Report of the Panel

(Continued)

2. The "as such" claims of the European Communities with respect to "standard zeroing procedures"

(a) Measures at issue

7.70 The European Communities uses the term "Standard Zeroing Procedures" in this proceeding to refer to specific lines of computer code contained in the AD Margin Programme, one of the computer programmes that incorporate USDOC's current calculation methodology, which separate sales with "positive margins" and sales with "negative margins" and subtotal only the dumping amounts for sales with "positive margins". The European Communities refers in particular to the line WHERE EMARGIN GT 0.¹⁷⁶

7.71 The term "Standard Zeroing Procedures" is not used in United States anti-dumping laws and regulations.

7.72 Since the claim of the European Communities is with respect to the "'standard zeroing procedures' (or the United States practice or methodology of zeroing)", the Panel has sought clarification from the European Communities of whether it is challenging a practice or methodology as a measure distinct from the standard zeroing procedures. In response, the European Communities has stated that, although the computer programme is "the root of the problem", in order " to cover all the bases", it is also challenging the "consistent practice" as such of the United States with regard to zeroing.¹⁷⁷ It is therefore our understanding that the European Communities claims that both the "Standard Zeroing Procedures" and the United States' practice or methodology of zeroing are WTO-inconsistent as such.

(b) Arguments of the parties

7.73 The European Communities submits that the measure that consists of or includes the Standard Zeroing Procedures (and the [Anti-Dumping Procedures] Manual to the extent it refers to the Standard Computer Programmes), or the United States practice or methodology of zeroing¹⁷⁸ is as such inconsistent with Articles 2.4, 2.4.2, 5.8, 9.3, 1 and 18.4 of the AD Agreement; Articles VI:1 and VI:2 of the GATT 1994 and Article XVI:4 of the WTO Agreement.

7.74 The European Communities argues that any act or omission attributable to a WTO Member can be a measure that can be challenged in WTO dispute settlement, including not only particular acts applied to a specific situation but also acts setting forth rules or norms that are intended to have general and prospective application. There is no basis in the practice of the GATT and WTO generally or in the provisions of the AD Agreement for finding that only certain types of measure can be challenged as such under the AD Agreement. In order to demonstrate that the "Standard Zeroing Procedures" are a "measure" the European Communities submits inter alia that the Anti-Dumping Procedures Manual and the "Standard Zeroing Procedures" are "administrative procedures" within the meaning of Article 18.4 of the AD Agreement; that they have at least potentially some legal effects; that in practice USDOC treats the "Standard Zeroing Procedures" as binding, at least until changed; and that the purpose and effect of the "Standard Zeroing Procedures" is to ensure that equivalent situations are automatically and consistently treated equally.¹⁷⁹

7.75 The European Communities argues that, in light of Article 18.4 of the AD Agreement and Article XVI:4 of the WTO Agreement, the correct legal test to apply to "as such" claims is whether or not a measure that is being challenged as such is "in conformity" with the relevant provisions of the AD Agreement, and not whether the measure is mandatory or discretionary, whether it is binding or whether it requires a Member to act in a WTO-inconsistent manner in all cases. In any event, whichever test is used, the mathematical character of the "Standard Zeroing Procedures" necessarily means that they are WTO-inconsistent as such because they provide for the application of a standard that differs from the standard set out in Article 2.4.2 of the AD Agreement.¹⁸⁰

7.76 The European Communities states that it is challenging, first and foremost the Standard AD Margin Programme but that to the extent necessary, and in order to "cover all the bases" it is challenging each part of the Anti-Dumping Procedures Manual referenced in the factual part of its first written submission, including the instruction to use the Standard AD Margin Programme.¹⁸¹

7.77 The European Communities submits that it is not necessary for a measure to set out or establish rules or norms in order for the measure to be reviewable under the DSU. In any event, the Standard Anti-Dumping Margin Program does set out or establish rules or norms because it is repeatedly referred to as a standard, is published or issued under the authority of USDOC and provides guidance to USDOC officials. The fact that the "Standard Zeroing Procedures" are contained in a computer programme and that they can be seen as implementing something else does not mean that they cannot be challenged as a measure. That the Standard Anti-Dumping Margin Programme can be changed is irrelevant; what matters is that, as it stands, it contains a standard rule different from the rule prescribed by Article 2.4.2 of the AD Agreement.¹⁸²

