12.4.2006   

EN

Official Journal of the European Union

L 103/1

(1)

By Regulation (EC) No 1480/2003 (3) (the definitive Regulation), the Council imposed a definitive countervailing duty of 34,8 % on imports of certain electronic microcircuits known as dynamic random access memories (DRAMs) originating in the Republic of Korea and manufactured by all companies other than Samsung Electronics Co., Ltd (Samsung), for which a 0 % duty rate was established. The definitive Regulation was preceded by Commission Regulation (EC) No 708/2003 of 23 April 2003 imposing a provisional countervailing duty on imports of certain electronic microcircuits known as DRAMs (dynamic random access memories) originating in the Republic of Korea (4) (provisional Regulation).

(2)

Two exporting producers located in the Republic of Korea cooperated in the investigation which led to the imposition of the existing measures (the original investigation): Samsung, which was not found receiving subsidies, and Hynix Semiconductor Inc. (Hynix). The European Community (EC) industry in the original investigation consisted of two producers which accounted for the major proportion of the total Community production of DRAMs, Infineon Technologies AG, Munich, Germany (Infineon) and Micron Europe Ltd, Crowthorne, United Kingdom (Micron).

(3)

The investigation of subsidisation and injury covered the period from 1 January 2001 to 31 December 2001 (IP). The examination of trends relevant for the assessment of injury covered the period from 1 January 1998 to the end of the IP (the period under consideration).

(4)

On 3 August 2005, the Dispute Settlement Body (DSB) of the World Trade Organisation (WTOmm) adopted a Panel report in ‘European Communities — Countervailing measures on dynamic random access memory chips from Korea’ (5).

(5)

According to the Panel report, the EC acted in a manner inconsistent with its WTO obligations under:

(6)

On the other hand, the EC was found to have acted consistently with its WTO obligations under:

(7)

In the light of the recommendations set out in the Panel report, the Commission conducted a thorough and objective reassessment of the relevant findings in the definitive Regulation on the basis of information on the record. The scope of the reassessment is appropriately limited to an examination of the five specific aspects of the existing measure, as described above, that were found to be inconsistent with the SCM Agreement. This examination is based on the re-evaluation of the existing evidence on the record of the original investigation in light of the relevant findings in the Panel report.

(8)

The reassessment set out below shows that injurious subsidisation still exists, although at a slightly lower level.

(9)

The product under consideration and the like product are the same as that covered by the original investigation, i.e. certain electronic microcircuits known as dynamic random access memories (DRAMs), of all types, densities and variations, whether assembled, in processed wafer or chips (dies), manufactured using variations of metal oxide-semiconductors (MOS) process technology, including complementary MOS types (CMOS), of all densities (including future densities), irrespective of access speed, configuration, package or frame etc., originating in the Republic of Korea. The product concerned also includes DRAMs presented in (non-customised) memory modules or (non-customised) memory boards, or in some other kind of aggregate form, provided the main purpose of which is to provide memory (7).

(10)

The product concerned is currently classifiable within CN codes 8542 21 11, 8542 21 13, 8542 21 15, 8542 21 17, ex 8542 21 01, ex 8542 21 05, ex 8548 90 10, ex 8473 30 10 and ex 8473 50 10.

(11)

The Panel report did not cover any claims affecting the findings set out in the definitive Regulation concerning the product under consideration and the like product.

(12)

This part sets out the reassessed findings based on the conclusions and recommendations of the Panel report concerning:

(13)

All other findings concerning subsidies remain as established in the original investigation and set out in the provisional Regulation and the definitive Regulation.

(14)

The Panel found that the determination that the creditor banks were directed to participate in the May 2001 restructuring programme and that their purchase of KRW 1 trillion of convertible bonds (CBs) as part of the programme therefore constituted a financial contribution by the GOK was not consistent with Article 1.1(a) of the SCM Agreement (8). More specifically, the Panel found that ‘while the May 2001 restructuring programme is a close case, we consider that the evidence before the investigating authority was insufficient to reasonably conclude that the government entrusted or directed the private banks to purchase the Hynix CBs.’ (9).

(15)

In accordance with the Panel's assessment, the funds provided to Hynix in the context of the May 2001 restructuring programme shall not be countervailed (10).

(16)

With respect to the syndicated loan of KRW 800 billion, the Panel found that the EC had properly established that the GOK directed the Korea Exchange Bank (KEB) to participate in the loan and that this constituted a financial contribution via entrustment or direction of KEB in accordance with Article 1.1(a) of the SCM Agreement. The Panel also found that the Korea Development Bank (KDB) provided a direct financial contribution through its participation in the syndicated loan, and, because the KDB was a public body, the EC was not required to make a finding of entrustment or direction (11). However, the Panel found that the EC failed to establish the existence of a benefit from the financial contribution provided by the KDB and the KEB within the meaning of Article 1(b) of the SCM Agreement (12). Specifically, the Panel stated that the EC had not examined whether there were comparable commercial loans that could actually have been obtained by Hynix on the market and stated that the EC had ignored the other participants in the syndicated loans as possible benchmarks (13).

(17)

In the view of the Commission, the participation by the other banks in the syndicated loan cannot serve as a proper benchmark for determining the existence of a benefit. As the Panel itself noted, in determining the existence of a commercial benchmark, authorities must examine whether such loans were so influenced by the alleged government intervention that they do not represent loans that could actually be obtained on the market (14). The Commission has conducted a thorough re-examination of the extensive record in this case and has employed the type of analysis discussed by the Panel. On this basis, the Commission has determined that the loans from the other banks participating in the syndicated loan could not serve as a proper benchmark.

(18)

First, bearing in mind the financial condition of Hynix at the time that the syndicated loan was arranged, as noted in recital 19 of the definitive Regulation, at the time the loan was agreed to, at the end of 2000, Hynix was in severe financial distress. It had over USD 9,46 billion in liabilities at the end of 2000, which was over twice its net worth and over four times the market capitalisation of the company. It suffered a loss in 2000, a year in which other large DRAM producers registered profits and which was generally considered to have been a very good year in the DRAM industry. The majority of Hynix liabilities were coming due in 2001 and the company was facing intense liquidity problems. Many independent analysts and brokerage houses were saying at that time that the company would not be able to generate sufficient cash flow or asset disposals to pay off its maturing liabilities. Under such circumstances, profit maximising banks operating under normal commercial conditions would not voluntarily participate in loans with such enormous risk of default (15).

(19)

Second, there is ample evidence on the record that other participants in the syndicated loan were pressured to participate in the overall restructuring of Hynix's debt. For example, the record contains evidence that one of the participants, Kookmin Bank, was required to lend directly to ‘troubled corporate borrowers’. Specifically, in a statement to the Securities and Exchange Commission of the United States, Kookmin stated:

‘The Korean government has promoted, and, as a matter of policy, may continue to promote lending to certain types of borrowers. It generally has done this by requesting banks to participate in remedial programs for troubled borrowers (such as Hynix) and by identifying sectors of the economy it wishes to promote and making low interest loans available to banks and financial institutions who lend to borrowers in these sectors . The government has in this manner promoted low-income mortgage lending and lending to technology companies. We expect that all loans made pursuant to government policies will be reviewed in accordance with our credit review policies. However, government policies may influence us to lend to certain sectors or in a manner in which we otherwise would not in the absence of government policy .’ (16).

It is clear that Kookmin refers to Hynix in mentioning ‘troubled borrowers’, because the bank's 2001 annual report lists Hynix as the single largest financially troubled borrower (17). Further, Kookmin, as a GOK-invested enterprise, was subject to audit and inspection by the GOK Board of Audit and Inspection and as such was under pressure to comply with GOK policy.

(20)

