(ii) The Plate Investigation

6.21. In order to evaluate whether an unbiased and objective investigating authority could properly have determined that the local sales in the Plate investigation were made in won rather than dollars, we must first consider in detail the determination of the DOC and the evidence that was before the DOC at the time it made its determination.

6.22. The arguments of the parties and conclusions of the DOC on the issue of local sales are addressed in detail in the final determination in the Plate investigation. We will quote from this determination at length:

"Comment 10. Local letter of credit sales. . . . . [R]espondent maintains that its calculation of normal value should be based on the US dollar price at which the local sales were invoiced. Respondent states that local customers pay POSCO in Korean won based on the US dollar invoiced price. Moreover, respondent notes that it reported the US dollar price for local sales consistent with the Department's requirement and practice. Respondent explains that in Fresh Cut Roses from Colombia, the respondent also invoiced home market sales in US dollars but like POSCO, received payment from the customer in the home market currency, pesos in that case. . . . . POSCO states that in this case, the Department accepted the US prices for the calculation of the normal value . . . . Respondent contends that not using the US dollar value in its calculation of normal value in this investigation could have a potentially significant distortive effect on the margin.

[Petitioners] object to POSCO's request to use nominal dollar prices for home market customers. Instead, they recommend that the Department use the won prices that the customers actually pay. They argue that it would be bad policy to use nominal prices in the margin analysis. Further, they continue that even if the Department were to find that in some instances use of the nominal price is warranted, the facts in this case do not support such a methodology. Petitioners allege that the Fresh Cut Roses from Colombia case cited by POSCO differs from this case in several important respects. Specifically, in Fresh Cut Roses from Colombia: (a) the effect of inflation in Colombia was being taken into account in the Department's cost of production analysis and costs were being converted to dollars; (b) the Department stated that it had verified that the payment in pesos had reflected the prevailing dollar/peso exchange rate at the time of payment; and (c) all home market sales were invoiced in dollars and paid in pesos . . . . In contrast, they note that the Department did not verify whether the exchange rates used were proper. Finally, they note that in this case, home market sales were also quoted in won. Therefore, because the Department has not accounted for inflation and did not verify the dollar won exchange rates used, petitioners argue that POSCO's dollar prices are meaningless because POSCO's customers pay in won.

Department's Position. . . . [W]e disagree with Respondent that the Department should use the US dollar invoiced price for the purposes of calculating normal value. Based upon the facts of the record, as discussed below, we find that it is more appropriate to use the won price in which the customer pays.

For HM sale number 1, POSCO provided an internal document which showed the exchange rates used by POSCO to convert US dollar prices into Korean won prices for the month of November 1997. . . . . The record indicates that although customers are invoiced in US dollars (for HM Channel 2 sales the shipping invoice also shows the won price), the customer pays in won not US dollars, and the sales value of the merchandise is charged to the sales ledger in won, based on the aforementioned exchange rate. . . . Moreover, a comparison of the internal exchange rate used by POSCO to the market exchange rate used by the Department shows that the two exchange rates are quite dissimilar (See Analysis Memorandum). We note that this is in contrast to Fresh Cut Roses from Colombia in which the Department verified that the payment in pesos reflected the market exchange rate at the date of payment . . . . Therefore, for the final determination, we have used the won price for home market sales³².

6.23. On the basis of this discussion, we can perceive the core issue faced by the DOC and the manner in which it resolved it. It will be recalled that POSCO initially reported local sales to be won sales but subsequently reported them in dollars. In considering how to treat these sales, the DOC had before it evidence that local sales were ordered and invoiced in dollars (and, on some invoices, also in won) but paid in won. The question it faced was whether it should consider the sale to be a sale in dollars in the amount shown on the invoice (such that no currency conversion would be required) or a sale in won based upon the amount charged to the sales ledger. The DOC concluded that use of the won amount was appropriate because the customers paid in won, the merchandise was charged to the sales ledger in won, and the exchange rate used by POSCO to calculate the won equivalent of the dollar amount was dissimilar to the "market" exchange rate used by the DOC.

