare LG Semicon Case Br. at 52-54 with id. at 55-59 (Ex. USA-99).

³⁵¹ See, e.g., LG Semicon Questionnaire Response, Section C, 16 August 1996, at 7 ( LG Semicon has reported invoice date as date of sale ) (Ex. USA-100); id., Section A, at 15 ( A variety of changes occur after the initial {purchase order} that affect the terms of the sale. Price. As a result of changing market conditions, the customer may demand a lower price. . . . ) (Ex. USA-100); Hyundai Questionnaire Response, Section A, 19 August 1996, at A-10 ( the product, unit price and quantity are shown in the purchase order from the customer. Once a sales order has been issued, the terms, i.e., price, product, and quantity, may be modified. If so, then a new sales order is issued. ) (Ex. USA-101).

³⁵² See Anti-dumping Duties; Countervailing Duties; Final Rule, 62 Fed. Reg. 27296, 27348-50 (Ex. USA-102); id. at 27411 (codifying practice in new regulation, 19 CFR � 351.401(i)) (9 May 1997) (Ex. USA 102)

³⁵³ Letter from Micron Technology, Inc. to Assistant Secretary LaRussa, 28 January 1997, at Ex. 1, p.18 (emphasis added by the United States) (Ex. USA-103).