Other participants in the syndicated loan were also under pressure from the GOK to assist Hynix. In this respect, recital 55 of the definitive Regulation provides extensive information regarding the extent to which the GOK put pressure on Korea First Bank (KFB) to provide ongoing financing to Hynix. In a high-level economic ministerial meeting in November 2000 the GOK took steps to ‘alleviate Hynix's cash crunch’ and waived statutory credit limits for KFB, KEB and KDB. The Panel did not examine the evidence in relation to KFB, because the Commission concluded its involvement to be negligible. However, given that the Panel found that the same GOK tactics employed in relation to KEB amounted to direction and that the GOK owns 49 % of KFB, it is reasonable to conclude that the GOK was in a position to and did exert influence over KFB in relation to its decision to participate in the syndicated loan. KDB is a public body and KEB was found by the Panel to have been directed by the GOK. The participation of these banks in the syndicated loan is thus not a reliable commercial benchmark. Other banks that were part of the syndicated loan were also subject to similar pressure from the GOK. The GOK put very direct pressure on both Shinhan and KorAm banks regarding the extension of the export credit arrangements, which was made at the same time as the syndicated loan. These banks had concluded that they were already over-exposed to Hynix and reversed their earlier decision to increase the export credit ceiling available to Hynix. In response, an FSS official criticised their move as ‘unthinkable’ (18). The GOK held 18,3 % of Shinhan and was its largest shareholder, thus able to exert considerable influence over the bank. It is clear that Shinhan was not acting commercially given that it increased its loan loss reserves for Hynix even as it continued to assist the company and that it offered no legitimate credit risk analysis for assisting it. KorAM Bank again faced GOK threats when the bank refused to purchase Hynix CBs, because it decided that doing so was not financially sound. In response to the bank's decision not to purchase the bonds, an FSS official rebuked KorAm, using language uncharacteristic of a government apparently uninvolved in credit decisions. The official stated: ‘If KorAm does not honour the agreement, we will not forgive the bank’ (19). He further stated that the FSS would take stern measures against the bank, such as disapproving new financial instruments and subjecting the bank to a tighter audit (20). As a result of these threats, KorAm Bank capitulated and reversed its decision (21), while at the same time increasing its loan loss reserves for Hynix and offering no legitimate credit analysis for providing such financial assistance. Of the remaining banks which participated in the syndicated loan, Chohung Bank, Woori Bank, the NACF, Seoul Bank, Kwangju Bank, Peace Bank, Kyongnam Bank and the Industrial Bank of Korea are all GOK-owned and thus their credit decisions are tainted by GOK policy. The Housing and Commercial Bank, of which the GOK has a 20 % share, admitted government direction in its credit decisions in a sworn statement to the US SEC (22). Finally, although Pusan Bank had no GOK ownership, it was nevertheless ordered by the FSS in 1999 to fire or replace its executives and to increase its paid-in capital by USD 86,2 million before the end of March 2000 (23), to ‘slash their manpower, the number of branches and organisations, while disposing of ailing subsidiaries’ (24) and, in 2000, the FSS sanctioned its former and current officials for mismanagement (25). Given this degree of GOK control over Pusan Bank, the bank's decision to participate in the syndicated loan cannot be considered to be a safe commercial benchmark.

(21)

Third, it would clearly not be appropriate to use Citibank Seoul as a benchmark to determine the commercial reasonableness of participating in the syndicated loan, for the reasons set forth in recitals 130 to 133 of the definitive Regulation. First, Citibank Seoul failed to cooperate in this investigation, thus forcing the Commission to rely on facts available regarding the participation of Citibank Seoul. Second, the evidence on the record discussed in recital 133 of the definitive Regulation reflects an unusually close relationship between the GOK and Citibank that calls into question the impartiality of Citibank in lending to Hynix. Third, Citibank itself was a financial advisor to Hynix regarding its restructuring and thus had an inherent conflict with respect to participation in new lending to Hynix (26). It had not given financing to Hynix prior to becoming Hynix's financial advisor. Fourth, the Panel itself recognised that Citibank cannot serve as a basis for determining whether any of the loans to Hynix were made on the basis of commercial considerations (27).

(22)

Finally, it is important to note that Hynix defaulted on its interest payments on the syndicated loan almost as soon as the loan was given. As noted in recital 87 of the definitive Regulation, Hynix had not been servicing the syndicated loan that it received in January 2001.

(23)

Based on the foregoing, it may be concluded that none of the other seven participants in the syndicated loan can serve as an appropriate benchmark for measuring the commercial reasonableness or the amount of benefit to Hynix from the portions of the loan provided by KEB and KDB.

(24)

On this basis, the participation of KEB and KDB in the syndicated loan may still be considered to be a countervailable subsidy that provides a benefit to Hynix. However, as set out in recital 37 of the definitive Regulation, no subsidy amount has been countervailed under the syndicated loan in order to avoid double-counting, as the syndicated loan amount was subsequently rolled-over into the October 2001 measures. Thus, no amendment of the findings is necessary with regard to this measure.

(25)

The Panel found, at para. 7.215 of its report, that the calculation of the amount of the benefit conferred by each of the financial contributions was inconsistent with Articles 1.1(b) and 14 of the SCM Agreement. The GOK did not raise any claims challenging ‘as such’ provisions of the basic Regulation or the Guidelines (28) relating to the calculation of the amount of the benefit. Thus, the Panel report only addressed how the EC applied these provisions and calculated the amount of the benefit with respect to the measures at issue in this proceeding (the definitive Regulation).

(26)

In particular, the Panel considered that the EC acted inconsistently with Article 1.1(b) and 14 of the SCM Agreement by taking as a starting point that the amount of the benefit conferred by each of the financial contributions, which were generally in the form of loans, guarantees or equity, was to be calculated as if those financial contributions had been grants, irrespective of their terms and conditions. The Panel reasoned that the calculation of the amount of the benefit should be based on the benefit to the recipient, not the cost to the provider. It further reasoned that the provision in this case of funds to Hynix in the form of loans, guarantees and equity may have been of less benefit to Hynix than if the funds had been provided in the form of grants. In particular, the Panel reasoned that the funds would appear as debt in Hynix's balance sheet or would have given rise to the dilution of the ownership claims of existing shareholders in Hynix (29).

(27)

The Panel considered that, as a matter of WTO law, the correct starting point for the EC in this case must be the search for an alternative benchmark as a base for the calculation of the amount of benefit. The Panel particularly referred, at para. 7.213 of its report, to ‘a comparable commercial loan which the firm could actually obtain on the market’ (Article 14(b) SCM Agreement) and ‘a comparable commercial loan absent the government guarantee’ (Article 14(a) SCM Agreement). The Panel also referred, by way of example, to so-called ‘junk bonds’ or ‘vulture funds’ (30).

(28)

The Panel observed that there may be circumstances in which there is an absence of any comparable benchmark, and that the investigating authority was and is entitled to considerable leeway in adopting a reasonable approach to calculating the amount of the benefit to Hynix (31).

(29)

In the light of the Panel's findings, for each of the relevant financial contributions in the measure at issue, the Commission has re-examined all of the relevant facts and evidence on the record, taking into account all of the submissions of interested parties, specifically with respect to possible alternative benchmarks, including ‘junk bonds’ and ‘vulture funds’, and, for each relevant financial contribution, redetermined the amount of the benefit to Hynix.

(30)

The Commission has not re-examined or made any redeterminations with respect to any substantive or procedural matters confirmed by the Panel report or not found by the Panel to be inconsistent with the SCM Agreement, including, but not limited to: Hynix's financial situation or worth, conditions in the DRAM market, conditions in the financial markets and the use of facts available. The Commission has limited itself to the specific issue that is the subject of the Panel's findings: the availability of alternative benchmarks.

(31)

In making this redetermination, the Commission has based its calculation of subsidy on a ‘benefit to the recipient’ and not on a ‘cost to the government’ standard, as is required by Article 6 of the basic Regulation, consistent with the Panel's reasoning, with the determinations in the definitive Regulation and with the Commission's established practice (32). In this regard, the Commission notes that one of the key features of the SCM Agreement (Article 14) following the Uruguay Round was the confirmation that the ‘benefit to the recipient’ standard would normally prevail over the ‘cost to the government’ standard. The Commission further notes that the former will generally lead to a higher amount of subsidy than the latter, because it requires a comparison with commercial, rather than governmental, funding. Consequently, the Commission notes that the ‘cost to the government’ approach would not generally be expected to lead to a higher amount of subsidy than the ‘benefit to the recipient’ approach.

(32)

Thus, consistent with the Panel's reasoning, the Commission has examined the financial contributions ‘from the perspective of the recipient’. The Commission has done this by means of an objective examination of the information on the record of the case. The Commission has not enquired into what might have been the subjective state of mind of one or more Hynix executives at the time. Rather, it has enquired into what, objectively, the amount of the benefit would have been for a company in exactly Hynix's situation at the time. Calculating the amount of the benefit from the perspective of the recipient does not entail switching from an objective to a subjective basis of enquiry.

(33)

The Panel summarised the facts concerning the KEIC guarantee in the following terms:

(34)

The Panel found that the EC reached a reasonable and reasoned conclusion on the basis of the record before it that the KEIC Guarantee conferred a benefit to Hynix (34). In doing so, it recalled that the KEIC itself was hesitant to provide the guarantee on Hynix's debt and needed its own guarantee from the GOK (35). However, as regards the amount of the benefit, the Panel report indicates that the EC should, as a matter of WTO law, re-examine the question of the availability of alternative benchmarks for calculating the benefit to Hynix (36).

(35)

As the Panel observed, the question of benefit in the case of a government guarantee can be examined from more than one angle (37).

(36)

One reasonable method of enquiring into the amount of benefit from such a measure might be to compare the guarantee provided by the government with a comparable guarantee provided by the market (38).

(37)

A second reasonable method is indicated by Article 6(b) of the basic Regulation (which generally corresponds to Article 14(b) of the SCM Agreement), which refers to government loans; and Article 6(c) of the basic Regulation (which generally corresponds to Article 14(c) of the SCM Agreement), which refers to a loan guarantee by a government (39). In this case, the ‘loan guaranteed by the government’ is the revolving amount of debt outstanding at any time owed by the Hynix group, via its affiliated importers in various countries of export, to the banks, in respect of export transactions where payment has yet to be made, but funds have already been advanced by the banks to Hynix. The financial contribution under consideration related to the extension of the facility by USD 600 million. However, the facility itself had an overall ceiling of USD 1,4 billion.