7.78 The European Communities points out that while the standard computer programmes used by USDOC may have undergone various changes, they have never been changed with respect to the zeroing procedures at issue in this dispute.¹⁸³

7.79 Regarding the issue of whether "practice" can be a "measure", the European Communities submits that whatever panels may or may not have said on this point in past cases, the particular factual circumstances of the present case justify an "as such" finding of inconsistency because the practice has been lifted up by the investigating authority itself into the Standard AD Margin Programme and the Anti-Dumping Procedures Manual. That USDOC cannot depart from the "Standard Zeroing Procedures" without providing a reasoned explanation confirms that these procedures have a legal effect.¹⁸⁴

7.80 The United States submits that the measures referred to by the European Communities in its claims regarding "Standard Zeroing Procedures" either are not measures at all or are not mandatory measures within the meaning of the mandatory/discretionary test.¹⁸⁵ First, with respect to the Anti-Dumping Procedures Manual, even if this Manual is considered a measure, it does not preclude the USDOC decision maker from offsetting negative margins nor does it mandate that the decision maker ignore negative margins. The Manual is nothing more than a source of guidance and training for USDOC personnel. The European Communities provides no evidence to support its assertion that USDOC considers itself bound by the Manual, and the court decisions cited by the European Communities expressly state that the Manual is not legally binding. For purposes of the mandatory/discretionary test, the relevant decision maker is the Assistant Secretary of Commerce for Import Administration, not the staff that implement the decision maker's decisions. The Assistant Secretary clearly is not obligated to follow the Manual and can decide in a particular case to offset negative margins. Second, the Anti-Dumping Margin Program is not a measure for purposes of WTO dispute settlement. It does not set out or establish rules or norms but is a piece of computer software that, at most, implements rules or norms contained in some other instrument. Even assuming that the Anti-Dumping Margin Program is a measure, it does not preclude the USDOC decision maker from offsetting negative dumping margins nor does it require the USDOC decision maker to ignore negative dumping margins.¹⁸⁶

7.81 The United States further submits that, to the extent that the claims of the European Communities regarding the "Standard Zeroing Procedures" pertain to the "practice or methodology of zeroing", they must be rejected because "practice" is not a "measure" and, even if it were, it would not be a mandatory measure within the meaning of the mandatory/discretionary test. The European Communities fails to explain how "practice" - the repeated application of a particular measure � creates a new and separate "autonomous measure". The Panel in US � Steel Plate specifically rejected the view that a practice can be challenged as a measure. Even if past instances of not offsetting negative margins were deemed to be a measure, it would not be WTO-inconsistent under the mandatory/discretionary test because there is no principle of administrative stare decisis under United States' law.¹⁸⁷

7.82 The United States submits that the argument of the European Communities and Japan that the lines of computer programming code at issue in this dispute are covered by Article 18.4 of the AD Agreement as "administrative procedures" is contrary to the Vienna Convention rules on treaty interpretation because it is inconsistent with the ordinary meaning of the term "administrative procedures". The United States also submits that no Member has ever notified a computer programme to the Committee on Anti-Dumping Practices. The Standard Anti-Dumping Margin Programme has often been changed, and is revised in each case in which a margin calculation is performed. The Assistant-Secretary of Import Administration is not obligated to follow either the Standard Anti-Dumping Margin Programme or the Anti-Dumping Procedures Manual, and there is no evidence that the United States treats zeroing as binding or mandatory.¹⁸⁸

7.83 The United States submits that the position of the European Communities on the Standard Anti-Dumping Margin Program as a measure that can be found to be WTO-inconsistent as such is based on a flawed approach to the mandatory/discretionary distinction that has never been applied in any WTO panel or Appellate Body report. The United States also asserts that the European Communities is incorrect in arguing that it is not necessary for a measure challenged as such to prescribe a result and in suggesting that the status of the Standard Anti-Dumping Margin Program is similar to that of the USDOC Regulations.¹⁸⁹

(c) Arguments of third parties

7.84 Brazil argues that the measure that is the subject of the present dispute is zeroing as a "policy". As such, it clearly falls within the scope of the expression "rules, norms and standards" used by the Appellate Body in US � Corrosion-Resistant Steel Sunset Review. This policy is reflected in a combination of the consistent application of the policy in every case in which there is negative dumping and the USDOC Antidumping Manual. These, in combination, set forth a rule, norm or standard for applying the zeroing methodology in US anti-dumping proceedings, which the Appellate Body has determined may be challenged in accordance with Article 18.4 of the AD Agreement, and should be found to violate Article 2.4 of that Agreement. In view of the practice of consistent application of zeroing, it is irrelevant whether these measures mandate application of zeroing.