6.24. The heart of Korea's argument that these sales should have been treated as dollar rather than won sales is clearly stated in its first submission:

"As discussed in the Statement of Facts, POSCO had a significant quantity of "local sales" of both SSPC and SSSS during the investigation periods. These "local sales" were negotiated and invoiced in US dollars, but the payments were made in Korean won. Significantly, to ensure that payment accurately reflected the actual dollar value of the sales, the amount of the Korean won payment for these "local sales" was not fixed at the time of the sales negotiation or at the time of invoice. Instead, the payment in Korean won was determined by applying the market exchange rate (as announced by the official Korean Exchange Bank) for the date of [payment]³³ to the US dollar amount shown on the invoice. Thus, the economic reality is that the final payment for these sales is determined by the US dollar amount shown on the invoice, and not by the Korean won amount recorded in POSCO's accounting records at the time of invoice. In economic terms these "local sales" are equivalent to sales that are invoiced and paid in US dollars.[footnote omitted³⁴]

This argument � that the transactions in economic terms were dollar transactions because the amount of Korean won actually paid was established by converting the dollar amount invoiced by the exchange rate at the time of payment, rather that at the time of invoice � is repeated frequently by Korea in its submissions³⁵.

6.25. We consider that the reasoning advanced by Korea is compelling. The issue in the Plate investigation whether the local sales in question were dollar or won sales arose because the invoices were expressed in dollars³⁶ but the payments were made in won. Thus, if the amount of won actually paid was based on the dollar amount identified in the invoice at the market rate of exchange on the date of payment (which, because the local sales in question were letter of credit sales, came some months after the date of invoice), then the controlling amount would be the dollar amount appearing in the invoice. This was in fact a key element in the reasoning of the DOC in Fresh Cut Roses from Colombia³⁷ , an anti-dumping investigation cited by both the petitioners and respondents in the Plate investigation and discussed by the DOC in its final determination, as quoted above. In that case, the DOC "established that respondent invoiced its home market customers in US dollars and received the equivalent value in pesos at the date of payment".(emphasis added) The DOC "found it appropriate that respondent's home market sales were reported in dollar value since the dollar value was the currency in which the sales transactions were made. Furthermore, since home market sales were transacted in dollars and the payments made, although in pesos, were based on constant dollar value, there is no distortion³⁸."

6.26. Korea's argument, however, is dependent upon a particular factual predicate: that the amount of won actually paid in respect of local sales was determined by applying the market rate of exchange at the time of payment to the dollar amount stated in the invoice. Korea asserts, and the United States does not appear now to dispute, that the amounts paid were established based upon the application as of the date of payment of exchange rates established by the Korean Exchange Bank. The DOC was not however aware at the time of its final determination in the Plate investigation that this was the case. Rather, it is evident to us that the DOC considered at the time of its final determination that the actual won amount paid was the amount charged to the sales ledger, which in turn was derived by converting the dollar amount invoiced to won at the Korean Exchange Bank exchange rate prevailing at the time of invoice, and that there was no evidence in the record suggesting to the contrary.

6.27. In response to a question from the Panel, Korea acknowledges that "it is unaware of any evidence in the record of the Plate case that indicates that POSCO specifically informed the DOC that the amount paid in won differed from the converted won amounts shown in the invoice³⁹." It contends however that, once POSCO had informed the DOC that local sales were denominated in dollars, the burden was on the DOC to ask for the necessary information. We consider this argument to be unpersuasive. This is not a case where the DOC did not know the actual won amounts paid and did not bother to ask. Rather, POSCO in its initial questionnaire reported local sales in won, and the amount reported was the invoiced amount rather than the � in most cases probably substantially higher � amount actually paid⁴⁰. Even after POSCO reported dollar prices for the local sales and argued that those dollar prices should be used in place of the won prices, it never corrected the initial misimpression that the won amount reported was the amount actually paid⁴¹. Finally, the DOC did not become aware at verification that the won amounts paid might be different from the won amounts charged to the sales ledger because customers paid on a rolling basis such that verification of payments on an individual basis was not possible⁴².