(38)

The Commission has re-examined the extensive record in this case and considered whether there is any relevant benchmark against which the amount of the benefit could be calculated. First, the Commission considered whether Hynix had received new loans at the time that the USD 600 million increase had been granted. The record demonstrates, however, that no new commercially comparable lending had been provided. Second, the Commission looked for verified evidence regarding the cost and terms of a comparable commercial guarantee or a ‘comparable commercial loan absent the government guarantee’ that Hynix or another firm in a situation comparable to Hynix concluded at the relevant time. A comparable commercial loan would ideally be one of approximately the same size, would related to the same or a similar sector of economic activity, would involve a comparable type of activity, such as export credit insurance, and would have been concluded at approximately the same time. It could not involve any government guarantee, express or implied, and would need to involve a firm in a comparable financial situation to that of Hynix.

(39)

Having conducted this re-examination, the Commission has determined that there is no reliable information on the record regarding these matters. The Commission has not found any such benchmark as regards Hynix itself. The loans provided to Hynix were generally either provided by the GOK or by banks that were entrusted and directed by the GOK, and the rate on the syndicated loan would also be an inappropriate benchmark for the reasons stated in recitals 17 to 20 above. Lending from Citibank Seoul is also not an appropriate benchmark for the reasons stated in recital 21 above. Citibank's participation was insufficiently significant, was influenced by the participation of other lenders and the GOK and was distorted by both its interest, together with SSB, as Hynix's financial advisor and its long-term relationship with the GOK. In any event, Citibank did not cooperate in the investigation.

(40)

Given Hynix's financial condition, there are no Hynix-specific benchmarks that can be used and there is no data on the record regarding loans to companies in a similar financial condition to Hynix. Moreover, the available information indicates that Hynix did not obtain, and would not have been able to obtain, the relevant funds in the market (40). The KEIC export credit guarantee programme and the KDB debenture programme were in effect throughout 2001. During this period, Hynix was, by all measures, uncreditworthy. Hynix's already dismal financial condition worsened markedly in 2000, when the company suffered a net loss of over KRW 2 trillion. At year-end 2000, Hynix's total indebtedness was KRW 8 trillion, and Hynix's current ratio, which measures the degree to which a firm's current assets cover current liabilities, was 0,33 (41). Hynix's net profit margin fell from 2,08 % in 1999 to minus 24,03 % in 2000 (when the DRAM industry as a whole performed well) and minus 93,83 % in 2001 (42). Clearly, during 2000 and 2001, Hynix was unable to generate adequate cash flow to cover its existing debt, much less new debt.

(41)

Market analysts agreed that Hynix was not creditworthy during this period. Hynix's Standard & Poor's rating was B- in March 2001, CCC+ in August 2001 and ‘selective default’ in October 2001. In January 2001, UBS Warburg explained that Hynix's ‘debt refinancing agreement with KDB … merely puts off the evil day but does not solve (Hynix's) core problem, that it is not profitable and is not paying off debt at a sufficiently fast rate from internal cash flow or asset disposals’ (43). UBS Warburg subsequently stated that Hynix's lack of product diversity was a concern and that Hynix's excessive focus on commodity DRAMs would contribute to its ‘deteriorating fundamentals’ throughout the first half of 2001 (44). Similarly, in February 2001, Deutsche Bank posited the cash flow problem bluntly: ‘We believe the most critical concern for (Hynix) is its ongoing cash flow concerns. In our view, it is unlikely that the Korean Government would be willing to give up on the company yet ….We believe (Hynix's) financial crisis will continue to carry a heavy burden on its ability to recover, as the DRAM sector is perhaps the most capital-intensive sector within semiconductors, with a constant requirement to upgrade technology processes’ (45).

(42)

In February 2001, in an article in the Korea Herald (46), another market expert stated that: ‘Given the fact that credit risk is still pronounced and haunting investors, I don't think fluent capital will flow into corporate bonds rated below BBB+’. This coincides with the early period of the KEIC Guarantee and KDB Debenture Programme. The article goes on to point out the yield difference between investment grade bonds and junk bonds, concluding that ‘this indicates that investors continue to shun low-grade corporate bonds, expanding the spread between high-credited bonds and the so-called junk bonds to their widest margin’. A press report in June 2001 stated that ‘Investors have started gazing down the Asian credit curve, but they still lack the nerve to eye corporates with ratings below the bottom rung of investment grade at BBB-minus. Only companies with strong cash flows, ability to service debt and growth prospects are likely to make them look.’ (47). Another report stated that ‘Especially in the latter half of last year, it was nearly impossible for a company to issue junk bonds based on its own credit and the junk bonds market existed through primary collateralised bond obligations, collateral loan obligations and a government-initiated bond-refinancing plan.’ (48).

(43)

Thus, the Commission confirms its original finding that Hynix fell short of the investment grade required by investors for the junk bond market; and that, during the IP, the type of capital markets referred to by the Panel were inaccessible to companies with prospects as weak as those of Hynix. Accordingly, such information as is on the record does not permit the establishment of a reliable benchmark with respect to such markets.

(44)

The Commission would therefore be in a position to reasonably and objectively conclude that no alternative commercial benchmark would actually be available in this case. Accordingly, in these circumstances, the Commission finds that the amount of the benefit to Hynix could not exceed the full amount of the loan, plus compound interest and the costs of recovery or attempted recovery in the event of default. Because there is also no reliable information on the record regarding these latter additional costs, the Commission has confined itself to the full amount of the loan. The Commission believes that this conclusion is fully supported by its (unchanged) determinations regarding the financial situation and worth of Hynix, and the situation in the DRAMs market and the financial markets, and that it is consistent with the basic Regulation and the Guidelines.

(45)

However, mindful of the WTO Panel’s reasoning, the Commission has, exceptionally and for this reason only, also considered whether or not, even absent any relevant benchmarks, there is anything that might serve as the basis for a proxy benchmark, adjusted where necessary to take account of the particular circumstances of this case. Such a proxy could, for example, be based on a commercial interest rate, with the addition of a risk premium, reflecting the default rates of companies exhibiting comparable risk. Such a calculation would take into account the term of the loan, a benchmark interest rate that would be paid by a creditworthy company, the probability of default by an uncreditwothy company within a specified period of time and the probability of default by a creditworthy company within the same specified period of time. This approach is premised on the commonly understood notion that when a lender makes a loan to a company that is considered to be uncreditworthy, the lender faces higher probability that the borrower will default on repayment of the loan. As a consequence of this higher default probability, the lender will charge a higher rate of interest, and one that is in proportion to the increased probability of default. Essentially, this would entail taking an interest rate available in the market to creditworthy borrowers and adding a ‘risk premium’ to it in order to account for the increased risk associated with lending to a company that is not creditworthy.

(46)

For purposes of implementing the Panel's report in this case, the Commission has used the following formula to calculate an appropriate uncreditworthy benchmark interest rate:

ib= [(1-qn)(1+if)n/(1-pn)]1/n -1

Where:

(47)

For the purposes of this implementation, the Commission has used as the creditworthy (or commercial) rate for USD loans the average USD lending interest rate for 2001, which, according to the IMF International Financial Statistics Service, was 6,922 %. Although the credit received by Hynix by virtue of the guarantee was short term, the measure under consideration, i.e., the guarantee itself, was in fact a form of rolling credit. The Commission used data from Moody's Investor Service to obtain default rates for creditworthy and uncreditworthy companies. For the uncreditworthy rate, the Commission used the likelihood of default for companies with a credit rating of Caa to C during the term of the relevant measure (i.e., one year for the KEIC guarantee). Similarly, the Commission used the default rate for a creditworthy company during the term of the measure in question, calculated based on a cumulation of data for investment grade companies. The rate of default of an uncreditworthy company in 2001 within one year was, according to Moody's, 34,34 %. The rate of default of a creditworthy company in 2001 within one year was 0,13 %. This led to total interest rate of 62,70 %. In contrast, the interest rate payable by Hynix on the USD loans in question was 8,33 %, i.e., a total interest differential of 54,37 %.

(48)

The Commission recalls that, normally, having determined that the benefit is equivalent to the full amount of the financial contribution (in this case, a loan or loan guarantee), the Commission treats it as a ‘non-recurring’ subsidy and allocates it over a period of time corresponding to the normal depreciation of fixed assets in the industry concerned. The Commission uses a straight-line allocation and does not increase the total amount of benefit by using the ‘time value of money’ method; it merely adds interest to each annual allocated portion. In this case, the allocation period was five years and no additional costs were added to the full amount of the loan as explained in recital 44 above, which means that 20 % of the financial contribution was allocated to each year.