7.85 China submits that the notion of "measures" that can be submitted to dispute settlement must be interpreted broadly and refers in this respect to the Appellate Body's analysis in US � Corrosion-Resistant Steel Sunset Review. In light of the Appellate Body's findings on this issue, a measure can be challenged as such if it is a generally applicable rule, norm or standard adopted by a Member in connection with the conduct of anti-dumping proceedings. The panel reports in US � Section 301 and US � 1916 Act (Japan) and the Appellate Body reports in US � 1916 Act and US � Corrosion-Resistant Steel Sunset Review indicate that the mandatory nature of a measure is not the decisive factor in deciding whether legislation is inconsistent as such with WTO obligations.

7.86 Hong Kong, China submits that, in light of the objectives of the DSU, the term "measures" in Article 3.3 of the DSU must be interpreted broadly. Since the Appellate Body has affirmed that in principle any act or omission attributable to a Member can be a measure of that Member for purposes of WTO dispute settlement, USDOC's AD Margin Programme, and more generally its practice of zeroing, constitutes a measure. In order to challenge a measure as such, it is not necessary for that measure to be mandatory. The consistent use by USDOC of the computer instructions in the AD Margin Programme demonstrates that they are in the nature of norms or rules of general and prospective application. That they can be changed is irrelevant because all norms and rules, including mandatory rules, can be changed. Even if practice in and of itself cannot be considered to be setting forth rules or norms that are intended to have general and prospective application, a repeated pattern of similar responses to a set of circumstances must be regarded as clear evidence of the existence of rules or norms which are of general and prospective application.

7.87 India argues that it follows from the rulings of the Appellate Body in US � Corrosion-Resistant Steel Sunset Review and US � Oil Country Tubular Goods Sunset Reviews that the model zeroing and simple zeroing procedures of USDOC can be challenged as such because the standard computer programmes, which include the procedures that result in zeroing through model or simple zeroing, can be characterized as norms or rules that are applied on a generalized and prospective basis. India recalls that the Appellate Body has held that in principle non-mandatory measures can be challenged as such in WTO dispute settlement.

7.88 Japan submits that in considering the issue of the measures that can be challenged as such, the Appellate Body has held that the word "measure" has a broad meaning, that an alleged "measure" will be assessed in WTO law irrespective of its legal character in domestic law, and that a "measure" need not be binding or mandatory in domestic law. Model and simple zeroing are measures that can be challenged as such because they are rules, norms or standards applied by the United States in anti-dumping proceedings on a generalized and prospective basis.

7.89 Mexico argues that in principle, for the purposes of a dispute settlement proceeding any act or omission attributable to a WTO Member may be deemed to be a measure. The Anti-Dumping Procedures Manual, the standard computer programmes and the Standard AD Margin Programme are components of "Standard Zeroing Procedures" and are built into the overall structure in which anti-dumping proceedings are conducted in the United States. Furthermore, USDOC applies the "Standard Zeroing Procedures" in all investigations and duty assessment proceedings. Thus, in practice they are compulsory.

7.90 Norway argues that, as stated by the Appellate Body, there are no limitations on the types of measures that can be challenged as such in WTO dispute settlement and that even non-mandatory measures can be the subject of dispute settlement. The AD Margin Programme is a norm or standard adopted by the United States in connection with the conduct of anti-dumping proceedings. The programme is a set of normative rules that apply mechanistically and the effect of which is utterly predictable. The fact that the Assistant Secretary of Commerce for Import Administration can change the Anti-Dumping Procedures Manual and the Anti-Dumping Margin Programme without notice is irrelevant because all measures are subject to changes. The Manual and the AD Margin Programme are prescriptive for a certain WTO-inconsistent result until they are changed.

(d) Evaluation by the Panel

7.91 The European Communities claims that "Standard Zeroing Procedures", by which it means certain lines of computer programming code contained in what it terms standard computer programmes used by USDOC in the calculation of dumping margins, are WTO-inconsistent as such, i.e. independently from their application in specific cases.

7.92 As a starting point for our analysis of this claim, we briefly recapitulate the main points emerging from recent Appellate Body decisions regarding measures that can be challenged as such in WTO dispute settlement.