6.28. We now turn to a review of the DOC's determination on this issue in light of the facts as understood by the DOC on the basis of the record before it. As the DOC explained in its determination, it had verified certain local sales. For these sales, while the customers were invoiced in dollars (and sometimes also in won), payment was made in won, and the value of the merchandise was charged to the sales ledger in won, based upon the exchange rate prevailing on the date of invoice⁴³. The Plate final analysis memorandum that underlies the final determination indicates that the date of payment for these sales however was some months later⁴⁴. In other words, the record before the DOC indicated that the won amount the customer would pay on a given date was fixed some months earlier based upon the won/dollar exchange rate prevailing on the earlier date, irrespective of subsequent changes in the dollar/won exchange rate. The DOC's determination then notes that the exchange rate used by POSCO on the earlier date did not correspond to the market exchange rate used by the DOC. Again, the Plate final analysis memorandum indicates that the DOC was comparing POSCO's exchange rate on the date of invoice both to the DOC's exchange rate on that date and to its exchange rate on the date of payment⁴⁵. The determination notes that "this is in contrast to Fresh Cut Roses from Colombia in which the Department verified that the payment in pesos reflected the market exchange rate at the date of payment" (emphasis added⁴⁶). In short, the DOC concluded that local sales were made in won because the amount to be paid was fixed in won on the date of invoice irrespective of subsequent movements in the won/dollar exchange rate between the date of invoice and the date of payment.

6.29. In challenging the DOC's determination, Korea emphasizes an admitted error by the DOC regarding benchmark exchange rates. Specifically, although the DOC stated that it was comparing POSCO's internal exchange rates to US Federal Reserve exchange rates, it mistakenly compared POSCO's internal exchange rates � which were in fact the Korean Exchange Bank's official rates � to certain adjusted exchange rates prepared by the DOC on the basis of three-month rolling averages. This error however does not detract from the essential fact � as perceived by the DOC at the time it made its determination based upon the record before it � that the won price was determined as of the date of invoice and remained fixed irrespective of any subsequent changes in the won/dollar exchange rate between the date of invoice and the date of payment.

6.30. Korea also has emphasized that the order sheets for local sales, which were verified by the DOC, reflect a dollar amount but no won amount. Korea contends that this supports its view that local sales were negotiated as well as invoiced in dollars. While the fact that order sheets showed sales in dollars and not in won might be a relevant consideration in other circumstances, the fact remains that � based on the record as placed before the DOC � a won amount was fixed at the date of invoice and this won amount was controlling as to the amount to be paid several months later. We recall in any event that in this investigation the DOC determined that the date of invoice rather than the date of order confirmation was the "date of sale" because it was at the time of invoice that POSCO established the material terms of sale. In support of this view, POSCO had argued that "all POSCO's sales were subject to change between order and shipment⁴⁷". Thus, the fact that the orders for local sales were expressed in dollars is in our view less than conclusive as to whether the sales were won or dollar sales.

6.31. For the foregoing reasons, we conclude that an unbiased and objective investigating authority evaluating the evidence before the DOC in the Plate investigation could properly have determined that the local sales in question were made in won.

(iii) The Sheet Investigation

6.32. We now consider whether an unbiased and objective investigating authority evaluating the evidence before the DOC in the Sheet investigation could properly have determined that the local sales were made in won rather than dollars.