(49)

Had the Commission used an interest rate comparison as described in recitals 45 to 47 above, the benefit (i.e., the difference between the benchmark and the actual interest rate) would have been considered as a recurring subsidy. The amount of subsidy allocated to each year would have been the interest rate differential multiplied by the fact amount of the loan. Therefore, depending on the size of the interest rate differential, the terms and conditions of the loan, and the length of the depreciation period, the annual benefit obtained by making an interest rate comparison may be bigger or smaller than that which results from the application of the Commission’s original approach.

(50)

In this case, the Commission's original approach would result in an allocated annual benefit of 20 % of the financial contribution. Depending on the exact terms of the loan, i.e., whether the beneficiary repays interest and capital or interest only each year, this amount would be reached with an interest rate differential of no less than 20 % and no more than about 35 % per year. The actual interest differential in this case is 54,37 %. This means that a strict application of this method would result in a much higher rate of countervailable subsidy allocated on turnover, in fact 12,8 % instead of the 5,1 % originally calculated by the Commission.

(51)

Consequently, the Commission considers that its original approach was and is reasonable in this case. Indeed, it is only in extreme cases that the Commission will determine the amount of the benefit to be equivalent to the full amount of a loan (i.e., when there is no realistic possibility of repayment) and in such circumstances it is usual that the benchmark interest rate will be extremely high. Therefore, the interest rate differential will often give a higher amount of benefit than a simple allocation of the face amount of the loan over time. In these circumstances, the Commission has used the more conservative of the two options in order to be sure that it has not exaggerated the amount of the subsidy, and will treat the face amount of the loan as the basis for a cap on the maximum amount of the benefit.

(52)

On the above basis, the amount of benefit conferred by the KEIC guarantee ad valorem of export turnover was 5,1 %.

(53)

The Panel summarised the facts concerning the KDB debenture programme in the following terms:

(54)

The Panel found that the EC reached a reasonable and reasoned conclusion on the basis of the record before it that the KDB debenture programme conferred a benefit on Hynix. In doing so, it recalled that the KDB debenture programme was itself set up and effectively guaranteed by the GOK (54). However, as regards the amount of the benefit, the Panel report indicates that the Commission should, as a matter of WTO law, re-examine the question of the availability of alternative benchmarks.

(55)

As the Panel observed, the question of benefit, at least in the case of a government guarantee, can be examined from more than one angle (55), and in applying the guidelines contained in Article 14 of the SCM Agreement, an investigating authority is entitled to considerable leeway (56). The KDB debenture programme, for example, exhibits some of the characteristics of a guarantee, insofar as the entire measure was set up, sponsored and effectively guaranteed by the GOK or a GOK public body. Thus, one reasonable method of enquiring into the amount of benefit of such a measure might be to compare the guarantee provided by the government with a comparable guarantee provided by the market. The KDB debenture programme might also be considered in the light of Article 6(b) of the basic Regulation (which generally corresponds to Article 14(b) of the SCM Agreement), which refers to government loans, and Article 6(c) of the basic Regulation (which generally corresponds to Article 14(c) of the SCM Agreement), which refers to a loan guarantee by a government.

(56)

The Commission has re-examined the record to consider whether there is any relevant benchmark against which the amount of the benefit could be calculated. It has looked for verified evidence about a comparable government guarantee or ‘a comparable commercial loan’ or ‘a comparable commercial loan absent the government guarantee’ that Hynix or another firm in a situation comparable to Hynix concluded at the relevant time. A comparable commercial loan would ideally be one of approximately the same size, would relate to the same or a similar sector of economic activity, would involve a comparable type of activity, and would have been concluded at approximately the same time. It could not involve any government guarantee, express or implied, and would need to involve a firm in a comparable financial situation to that of Hynix.

(57)

Having conducted this re-examination, the Commission has determined that a commercial benchmark meeting these criteria was not available in this case. Hynix itself did not sell any corporate bonds into the commercial market at the time it participated in the KDB debenture programme. In fact, one of the threshold criteria for participation in the KDB debenture programme was that the company in question could not obtain loans or sell bonds on the normal commercial market. Moreover, as noted in recital 53 of the definitive Regulation, Hynix provided no evidence on the record that it tried to access foreign debt markets during this period. In addition, the fact that the international ratings agency Standard and Poor's downgraded Hynix's debt to a speculative rating of ‘B-’ makes it unlikely that Hynix would have been able to access foreign bond or loan markets, even if it had tried to do so. Even the Korean government itself noted in January 2001 that Hynix was ‘in a state of technical insolvency.’ (57).

(58)

Moreover, other lending to Hynix during the period that the KDB rolled over Hynix’s maturing bonds cannot serve as a basis for measuring whether a benefit to Hynix exists. For the reasons noted in recitals 17 to 22 above, the loans provided as part of the syndicated loan cannot serve as an appropriate basis of comparison. The lending made pursuant to the syndicated loan was so distorted by the government’s intervention in that loan that it cannot serve as a reasonable commercial benchmark by which to measure whether a benefit has occurred or the amount of that benefit. In this regard, the Commission recalls that the syndicated loan itself was made possible by the GOK waiving the lending limits for three key participants in that loan. Without the participation of these three banks in the syndicate, it is unclear whether the loan would have been able to go forward at all, given that it would have been significantly undersubscribed (58).

(59)

Further, the convertible bonds purchased by private bodies as part of the May 2001 restructuring programme cannot serve as a benchmark because, although the evidence on record did not rise to the level sufficient for a finding of entrustment or direction, it indicates a level of government involvement that makes the participation by private creditors an inappropriate measure of the commercial reasonableness of the CB purchases: the creditor banks themselves acknowledged in the 7 May 2001 creditor financial institution resolution that Hynix was unable to service its debts in May 2001 (59) and the GOK was clearly heavily involved in overseeing the completion of the May 2001 CB offering, going as far as having a high level FSS official attend the March 2001 meeting of the creditor council (60). Further, a comparable commercial loan which could serve as a benchmark is one which is concluded at approximately the same period of time; the CB offering took place five months after the KDB debenture programme, when both Hynix's financial situation and the state of the DRAMs market were markedly different.

(60)

As already indicated above (61), no other comparable commercial benchmark was actually available in the market. On the contrary, the available information indicates Hynix would not have obtained the relevant funds in the market.

(61)

The Commission would therefore be in a position to reasonably and objectively conclude that no alternative commercial benchmark would actually be available in this case. Accordingly, under these circumstances, the Commission finds that the amount of the benefit to Hynix could not exceed the full amount of the funds obtained by Hynix as a result of the operation of the KDB debenture programme, plus compound interest and the costs of recovery or attempted recovery in the event of default. Because there is also no reliable information on the record regarding these latter additional costs, the Commission has confined itself to the full amount of the funds. The Commission believes that this conclusion is fully supported by its (unchanged) determinations regarding the financial situation and worth of Hynix, and the situation in the DRAMs market and the financial markets, and that it is consistent with the basic Regulation and the Guidelines.

(62)

However, mindful of the WTO Panel's reasoning, the Commission has also considered whether or not, even without the availability of any appropriate commercial benchmarks, there is anything that might serve as the basis for a proxy benchmark, adjusted where necessary to take account of the particular circumstances of this case. Such a proxy could, for example, be calculated along the lines already indicated above in recitals 45 and 46 with respect to the KEIC Guarantee.

(63)

For the purposes of this calculation, and as suggested by GOK, the Commission has used as the creditworthy rate for KRW loans the IMF International Financial Statistics rate on short-term KRW denominated-lending, which was 7,7 %. The default rates for both creditworthy and uncreditworthy companies for the term of the measure in question were taken from Moody's Investor Services (0,13 % and 34,37 % respectively). This led to a total interest rate of 63,89 %. In contrast, the average interest rate payable by Hynix on the KDB debentures in question was 12,7 %, i.e., a total interest rate differential of 51,19 %.

(64)

Viewing the financial contribution as a recurring subsidy, rather than as a non-recurring subsidy to be allocated over time, and applying this model to the calculation of the amount of the benefit in the financial contribution under consideration, would lead to an amount of benefit of KRW 626,9 billion, resulting in a countervailable subsidy allocated on turnover of 11,8 %, instead of the 4,9 % originally calculated by the Commission. Consequently, the Commission considers that its original approach was and is reasonable in this case, as the interest rate differential would give a much higher amount of benefit than a simple allocation of the face amount of the loan over time. In these circumstances, the Commission has used the more conservative of the two options in order to be sure that it has not exaggerated the amount of the subsidy and will treat the face amount of the loan as the basis for a cap on the maximum amount of the benefit.

(65)

Taking the amount of the funds obtained by Hynix as a result of the operation of the KDB debenture programme as a cap on the maximum amount of the benefit, the Commission finds that the amount of benefit to Hynix in the financial contribution at issue ad valorem of turnover was 4,9 %.

(66)

The Panel summarised the facts concerning the October 2001 restructuring programme in the following terms:

(67)

The Panel found that the EC reached a reasonable and reasoned conclusion on the basis of the record before it that the October 2001 restructuring programme conferred a benefit on Hynix (64). However, as regards the amount of the benefit, the Panel report indicates that the EC should, as a matter of WTO law, re-examine the question of the availability of alternative benchmarks.