7.93 In US � Corrosion-Resistant Steel Sunset Review, the Appellate Body discussed the issue of "the type of measures that can, as such, be the subject of dispute settlement proceedings" under the AD Agreement. The Appellate Body held that "in principle any act or omission attributable to a WTO Member can be a measure of that Member for purposes of dispute settlement", and that, in addition, GATT and WTO dispute settlement practice confirmed that the term "measure" also comprised "acts setting forth rules or norms that are intended to have general and prospective application" or, in other words, "instruments of a Member containing rules or norms", irrespective of their application in particular instances. Specifically with respect to the AD Agreement, the Appellate Body held that the phrase "laws, regulations and administrative procedures" used in Article 18.4 of that Agreement implies that "the entire body of generally applicable rules, norms and standards adopted by Members in connection with the conduct of anti-dumping proceedings" can be challenged as such.¹⁹⁰ The Appellate Body concluded that:

"..there is no basis either in the practice of the GATT and the WTO generally or in the provisions of the AD Agreement, for finding that only certain types of measures can, as such, be challenged in dispute settlement proceedings under the AD Agreement".¹⁹¹

7.94 With regard to the particular measure at issue in that case, USDOC's Sunset Policy Bulletin ("SPB"), the Appellate Body found that certain provisions of the SPB were not as such inconsistent with Articles 6.10 and 11.3 of the AD Agreement, and that it was unable to rule on the claim that certain other provisions of the SPB were as such inconsistent with Article 11.3 because of "the lack of relevant actual findings by the Panel or uncontested facts on the Panel record".¹⁹²

7.95 The Appellate Body confirmed that the SPB is a measure that can be challenged as such in US � Oil Country Tubular Goods Sunset Reviews, when it held that the SPB was covered by the concept of "acts setting forth rules or norms that are intended to have general and prospective application":

"In our view, the SPB has normative value, as it provides administrative guidance and creates expectations among the public and among private actors. It is intended to have general application, as it is to apply to all the sunset reviews conducted in the United States. It is also intended to have prospective application, as it is intended to apply to sunset reviews taking place after its issuance. Thus, we confirm�once again�that the SPB, as such, is subject to WTO dispute settlement."¹⁹³

In this connection, the Appellate Body rejected as irrelevant the arguments of the United States that the SPB is not a legal instrument under United States law, that the SPB does not bind USDOC and that USDOC is entirely free to depart from the SPB at any time.¹⁹⁴

7.96 It can thus be inferred from the Appellate Body's reasoning with regard to the SPB as a measure that can be challenged as such in WTO dispute settlement that it is possible for a measure to be challenged as an act or instrument that "sets forth rules or norms that are intended to have general and prospective application" even where the measure in question is not "a legal instrument" under the law of a Member and does not bind an administering agency.¹⁹⁵

7.97 We note that a prominent theme in the arguments of the parties (and third parties) on the "Standard Zeroing Procedures" as a measure has been whether these procedures are an act or instrument that sets forth rules or norms intended to have general and prospective application. In this regard, we consider that to characterize the "Standard Zeroing Procedures" as an act or instrument that sets forth rules or norms intended to have general and prospective application is somewhat difficult to reconcile with the fact that the "Standard Zeroing Procedures" are only applicable in a particular anti-dumping proceeding as a result of their inclusion in the computer programme used in that particular proceeding. The need to incorporate these lines of computer code into each individual programme indicates that it is not the "Standard Zeroing Procedures" per se that set forth rules or norms of general and prospective application. For this reason, we also question whether these "Standard Zeroing Procedures" are "administrative procedures" within the ordinary meaning of that term as used in Article 18.4 of the AD Agreement. The "Standard Zeroing Procedures" by themselves do not create anything and are simply a reflection of something else.

7.98 We note in this regard, however, that the European Communities has not limited its claim to the "Standard Zeroing Procedures" in and of themselves but has also challenged what it describes as "the United States practice or methodology of zeroing". We will therefore proceed to consider whether there exists what the European Communities terms methodology and whether this methodology can be found to be WTO-inconsistent.

7.99 In the latter regard, we note that the SPB, which the Appellate Body has found to be a measure challengeable as such, notwithstanding that it has no legally binding status under United States law, embodies a norm to act in a given way in a given situation. That is precisely what the Appellate Body focuses on when it refers to the SPB as an instrument with normative value that contains rules or norms of general and prospective application. If, as confirmed by the Appellate Body, a non-legally binding policy instrument such as the SPB is a measure that can be challenged as such, it must logically also be possible to challenge as a measure a norm that is not expressed in the particular form of an official written statement but the existence of which is made manifest on the basis of other evidence.