6.33. The arguments of the parties and the conclusions of the DOC are set forth in some detail in the DOC's final determination:

"Respondent argues that the Department should calculate normal value for "local" sales made in the home market based on the US dollar price at which those sales were invoiced. Local sales are sales of subject merchandise to home market customers who will further process the merchandise into non-subject products for export. Respondent maintains that although POSCO is paid in Korean won, the amount of payment is based on the US dollar invoiced price. Respondent contends that because POSCO's local sales are denominated and invoiced in US dollars, the invoiced prices do not require conversion to won for US comparison purposes, and that the conversion of the US dollar price to won and then back to dollars is not only unnecessary, but would significantly distort the margin. Respondent cites to . . . Roses from Columbia [sic] . . ., noting that the Department agreed and accepted the US prices for sales invoiced in US dollars, notwithstanding that the respondent received payment from the customer in the home market currency. Respondent argues that in the final determination in SSPC from Korea, the Department's concern was that POSCO's customers paid for sales in won, the sales amounts were recorded in won in POSCO's accounting records, and that the exchange rates utilized by POSCO to determine the won equivalents were different from those exchange rates used by the Department. Respondent contends that the fact that payment is made in won is irrelevant, since both the contract and the invoice reflect a US dollar price, and that sales are converted to won for the purposes of consistency with POSCO's accounting records, which are maintained in won.

Petitioners claim that the use of the dollar value for local sales in the home market would be inappropriate, given that POSCO receives payment in won. Petitioners distinguish this case from Roses from Columbia [sic] by noting that in that case, the Department was factoring in the effects of inflation in the cost-of-production analysis, costs were converted into dollars; the payments in local currencies had reflected the prevailing exchange rate, and all home market sales had been invoiced in dollars and paid in pesos. Petitioners further contend that in Roses from Columbia [sic], the decision to use US dollar-based prices was presumably made for convenience and consistency, as costs were also dollar-denominated. Petitioners further note that the disparity between the exchange rates reflected in the price conversion and the rates used by the Department is too great to reconcile, and is in contrast to the situation in Roses from Columbia [sic]. Petitioners argue that the use of a constant index such as the dollar is used by the Department in the face of currency depreciation or significant deflation, and should not be applied selectively to reduce a dumping margin.

Department's position. We agree with petitioners. First, we believe that respondent's reliance on Roses from Columbia [sic] is misplaced. In that case, all prices and costs, both in the home market and in the U.S., were dollar denominated, and the exchange rates reflected in the dollar to peso conversion coincided with the exchange rates used by the Department. Given these facts, the use of dollar-denominated prices provided consistency throughout the Department's analysis in that case. Neither of these facts are present in the instant case. At verification, we found that local sales are the only sales made in the home market that are expressly linked to a dollar value, but that the sale is ultimately a won-denominated sale. Additionally, the vast majority of the costs incurred for home market and US sales are denominated and paid by POSCO in won. . . . Finally, as we note above, there is a disparity between exchange rates reflected in POSCO's accounting records and those used by the Department . . . . Although the sales are linked to a dollar value, there is no question that the respondent pays in won, and therefore, the use of the dollar-denominated gross unit price for local letter of credit sales in the home market is unwarranted. In addition, in recent cases involving POSCO (e.g., SSPC from Korea and Carbon Steel from Korea � 3rd Review), the Department has used the won-denominated price for local letter of credit sales in the home market because we found that, as in the instant case, the local sales were paid in won and recorded in POSCO's accounting records in won, and the exchange rates used by POSCO were dissimilar from those used by the Department. . . (emphasis added⁴⁸).

6.34. On its face, the issue before the DOC in the Sheet investigation was similar to that in the Plate investigation. As in the Plate investigation, the issue before the DOC in Sheet was whether it should use the dollar-invoiced price for local sales or whether it should use the won price which was charged to the sales ledger and reflected in the tax and certain shipping invoices.