(68)

As the Panel indicated, in approaching the question of benefit in the case of the government provision of equity capital and in applying the guidelines contained in Article 14 of the SCM Agreement, an investigating authority is entitled to considerable leeway (65). The Commission considers that the debt-to-equity swap in the October 2001 restructuring programme may be considered in the light of Article 6(1) of the basic Regulation (which generally corresponds to Article 14(a) of the SCM Agreement), which refers to government provision of equity capital inconsistent with the usual investment practice, including for the provision of risk capital, or private investors in the territory of the relevant Member. The Commission notes that, unlike subparagraphs (b) and (c), subparagraph (a) contains no final sentence providing further guidance on how the amount of the benefit is to be calculated.

(69)

The Commission has re-examined the record to consider whether there is any relevant benchmark against which the amount of the benefit could be calculated. It has looked for verified evidence about a comparable private provision of equity capital that Hynix or another firm in a situation comparable to Hynix concluded at the relevant time. A comparable provision of equity ideally would be one of approximately the same size, would relate to the same or a similar sector of economic activity, would involve a comparable type of activity and would have been concluded at approximately the same time. It could not involve any government guarantee, express or implied, and would need to involve a firm in a comparable financial situation to that of Hynix.

(70)

Having conducted this re-examination, the Commission has determined that there is no verified information on the record regarding these matters. The Commission has not found any such benchmark as regards Hynix itself. The equity provided to Hynix was generally either provided by the GOK or by banks that were entrusted or directed by the GOK.

(71)

Moreover, no other comparable commercial benchmark was actually available in the market. On the contrary, the available information indicated Hynix would not have obtained the relevant funds in the market. As noted in recital 103 of the definitive Regulation, in October 2001 Hynix's credit rating from Standard & Poor’s was ‘SD’ (selective default). Hynix went into default of interest payments on the syndicated loan in May 2001 and several creditor banks raised their loan-loss provisions as regards Hynix already in May 2001 and again in October 2001, before the measures under consideration were put in place. The six banks concerned by the October 2001 measures rated Hynix internally from ‘precautionary’ to ‘doubtful’ and the Option 2 banks involved in the debt-to-equity swap were given very little choice: they either took the shares or they effectively wrote off Hynix’s debts completely. The GOK suggestion that the difference between the swap price paid and the market price should serve as the amount of benefit ignores the realities of Hynix’s abysmal financial situation; no-one would invest in a company with an SD rating. The conclusions set out in recitals 101 to 112 of the definitive Regulation supporting the finding that no comparable equity financing that might serve as a benchmark was available to Hynix on the commercial market are hereby confirmed.

(72)

The Commission would therefore be in a position to reasonably and objectively conclude that no alternative commercial benchmark would actually be available in this case. Accordingly, under these circumstances, the Commission finds that the amount of the benefit to Hynix could not exceed the full amount of the funds obtained by Hynix as a result of the operation of the debt to equity swap in the October 2001 restructuring programme. The Commission believes that this conclusion is fully supported by its (unchanged) determinations regarding the financial situation and worth of Hynix, and the situation in the DRAMs market and the financial markets, and that it is consistent with the basic Regulation and the Guidelines.

(73)

However, mindful of the WTO Panel's reasoning, the Commission has also considered whether or not, even without any relevant benchmarks, there is any other method by which the amount of the benefit might be adjusted to take into account the particular circumstances of this case. The Commission considers that the uncreditworthy analysis conducted with respect to debt cannot provide an appropriate basis for comparison, precisely because this particular financial contribution is concerned with equity, which is a fundamentally different type of measure from debt.

(74)

The Commission has taken into consideration the possibility that, in the particular circumstances of the present reassessment, the cost to the recipient of receiving the financial contribution should be taken into account. The Commission notes that, in order to examine the cost, if any, of the swap from the point of view of the recipient, the obligations imposed on a company when issuing new equity would have to be considered. The Commission notes in this respect that injections of ordinary share capital do not generally impose particular obligations on a company, as there is neither a fixed rate of return which the company must achieve nor, in fact, do any payments need to be made. However, the company has at least a theoretical obligation to distribute its profits or part thereof to its shareholders; the return on equity ratio (ROE) could provide some indication of the level of return the company is expected to provide to its shareholders and could be used in the calculation of the amount of the benefit to Hynix. However, the ROE of Hynix for 2001 was minus 96,75 %, as calculated by reference to its financial statements for the year. Therefore, on the basis of an objective evaluation of Hynix's financial standing at the time of the debt-to-equity swap, it is unreasonable to expect that Hynix would be in a position to distribute any sort of return to its shareholders for the relevant year. Hence, using the ROE as a benchmark, it would still be concluded that the entire amount of the debt-to-equity swap is a countervailable subsidy.

(75)

The Commission has also examined whether there are costs to Hynix associated with having to surrender ownership shares as a result of the swap or the cost to Hynix from potential dilution of the share price resulting from the issuance of new shares (66). Based on its review of the record evidence, the Commission does not believe that the net benefit to Hynix of the full debt-to-equity swap is diminished as a result of the issuance of new shares. In this regard, the Commission believes it is essential to recognise the impact that the debt-to-equity swap had on Hynix’s financial situation.

(76)

As a threshold matter, it is key to recognise that this was not a straight equity infusion. The fundamental benefit to Hynix resulted from the massive amount of debt that was extinguished as a result of the swap, namely KRW 2,483 billion. This effectively absolved Hynix from having to repay any of this principal amount and from having to pay interest. Instead, this massive amount of debt was replaced with equity shares issued to the Option 1 creditor banks. However, there was no real quantifiable cost to Hynix of issuing the new shares. While the existing share values were diluted as a result of the issuance, this had absolutely no balance sheet impact on Hynix. It did not result in any cash outlays (apart from the expense associated with the issuance of the new shares) and did not obligate Hynix to any sort of cash payments in the future, as would have been the case with debt instruments.

(77)

Consequently, in light of all these considerations and since Hynix did not report any costs associated with the issuance of equity, the appropriate approach to measuring the benefit from the debt-to-equity swap is to take the full amount of the swap. This results in an overall benefit to Hynix of KRW 358,2 billion and a countervailable subsidy allocated on turnover of 9,6 %.

(78)

The Commission considers that the debt roll-over and new loan in the October 2001 restructuring programme may be considered in the light of Article 6(b) of the basic Regulation (which generally corresponds to Article 14(b) of the SCM Agreement), which refers to the amount which the firm would pay on a comparable commercial loan that the firm could actually obtain on the market.

(79)

The Commission has re-examined the record to consider whether there is any relevant benchmark against which the amount of the benefit could be calculated. It has looked for verified evidence about a comparable commercial loan that Hynix or another firm in a situation comparable to Hynix concluded at the relevant time. A comparable commercial loan ideally would be one of approximately the same size, would relate to the same or a similar sector of economic activity, would involve a comparable type of activity, and would have been concluded at approximately the same time. It could not involve any government guarantee, express or implied, and would need to involve a firm in a comparable financial situation to that of Hynix.

(80)

The bonds purchased as part of the May 2001 restructuring programme cannot serve as a basis for determining the commercial reasonableness of the new loan and the debt roll-over, as Hynix's financial situation deteriorated substantially between May and October 2001. Therefore, the two situations cannot be considered to be comparable.

(81)

Having conducted this re-examination, the Commission has determined that there is no verified information on the record regarding these matters. The Commission has not found any such benchmark as regards Hynix itself. The loans provided to Hynix were generally either provided by the GOK or by banks that were entrusted or directed by the GOK.

(82)

As already indicated above (67), no other comparable commercial benchmark was actually available in the market. On the contrary, the available information indicates Hynix could not have obtained the relevant funds in the market.

(83)

The Commission would therefore be in a position to reasonably and objectively conclude that no alternative commercial benchmark would actually be available in this case. Accordingly, under these circumstances, the Commission finds that the amount of the benefit to Hynix could not exceed the full amount of the funds obtained by Hynix as a result of the operation of the debt roll-over and new loan in the October 2001 restructuring programme. The Commission believes that this conclusion is fully supported by its (unchanged) determinations regarding the financial situation and worth of Hynix, and the situation in the DRAMs market and the financial markets, and that it is consistent with the basic Regulation and the Guidelines.

(84)

However, mindful of the WTO Panel's reasoning, the Commission has also considered whether or not, even in the absence of any relevant benchmarks, there is any other method by which the amount of the benefit might be adjusted to take into account the particular circumstances of this case. In this respect, the same methodology as the one outlined in recitals 45 and 46 was used to calculate the amount of the benefit received by Hynix from these two measures.

(85)

For the purposes of this calculation, the Commission has used as the creditworthy rate for KRW loans the average Bank of Korea interest rate on investment grade corporate bonds in 2001, which was 11,38 %. The default rates for both creditworthy and uncreditworthy companies for the term of the measures in question were taken from Moody's Investor Services. The rate of default of an uncreditworthy company in 2001 within three years (the term of the new loan and of the roll-over) was 57,97 %, whereas the rate of default of a creditworthy company within three years in 2001 was 0,94 %. This led to a total interest rate of 48,22 %. In contrast, the interest rate payable by Hynix on the new loan was 7 % and on the roll-over 6,5 %, i.e., an interest rate differential of 41,2 % and 41,7 % respectively.