7.100 We find support for this approach in the reasoning of the Appellate Body regarding the rationale for allowing "as such" claims. We note in particular the following statement in paragraph 82 of the Appellate Body Report in US � Corrosion-Resistant Steel Sunset Review:

"... the disciplines of the GATT and the WTO, as well as the dispute settlement system, are intended to protect not only existing trade but also the security and predictability needed to conduct future trade. This objective would be frustrated if instruments setting out rules or norms inconsistent with a Member's obligations could not be brought before a panel once they have been adopted and irrespective of any particular instance of application of such rules or norms. It would also lead to a multiplicity of litigation if instruments embodying rules or norms could not be challenged as such, but only in the instances of their application. Thus, allowing claims against measures, as such, serves the purpose of preventing future disputes by allowing the root of WTO-inconsistent behaviour to be eliminated."¹⁹⁶

In our view, the objective of protecting the security and predictability needed to conduct future trade can just as readily be frustrated if well-established norms that systematically and predictably lead to WTO-inconsistent actions cannot be challenged or if they can be challenged only if they are embodied in a particular type of instrument. Similarly, not allowing "as such" claims against such norms that are "the root of WTO-inconsistent behaviour" might well entail a multiplicity of litigation. While an agency probably can in most cases more easily depart from its own established norm than from a law or regulation, the argument that there cannot be WTO-inconsistency as such if an agency has discretion to make a change strikes us as artificial, at the very least in the case of a norm that has been applied invariably for a considerable period of time. In such a case, WTO-inconsistent conduct may be as predictable as when WTO-inconsistent conduct is envisaged in a law or regulation. We also consider that to accord decisive weight to the nature of a particular instrument in which a norm manifests itself creates a risk of addressing symptoms rather than causes.

7.101 In light of these considerations, we must determine whether in the present case what is challenged by the European Communities as methodology constitutes a norm that is WTO-inconsistent as such. We note that the idea that methodology can be challenged as such is not new.¹⁹⁷

7.102 We realize that "as such" challenges are "serious challenges" in that they "seek to prevent Members ex ante from engaging in certain conduct".¹⁹⁸ In this regard, we consider that a finding that a norm is as such WTO-inconsistent must rest on solid evidence that enables a panel to determine the precise content of that norm and the conduct to which that norm will necessarily give rise in future. We are cognizant that norms are not always susceptible of such a clear definition. In the case of the SPB, the necessary precision and predictability resulted from the availability of an official policy statement that set out with a considerable degree of detail the methodology the USDOC intended to apply in certain situations. There are, however, other types of evidence that can be used to establish with the necessary degree of precision the content of a norm and the future conduct it will generate. While we do not focus on the "Standard Zeroing Procedures" as a measure per se, we consider that they can be relevant evidence to ascertain the existence of a methodology.¹⁹⁹

7.103 Thus, in the present case, the instruction not to include comparison results with negative margins in the numerator of the dumping margin is reflected in certain lines of computer code that are always included in the computer programmes used by USDOC in anti-dumping proceedings. Although the United States has emphasized that what the European Communities refers to as standard computer programme has frequently been revised, the United States does not contest that the lines of computer code identified by the European Communities as "Standard Zeroing Procedures" are a constant feature of the computer programmes used by USDOC to perform dumping margin calculations. The evidence before the Panel also indicates that this exclusion of comparison results with negative margins has been invariably performed by USDOC for an extended period of time. In response to a panel question whether there have been cases in which the Anti-Dumping Margin Program has been applied without zeroing, the United States states that, whether calculations have been done by hand or by computer, it is unable to identify any instance where USDOC had given a credit for non-dumped sales.²⁰⁰ The United States has not contested in this proceeding that USDOC's zeroing methodology reflects a deliberate policy.

7.104 We thus consider that the evidence before us indicates that the zeroing methodology manifested in the "Standard Zeroing Procedures" represents a well-established and well-defined norm followed by USDOC and that it is possible based on this evidence to identify with precision the specific content of that norm and the future conduct that it will entail.²⁰¹ In our view, the situation is the same as in the case of the Sunset Policy Bulletin, except that the zeroing methodology is not expressed in writing.