6.35. There was however in our view a significant factual difference between the two investigations. As previously discussed, the record in the Plate investigation suggested that the won prices initially reported by POSCO represented the won amounts actually paid. In the Sheet investigation, on the other hand, the DOC verified, and recorded in its Verification Report, that the won amounts reported by POSCO were not in fact the amounts actually paid:

"Observation [xxxx](HM#1)

This observation represented a local sale from POSCO to [xxxxxx], which involved a recognized loss on foreign currency exchange. We tied the individual sale from POSCO's order sheet to a cumulative shipping list/invoice and tax invoice. As local sales are dollar-denominated, and paid in won equivalents, POSCO records the sale in its won equivalent, reflected in the tax invoice to the customer. On payment, the exchange rate is determined based on the rates given to POSCO by Korean Exchange Bank for Inward Remittance. For this sale, POSCO recognized an exchange rate loss of [xxxx]. This difference between the recorded sales amount and payment amount is reflected in the "Foreign Exchange Currency Loss of Transaction for Local Sales" account. We examined all journal entries, ledgers, bank documents, and publicly available information (re: Korea Exchange Bank, Foreign Exchange Rates), and found no discrepancies⁴⁹."

On the basis of this Verification Report, it seems clear to us that the DOC in the Sheet investigation was fully aware that the won amounts reported by POSCO in respect of local sales were in fact different from the won amounts actually paid.

6.36. The United States argues that "the only suggestion that [the invoiced and paid] amounts differed came late in the proceeding amid conflicting information". It contends that POSCO only reported the invoice price and neither informed the United States nor claimed that the amount paid differed from the amount invoiced⁵⁰. We note however that, unlike in the Plate investigation, POSCO in the Sheet investigation reported dollar amounts ("for ease of verification") in addition to won amounts in its initial questionnaire response⁵¹. This triggered a supplemental questionnaire from the DOC in response to which POSCO provided information suggesting that the won price paid was not the same as that invoiced⁵². More importantly, the sales verification report quoted above was dated 6 April 1999 and related to a verification performed in February. It thus predated the final determination in the Sheet investigation by several months. Moreover, the Verification Report quoted above does not suggest a hint of doubt about the manner in which local sales were handled. A page extracted from POSCO's accounts and attached to the Verification Report listed dozens of exchange rate losses related to local sales⁵³. At this point, therefore, the record clearly showed that the amount of won actually paid in the case of local sales differed from the amount initially reported by POSCO and appearing on POSCO's tax invoices⁵⁴.

6.37. In light of the foregoing, it is difficult to understand the basis for the DOC's statement, in its final determination, that, "[a]t verification, we found that local sales are the only sales made in the home market that are expressly linked to a dollar value, but that the sale is ultimately a won-denominated sale." Leaving aside that we can locate no such "finding" in the Verification Report, we are convinced that an unbiased and objective investigating authority could not so find on the basis of the facts before it in the Sheet case. As we have seen, it was clear from the record in the Sheet investigation that the won price which the DOC considered to be the price in which local sales were denominated was in no sense controlling. Rather, the won amount ultimately paid would be determined by converting the dollar amount appearing on the invoice into won at the rate of exchange prevailing on the date of payment. Thus, the dollar amount appearing on the invoices was controlling, while the won equivalent appearing on the tax and certain shipping invoices and noted in POSCO's accounts⁵⁵ played no role in determining the amount the purchaser ultimately would pay. As explained above, we agree with Korea that there is no logical basis under these circumstances to consider that the sales in question were denominated in won.

6.38. In its determination, and before this Panel, the United States has emphasized that there were differences between the exchange rates used by POSCO and the market exchange rates relied upon by the DOC. We agree that the DOC might have been entitled to disregard the dollar prices stated in the invoices if the won amounts actually paid had been based on fictitious or inaccurate exchange rates, as this would have indicated that the economic value of the transactions was not in fact determined in dollars at all⁵⁶. In this case, however, the DOC verified, and the United States does not now dispute, that the "internal exchange rates" used by POSCO were in fact the official exchange rates published by the Korea Exchange Bank⁵⁷. Further, we consider that the differences between the Korea Exchange Bank rates and the Federal Reserve rates cited in the Sheet investigation⁵⁸ are not the result of fictitious or inaccurate exchange rates, but merely reflect the existence of a 14-hour time difference between New York and Seoul⁵⁹.

6.39. For the foregoing reasons, we conclude that an unbiased and objective investigating authority evaluating the evidence before the DOC in the Sheet investigation could not properly have determined that the local sales in question were made in won.