(86)

Viewing the financial contributions as recurring subsidies, rather than as non-recurring subsidies to be allocated over time, and applying this model to the calculation of the amount of the benefit in the financial contributions under consideration, would lead to an amount of benefit of KRW 271,5 billion for the new loan and KRW 761,4 billion for the roll-over. This would result in countervailable subsidies allocated on turnover of 6,8 % for the new loan and 19,1 % for the roll-over, compared to 3,5 % and 9,8 % respectively, as originally calculated by the Commission. Consequently, the Commission considers that its original approach was and is reasonable in this case, as the interest rate differential would give a much higher amount of benefit than a simple allocation of the face amount of the loans over time. In these circumstances, the Commission has used the more conservative of the two options in order to be sure that it has not exaggerated the amount of the subsidy and will treat the face amount of the loans as the basis for a cap on the maximum amount of the benefit. On the basis of these considerations, the Commission had determined that the amount of the benefit to Hynix ad valorem of turnover is, for both measures, 13,3 %.

(87)

It is noted that, in relation to the debt roll-over, contrary to recital 166 of the definitive Regulation, the amount of CBs purchased by Option 1 banks has not been deducted, as the May 2001 measures are no longer countervailed.

(88)

The subsidy amount thus calculated for the debt-to-equity swap, the debt roll-over and the new loan was considered to confer a large non-recurring benefit and was spread across a five-year period corresponding to the normal depreciation period of assets in the DRAMs industry. The resulting amount was then allocated on Hynix’s turnover during the IP and interest based on the average commercial rate applicable in Korea during the IP was added. The revised amount expressed ad valorem is 22,9 %.

(89)

In summary, the calculation of the total amount of the subsidy has thus been revised on the basis of the Panel's recommendations and is now as follows:

(90)

This part sets out the reassessed findings based on the recommendation of the Panel report concerning the evaluation of the factor ‘wages’ as a relevant factor affecting the state of the Community industry within the meaning of Article 8(5) of the basic Regulation (which generally corresponds to Article 15.4 of the SCM Agreement).

(91)

All other findings concerning injury remain as established in the original investigation and set out in the provisional Regulation and the definitive Regulation.

(92)

In the original investigation the EC had considered and evaluated the wages factor, but found wages to be irrelevant to its injury determination. However, the Panel found that:

‘without a written record of the analytical process undertaken by the investigating authority to evaluate “wages”, we cannot examine whether an adequate and reasoned explanation has been provided of how the facts support the injury determination made by the EC and, consequently, whether that determination was consistent with Article 15.4. Without that written record, we would be forced to embark on post hoc speculation about the thought process by which the investigating authority arrived at its ultimate conclusions. We consider that the standard of review in Article 11 of the DSU prevents us from embarking upon this exercise.’ (68).

Consequently, the Panel concluded that the EC acted inconsistently with Article 15.4 of the SCM Agreement in not evaluating the ‘wages’ factor when examining the impact of the subsidised imports on the state of the domestic industry (69).

(93)

In the light of the Panel's finding, the Commission, having again reviewed the information on the record concerning Community wages, both in isolation and together with the other injury factors set out in Article 8(5) of the basic Regulation, provides herein a written record of its evaluation of the wages factor.

(94)

The Commission examined the average cost of wages per employee for the Community industry during the IP, and in the three prior years. The indexed results are shown in the following table (70).

(95)

The table shows that the average cost of the Community industry for wages and salaries during the period under consideration varied, decreasing and increasing over time. In 2000, it increased by 15 %. In 2001, compared to 1998, it had increased by 9 %. Overall, the increase was lower than the increase of the total hourly labour costs in the euro area during this period (71). Thus, there were no unusual developments in wages during the relevant period, and any developments only had very limited effects on the financial situation of the domestic industry as a whole (72).

(96)

Furthermore, the Commission recalls that DRAMs are capital intensive products and, thus, labour may only have a limited contribution to their cost of production (see also recital 41 above). Consequently, this relatively small increase in wages had hardly any effect on the overall financial situation of the Community industry.

(97)

Given its negligible impact, consideration of the wages factor together with the other factors referred to in Article 8(5) of the basic Regulation is not such as to alter the conclusion that the Community industry suffered material injury (73).

(98)

Based on the foregoing, the conclusion set out in recital 188 of the definitive Regulation that the Community industry suffered material injury within the meaning of Article 8 of the basic Regulation is hereby confirmed.

(99)

This part sets out the reassessed findings based on the recommendations of the Panel report concerning the non-attribution to the subsidised imports of any injury caused by other factors (in particular, decline in demand, overcapacity and other, non-subsidised imports) within the meaning of Article 8(7) of the basic Regulation. The Panel's determination in paragraph 7.437 of its report does not preclude a conclusion of a causal relationship between subsidised imports and the material injury suffered by the Community industry, but added that a satisfactory explanation of the nature and extent of the injurious effects of those other factors, preferably with quantitative analysis, as distinguished from the effects of the subsidised imports, is required.

(100)

Article 8(7) of the basic Regulation (which generally corresponds to Article 15.5 of the SCM Agreement) requires examination of any known factors other than subsidised imports which at the same time are injuring the Community industry, and not to attribute injury caused by these other factors to subsidised imports. One of the factors referred to is ‘contraction in demand’.

(101)

The Panel found in paragraph 8.1(e) of its report that the EC had failed to make sure that the injury caused by ‘decline in demand’ was not attributed to the subsidised imports. This conclusion was reached in the light of the Panel’s findings in paragraphs 7.411 to 7.414 of its report, which treat a ‘decline in demand’ as synonymous with a ‘general economic downturn’.

(102)

The Commission recalls that, all other things being equal, demand (i.e., the willingness and ability to obtain an amount of a particular good) may generally be considered to be equivalent to consumption (i.e., the actual physical process of using a good irrespective of the level of trade). The above table sets out EC consumption during the relevant period in terms of Mbits. The use of this unit of measurement was not disputed before the Panel. It appears from the above table that because the actual consumption in the Community for DRAMs increased constantly and significantly during the period under consideration, it is incorrect to allege that the level of demand for DRAMs in the Community declined during the same period. Thus, all of the relevant argumentation concerning the impact of the ‘decline in demand’ on the injurious situation of the Community industry (in particular its prices) raised by the GOK and Hynix is irrelevant, because, as shown above, the level of demand increased. With respect to the claim by the GOK and Hynix that the alleged decline in demand affected the Community industry's prices, this is unfounded in light of the fact that there was no decline in the level of demand. On the contrary, the subsidised imports had a significant impact on the prices of the Community industry and, thus, on its injurious condition (see recitals 191 and 192 of the definitive Regulation).

(103)

Hynix and the GOK noted that worldwide demand for DRAMs grew in 2001 by only 59 % compared to the average year on year growth rate of 75 %, apparently contending that in 2000 to 2001 there was a huge slowdown in the growth rate of demand (see recital 193 of the definitive Regulation). The Commission does not agree. Hynix and the GOK correctly point out that the DRAMs market is cyclical (see recital 193 of the definitive Regulation). However, in cyclical markets, demand growth rates vary sharply from one year to another, and it would be unreasonable to expect the growth rate of demand in a particular year to reflect the average growth rate of demand for an entire business cycle. A reasonable way to assess whether demand growth slowed significantly in 2000 to 2001, taking into account the cyclical nature of the market, would be to compare the growth rate of demand in 2000 to 2001, which was a cyclical trough, with the average growth of demand during troughs in previous cycles. Neither Hynix nor the GOK made any arguments nor submitted any data in support of such a comparison. Accordingly, record evidence does not support the conclusion that in 2000 to 2001 the growth rate of demand slowed down significantly, considering the cyclical nature of the DRAMs market.

(104)

It was also considered whether the growth rate of demand in the Community slowed down in 2000 to 2001 as compared to the actual growth rates observed in previous years. The demand for DRAMs in the Community, measured in terms of EC consumption, increased by 75 % in 1998 to 1999, 58 % in 1999 to 2000 and 51 % in 2000 to 2001. Thus, while there was a constant increase in demand during the period under consideration, this increase followed a declining trend. Nevertheless, demand growth rate remained significant in the IP (51 %).