7.105 We recall our finding that the use of model zeroing in the anti-dumping investigations at issue in this dispute is inconsistent with Article 2.4.2 of the AD Agreement. Therefore, in light of the considerations in the preceding paragraphs, we find that USDOC maintains a norm that will necessarily produce WTO-inconsistent actions.

7.106 In light of all the foregoing considerations, the Panel finds that the United States' zeroing methodology, as it relates to original investigations, is a norm which, as such, is inconsistent with Article 2.4.2 of the AD Agreement.

7.107 The Panel considers that it is not necessary to make a finding on whether the Anti-Dumping Procedures Manual is WTO inconsistent as such. The Manual has been referred to by the European Communities principally as evidence to confirm the "standard" character of the "Standard Zeroing Procedures".

7.108 In light of its finding that the United States' zeroing methodology, as it relates to original investigations, is a norm which, as such, is inconsistent with Article 2.4.2 of the AD Agreement, the Panel considers that it is not necessary to address the claim of the European Communities that this methodology is also inconsistent as such with Article 2.4 of the AD Agreement.

7.109 The Panel also perceives no need to pronounce on the dependent claims raised by the European Communities under Articles 1; 3.1, 3.2 and 3.5; 5.8; 9.3; and 18.4 of the AD Agreement, Articles VI:1 and VI:2 of the GATT 1994 and Article XVI: 4 of the WTO Agreement. Deciding such dependent claims would provide no additional guidance as to the steps to be undertaken by the United States in order to implement our recommendation regarding the violation on which it is dependent.

E. CLAIMS OF THE EUROPEAN COMMUNITIES IN RESPECT OF CERTAIN ADMINISTRATIVE REVIEWS

1. Measures at issue

7.110 The European Communities requests the Panel to find that the United States has acted inconsistently with its WTO obligations in 16 "anti-dumping duty administrative review" proceedings listed in Exhibits EC-16 to EC-31²⁰² because: (a) USDOC compared export price and normal value on an asymmetrical, average-to-transaction basis; and (b) in calculating a weighted average dumping margin based on these average-to-transaction comparisons, USDOC did not take into account any amounts by which individual export transactions exceeded average normal value. The European Communities uses the term simple zeroing to refer to this second aspect of the calculation methodology used by USDOC. The United States rejects the simple zeroing terminology employed by the European Communities but does not contest that the European Communities has accurately described the comparison methodology used by USDOC in these administrative reviews.

2. Order of analysis

7.111 The claims of the European Communities with respect to the administrative reviews at issue are based on: (1) Articles 2.4, 2.4.2, 11.1, 11.2, 9.3, 1 and 18.4 of the AD Agreement; (2) Articles VI:1 and VI:2 of GATT 1994; and (3) Article XVI:4 of the WTO Agreement.

7.112 As in the case of the claims of the European Communities with regard to the model zeroing methodology used by the USDOC in original investigations, we consider that it is logical to address first the claim of the European Communities under Article 2.4.2 of the AD Agreement.

3. Claim of the European Communities under Article 2.4.2 of the AD Agreement

(a) Arguments of the parties

7.113 The European Communities submits that the United States acted inconsistently with Article 2.4.2 of the AD Agreement because USDOC used the asymmetrical, average-to-transaction comparison referred to in the second sentence of Article 2.4.2 when the conditions for the use of that exceptional method were not met and because it failed to use one of the two symmetrical methods set out in the first sentence of Article 2.4.2. With respect to simple zeroing, the European Communities submits that the reasoning of the Appellate Body in EC � Bed Linen and US � Softwood Lumber V with regard to model zeroing applies to any parameter used to define the scope of an investigation, be it product, time, level of trade or region. Thus, the simple zeroing methodology employed by USDOC is inconsistent with Article 2.4.2 because instead of treating all export transactions as a whole, it treats each transaction individually. The European Communities also refers to its arguments presented in connection with its claims regarding model zeroing.²⁰³