(105)

The declining trend in demand growth was more pronounced between 1999 and 2000 (where demand growth declined by 17 percentage points) than between 2000 and the IP (where demand growth declined by 8 percentage points). This evolution of demand growth has to be assessed by comparison with the evolution of the sales prices of the Community industry, the primary cause of injury to the Community industry during the period under consideration (see recital 188 of the definitive Regulation). In particular, in 1999 to 2000, demand growth declined by a sizable 23 %, but the average sales price of the Community industry dropped by only 5 %. Conversely, between 2000 and the IP demand growth declined moderately by 12 %, but the average sales price of the Community industry plummeted by 77 % (see recital 128 and the relevant table of the provisional Regulation). If slowing demand growth had been a cause of injury, one would have expected the sales prices of the Community industry to decline more sharply in the period when demand growth declined more steeply. Record evidence shows no such relationship between trends in demand growth and trends in the sales prices of the Community industry. It is therefore concluded that there is no apparent link between demand growth and the sales prices of the Community industry which could indicate that slowing demand growth contributed to the injury suffered by the Community industry.

(106)

This approach is consistent with the findings of recent WTO panels reviewing determinations in respect of non-attribution. In particular, WTO panels have upheld non-attribution findings where record evidence showed that the factor concerned could not have caused injury because trends in such factor were independent from trends in key indicators of the domestic industry. Specifically, in US-Steel, the Panel noted that the facts appeared to support the investigating authority’s finding, in respect of hot-rolled bar, that cost increases had not caused injury because trends in operating margin (a key measure of profitability) were unrelated to trends in the cost of goods sold (COGS) (74). Similarly, while the Panel concluded that the investigating authority’s finding, in respect of stainless steel bar, concerning demand as a factor of injury, did not satisfy the non-attribution requirement, the Panel noted that the investigating authority could have met this requirement, through a ‘reasoned and adequate explanation’, simply by showing that trends in demand had no connection with trends in a key variable attesting to the condition of the domestic industry (75).

(107)

Accordingly, the Commission is of the view that its redetermination satisfies the non-attribution requirement under Article 8(7) of the basic Regulation (which generally corresponds to Article 15.5 of the SCM Agreement) in respect of demand growth by showing that, during the period under consideration, trends in demand growth were independent from trends in the sales prices of the Community industry. The Commission is also of the view that its redetermination satisfies the non-attribution requirement in respect of declines in the level of demand by showing that during the period under consideration no such demand declines took place.

(108)

The Panel noted in paragraph 7.421 of its report that the question of separating and distinguishing the effects of overcapacity in driving prices down should be addressed, irrespective of who is to blame for such overcapacity.

(109)

The Commission notes that ‘overcapacity’ is not expressly referred to in Article 8(7) of the basic Regulation (which generally corresponds to Article 15.5 of the SCM Agreement). The concept is discussed in the provisional and definitive Regulations insofar as it is invoked by certain of the interested parties. It is a relative concept, insofar as it depends on how much capacity is available and how much capacity is being used – that is, on the level of demand. ‘Overcapacity’ may thus arise either because capacity is increased, or demand falls.

(110)

‘Overcapacity’ as a possible factor of injury can be examined in the context of the world market or in the context of the Community market. The Commission considers both situations below.

(111)

As set out in recital 153 of the provisional Regulation and recital 197 of the definitive Regulation, some overcapacity existed worldwide during a number of years including the period under consideration. No major change in this overcapacity situation has been observed during the period under consideration and in particular at the end of this period leading to the IP. This worldwide overcapacity may have asserted pressure on prices during the period under consideration, which may have contributed to the injury suffered by the Community industry. This finding has not been disputed by any party concerned.

(112)

However, as set out in recital 147 of the provisional Regulation, the worldwide overcapacity was estimated to be approximately 20 % of the market. Hynix accounted for around 17 % of worldwide production during the IP. Given that Hynix was virtually bankrupt during the IP and remained afloat only through the subsidies provided by the GOK, it can be concluded that the worldwide overcapacity situation has been artificially maintained through the subsidisation of Hynix. Thus, the effect of worldwide overcapacity on the injurious situation of the Community industry was channelled through Hynix’s subsidised imports. This effect has been analysed in recitals 144 to 149 of the provisional Regulation and recitals 191 and 192 of the definitive Regulation. In this respect, the effect of worldwide overcapacity created by Hynix cannot be considered as a factor separate from the effects of the subsidised imports.

(113)

In any event, even if the effect of worldwide overcapacity could be considered separately from the effect of subsidised imports, the following can be observed. Information from the questionnaire responses of the four producers of DRAMs cooperating with the original investigation (i.e., Samsung, Hynix, Infineon and Micron, which together represented around 80 % of the worldwide DRAMs market) shows that their total capacity in Mbits increased by 129 % between 1999 and 2000 and by 85 % between 2000 and the IP. Their total production in Mbits increased in line by 124 % between 1999 and 2000 and by 81 % between 2000 and the IP. In other words, the capacity and the production of the four major worldwide producers increased significantly more between 1999 and 2000, the good year during which the Community industry maintained some price stability and made profits, than between 2000 and the IP (2001), during which the Community industry suffered significant losses because of the collapse of prices (see recitals 128 and 131 and relevant tables of the provisional Regulation). Consequently, the existing worldwide overcapacity throughout the period under consideration did not, in the Commission’s view, contribute significantly to the collapse of prices and the material injury suffered by the Community industry in the IP.

(114)

In addition, record evidence shows that Community overcapacity was not a factor of injury. During 1999 to 2000, Community capacity more than tripled (increasing by 252 %), and from 2000 to 2001 it nearly doubled (increasing by 92 %). These additions to Community capacity were accompanied by increases in Community production. While capacity utilisation in the Community did drop by 28 % from 1999 to 2000, it recovered by 4 % from 2000 to 2001 (see recital 125 and relevant table of the provisional Regulation).

(115)

Trends in Community overcapacity (defined as the inverse of Community capacity utilisation) show little or no correlation with trends in the sales prices of the Community industry. In particular, during 1999 to 2000, Community overcapacity rose by 28 %, but the sales prices of the Community industry fell by only 5 %. Further, from 2000 to 2001, Community overcapacity went down by 4 % but the sales prices of the Community industry plummeted by 77 % (see recitals 125 and 128 and relevant tables of the provisional Regulation). It may be concluded that, if rising overcapacity were correlated to falling price levels, the significant increase in overcapacity over 1999 to 2000 would have resulted in severe price depression, whereas the contraction in overcapacity in 2000 to 2001 would have led to rising prices, or at least to some degree of price stability. Record evidence shows the opposite. Over 1999 to 2000, there was very mild price depression, despite the significant increase in overcapacity, while from 2000 to 2001, there was substantial price depression, despite the contraction in overcapacity. This stands in sharp contrast with the close correlation between trends in the prices of the subsidised imports and trends in the sales prices of the Community industry. In fact, during 2000 to 2001 the trend in the sales prices of the Community industry mirrored almost perfectly the trend in the prices of the subsidised imports: over this period, the former dropped by 74 % and the latter did so by 77 % (see recitals 119 and 128 of the provisional Regulation).

(116)

Record evidence also shows that trends in Community overcapacity bear no correlation with trends in other key indicators of the condition of the Community industry. In particular, during 1999 to 2000, the profitability of the Community industry was rising, and returns on investment and cash flow were positive, in spite of the significant increase in Community overcapacity. In contrast, in 2001 the Community industry suffered heavy losses, and returns on investment and cash flow were negative, in spite of the contraction in Community overcapacity during 2000 and 2001. The performance of the Community industry over 2000 to 2001 in terms of profitability, returns on investment and cash flow is correlated instead with the trend in the prices of the subsidised imports over that period. As noted, in 2000 to 2001, as the Community industry suffered heavy losses and negative cash flow and return on investment, the prices of the subsidised imports decreased by 74 %.

(117)

Hynix claimed that Community overcapacity was another known factor contributing to material injury during the period under consideration. However, trends in Community overcapacity show no or only negligible correlation with trends in the condition of the Community industry. The Commission therefore concludes that Community overcapacity cannot be considered as a factor of material injury to the Community industry. Given that Community overcapacity did not adversely affect the condition of the Community industry, there is no need to distinguish any alleged injury caused by capacity from the injury caused by Hynix’s imports. The Commission believes this finding is consistent with the findings of recent WTO Panels reviewing determinations in respect of non-attribution described in recital 106 above.

(118)

As reviewed, there is no demonstrated causal relationship between overcapacity and either price declines or injury to the Community industry. In contrast, the Community industry suffered severe declines during 2001, which is the same period when subject import prices plummeted and when subject import volumes reached peak penetration in the Community market. This contrast establishes that subject import volumes and price effects are a separate and distinguishable cause of material injury to the Community industry as compared with overcapacity.

(119)

The Panel noted in paragraph 7.426 of its report that there is no requirement to consider non-subsidised imports on a country-per-country basis. The Panel also noted in paragraph 7.435 that a satisfactory explanation of the nature and extent of the injurious effects of non-subsidised imports, as distinguished from the injurious effects of the subsidised imports, should be provided. Accordingly, the effects of non-subsidised imports were examined by considering jointly imports from countries other than Korea (e.g. the United States, Japan, and Taiwan) and imports from Samsung. Recital 151 of the provisional Regulation and recitals 194 and 200 of the definitive Regulation provide the growth rates of import volume, market share, and average prices of non-subsidised imports during the period under consideration. These data are re-stated below in aggregated index form (76). Given the very limited number of players in the DRAMs market, for reasons of confidentiality, indices are used to show the evolution of trends.

(120)

During the period under consideration, non-subsidised imports increased significantly less than did subsidised imports, i.e., by 133 % as opposed to 361 % (see recital 177 and relevant table of the definitive Regulation). It is noted that during the same period Community consumption increased by 316 %. Hence, whereas the rate of growth in non-subsidised imports was only a fraction of the rate of growth of Community consumption, subsidised imports increased well in excess of Community consumption. Accordingly, during the period under consideration, the market share of non-subsidised imports contracted by 44 % whilst the market share of subsidised imports expanded by 20 % (see recital 185 and relevant table of the definitive Regulation). If the market share of non-subsidised imports heavily declined during the period under consideration, as shown by the above facts, it can be concluded that non-subject imports did not injure the Community industry by reason of their volume effects.

(121)

Record evidence also indicates that non-subject imports did not injure the Community industry by reason of their price effects. During the IP, the prices of non-subsidised imports exceeded not only the prices of subsidised imports but also the sales prices of the Community industry. In particular, during the IP the average price of non-subsidised imports was 19 % higher than the average price of subsidised imports and 24 % higher than the average sales price of the Community industry. In addition, during the period under consideration, the prices of non-subsidised imports fell, but not anywhere as much as the prices of subsidised imports and the Community industry. In particular, during the period under consideration the average price of non-subsidised imports decreased by 52 %. In contrast, over such period, the average price of subsidised imports decreased by 80 % (see recital 178 and relevant table of definitive Regulation) and the average sales price of the Community industry by 79 % (see recital 128 and relevant table of the provisional Regulation). On the basis of these facts, it can be concluded that the prices of non-subsidised imports did not affect adversely the sales prices of the Community industry.

(122)

As is the case with overcapacity, there is no demonstrated causal relationship between non-subject imports and price declines or injury to the Community industry. There is, at the same time, a clear deterioration in Community industry performance while subject import volumes steadily increased and while subject import prices undercut the Community like product. This contrast establishes that subject import volumes and price effects are a separate and distinguishable cause of material injury to the Community industry as compared with non-subject imports.

(123)

Therefore, the Commission is of the view that its redetermination satisfies the non-attribution requirement under Article 8(7) of the basic Regulation (which generally corresponds to Article 15.5 of the SCM Agreement) in respect of non-subject imports by separating and distinguishing this other factor from subsidised imports and showing that, during the period under consideration, non-subject imports did not cause material injury to the Community industry, either in terms of their volume effects or in terms of their price effects.

(124)

Having separated, where appropriate, and analysed further the effects, if any, of the decline in demand, overcapacity and other non-subsidised imports, it is considered that no injury caused by any of these factors has been attributed to the subsidised imports within the meaning of Article 8(7) of the basic Regulation. Furthermore, it is confirmed that the subsidised imports through the effects set out in recitals 144 to 149 of the provisional Regulation and recitals 191 and 192 of the definitive Regulation have caused material injury to the Community industry within the meaning of Article 8(6) of the basic Regulation.

(125)

The contents and conclusions reached in recitals 206 to 209 of the definitive Regulation remain unaffected by the Panel report.

(126)

The Panel report and the reassessed findings do not concern Samsung which was not found to be subsidised and thus was attributed a 0 % duty rate in the original investigation. The revised subsidy amount for Hynix was found to be lower than the injury elimination level established during the original investigation and, thus, the revised countervailing duty rate should be based on the revised subsidy amount expressed ad valorem, i.e. 32,9 %.

(127)

It is noted that the revised duty rate for Hynix should apply on the duty as specified by Regulation (EC) No 2116/2005.

(128)

The GOK, Hynix and the Community producers Infineon and Micron received disclosure of the reassessed findings and were given an opportunity to comment and to be heard. They each submitted comments (77).

(129)

The GOK generally submits that the revised findings do not reasonably address the Panel’s recommendations. The Commission does not agree, for the reasons set out in this Regulation. The Commission does not agree that the revised duty must necessarily be lower than the original duty, and notes that in any event it is. The GOK specifically submits that the Commission should have used seven banks in the syndicated loan as a benchmark. The Commission disagrees, particularly for the reasons set out in recitals 16 to 24 above. The GOK also submits that the Commission should have made its benchmark calculations differently. In the light of the GOK’s comments the Commission has made some adjustments to its calculations, as reflected in this Regulation. For the reasons set out in this Regulation, the other aspects of the calculations remain unchanged. Finally, the GOK submits that the Commission’s injury redetermination does not separate and distinguish the effects of other known factors, particularly with respect to economic downturn, overcapacity and other non-subsidised imports. The Commission disagrees, particularly for the reasons set out in recitals 100 to 123 above.

(130)

Hynix aligns itself with the submissions of the GOK, discussed in the preceding recital. In addition, Hynix submits that the period of time during which duties should be imposed should have been terminated on 1 January 2006. This is not a matter that was the subject of dispute or recommendations by the WTO Panel; and is not therefore part of the ‘measure taken to comply’ with the Panel's recommendations which is encapsulated within this Regulation. This Regulation does not alter the impending expiry date of the definitive Regulation. The question of an early termination of the measure rather falls within the ambit of an eventual interim review, not within the ambit of implementation of the Panel report. Hynix also submits that the revised rate of duty must apply retroactively from the date on which measures were first imposed. The Commission similarly considers that this issue is not part of the ‘measure taken to comply’ and in any event does not agree. Nothing in the relevant legal rules obliges the Commission to apply the revised duty retroactively, the Panel limiting itself to recommending that the measure be brought in conformity, which is achieved by the present Regulation.

(131)

Infineon substantially agrees with the redetermination. However, it submits that the Commission should impose a higher duty based on a proxy benchmark and the original injury determination, rather than the lower duty set out in this Regulation. The Commission has carefully considered that request. However, for the reasons set out in this Regulation and due to the specific purpose of this Regulation, notably as an implementation of the Panel's recommendations, the Commission has not altered its initial determination on this point.

(132)

Micron also substantially agrees with the redetermination, referring in this respect to facts and evidence that tend to confirm and corroborate the Commission’s conclusions, including an expert report prepared by Professor J. Hausman. Therefore, Micron’s comments do not necessitate any modification of those conclusions as such. Micron further submits that the Commission should reaffirm, based on current WTO precedent, that the GOK entrusted or directed the banks in the May 2001 restructuring programme and should countervail the subsidies provided by the public bodies. The Commission has carefully considered this submission which, in principle, has merit. However, as noted in this Regulation, the May 2001 restructuring programme would not in any event be countervailed, given that the largest proportion of the benefit would have been accounted for in the October 2001 measures. Mainly for this reason, the Commission has decided not to return to these issues in this implementing Regulation.

(133)

Thus, all comments submitted by the interested parties were considered and taken into account where appropriate. However, certain comments were outside the scope of this Regulation which is only intended to implement the DSB recommendations. Thus, such comments have not resulted in any changes to this Regulation,

1.

Article 1(2) and (3) shall be replaced by the following:

‘2.   The rate of the definitive countervailing duty applicable to the net free-at-Community-frontier price, before duty, shall be as follows:

3.   Upon presentation to the Member State’s customs authorities of the customs declaration for release into free circulation of multi-combinational forms of DRAMs, the declarant shall indicate in box 44 of the Single Administrative Document (SAD) the reference number corresponding to the below product/origin descriptions indicated in the following table. The rate of the definitive countervailing duty applicable to the net-free-at-Community-frontier price, before duty, shall be as follows:

2.

Article 1(5) shall be replaced by the following:

‘5.   In the case of no indication of any reference number in the SAD, as foreseen in paragraph 3, or if the customs declaration is not accompanied by any statement in the cases required in paragraph 4, the multi-combinational form of DRAM shall — unless the contrary is proved — be considered as originating in the Republic of Korea and as having been manufactured by all companies other than Samsung, and the countervailing duty rate of 32,9 % shall apply.

In cases where some of the DRAM chips and/or mounted DRAMs incorporated in multi-combinational forms of DRAMs are not clearly marked and their manufacturers are not clearly identifiable from the statement required in paragraph 4, such DRAM chips and/or mounted DRAMs shall — unless the contrary is proved — be considered as originating in the Republic of Korea and as having been manufactured by companies subject to the countervailing duty. In such cases, the countervailing duty rate applicable to multi-combinational forms of DRAMs shall be calculated on the basis of the percentage represented by the net free-at-Community-frontier price of the DRAM chips and/or mounted DRAMs originating in the Republic of Korea out of the net free-at-Community-frontier price of the multi-combinational forms of DRAM, as established in the table contained in paragraph 3, numbers 2 to 7. Should, however, the value of the aforementioned DRAM chips and/or mounted DRAMs be such to determine that the multi-combinational forms of DRAMs in which they are incorporated becomes of Korean origin, the countervailing duty rate of 32,9 % shall apply to the multi-combinational forms of DRAMs.’.