7.114 The European Communities rejects the argument of the United States that since Article 2.4.2 is expressly limited to the investigation phase of an anti-dumping proceeding it does not contain any obligations with respect to administrative reviews. Since the AD Agreement does not contain a definition of the word "investigation" the ordinary meaning of "investigation" in Article 2.4.2 must be determined on the basis of dictionary definitions, which indicate that "investigation" means "a systematic examination or inquiry or a careful study of or research into a particular subject". While the text of a provision may limit the particular subject of such an examination or study, as is the case in Article 5, the meaning of the word "investigation" by itself is not limited to investigations "to determine, the existence, degree and effect of any alleged dumping" referred to in Article 5. In the view of the European Communities, Articles 9.3.1, 9.5, 11.2 and 11.3 of the AD Agreement require an authority to conduct "investigations".²⁰⁴ The European Communities asserts that the AD Agreement generally uses "investigation" in accordance with the dictionary definition but that, having regard to text, context, object and purpose, the word "investigation(s)" in Articles 3.3, 5, 7, 9.5 and 10 of the Agreement has the more limited meaning of "an investigation to determine the existence, degree and effect of any alleged dumping".²⁰⁵

7.115 The European Communities submits various possible alternative interpretations of the phrase "during the investigation phase" in Article 2.4.2 in order to demonstrate that this phrase can be given meaning without interpreting it as restricting the application of the obligations contained in Article 2.4.2 to original investigations. First, the most natural reading of "during the investigation phase", given its placement in the sentence, is that it is associated with "the existence of margins of dumping". Linked to the words "the existence of margins of dumping", the phrase "during the investigation phase" can be interpreted to refer to the investigation period and to prohibit authorities from relying on data arising outside that period. This would apply to any kind of investigation conducted under the Agreement, not just an original investigation.²⁰⁶ Second, if "during the investigation phase" is associated with the word "established", it can be interpreted to refer to the period of time during which the investigating authority must make its determination. This would again apply not only to an original investigation but to any type of investigation under the Agreement. Under this interpretation, the phrase "during the investigation phase" requires authorities to make their determination within the time span identified in the relevant provision.²⁰⁷ Third, "during the investigation phase" can also be interpreted to mean that in the pre-initiation phase applicants are not require to apply the technical rules of Article 2.4.2. Fourth, it is possible to interpret the words "during the investigation phase" as being descriptive, in the same way in which the United States considers the words "the existence, degree and effect of any alleged dumping" in Article 5.1 to be descriptive.

7.116 The European Communities considers that the word "the" in "during the investigation phase" does not support the view of the United States that this phrase refers to investigations within the meaning of Article 2.4.2. If the word "the" suggests something that already "known ... or ... familiar" this would rather support the view that "investigation phase" in Article 2.4.2 means the "period of investigation", an expression frequently used in Article 2.²⁰⁸

7.117 The European Communities rejects the argument of the United States that Article 2.4.2 focuses upon the "existence" of dumping rather than on the "amount" of dumping. In an original investigation, it is impossible to determine the existence of dumping without calculating the degree of dumping. In an assessment proceeding, in calculating the degree of dumping, an authority is necessarily also determining whether dumping existed during the period under review. A margin of dumping can only exist if it has a defined magnitude. The word "existence" is used in various provisions of the AD Agreement without that being of any legal consequence. The European Communities also draws attention to the fact that the United States does not apply certain provisions of Article 2.4.1 to administrative reviews, notwithstanding that Article 2.4.1 does not include the word "existence" of dumping. The fact that the French and Spanish titles of Article 2 of the AD Agreement refer to the "existence" of dumping shows that Article 2 as a whole addresses the "existence" of dumping and that the connection between "existence" and dumping in Article 2.4.2 is not unique.²⁰⁹

7.118 Regarding the word "phase" in "during the investigation phase", the European Communities considers that while there can be phases other than an investigation phase, such as a pre-investigation phase or post-investigation phase, this applies to any type of investigation, original or otherwise. Therefore, the word "phase" in Article 2.4.2 does not support the conclusion that the word "investigation" is defined in Article 5.1 for the purposes of the entire AD Agreement, including Article 2.4.2. In any event, the fact that the United States has excluded certain provisions of Article 2.4.1 from administrative reviews because of the phrase "in an investigation" shows that the word "phase" in Article 2.4.2 cannot explain the position of the United States that Article 2.4.2 does not apply to such reviews. Panel and Appellate Body reports have used the word "phase" in different ways indicating that the word "phase" in Article 2.4.2 does not have the special or limited or defined meaning suggested by the United States.²¹⁰

7.119 The European Communities rejects as irrelevant the reference made by the United States to Article 9.4(ii) of the AD Agreement because none of the measures at issue in this dispute involved the application of a prospective normal value system within the meaning of that provision and because a prospective normal value system is in any event subject to the refund provisions of Article 9.3.2, which in turn is subject to Article 2, including Article 2.4.2.²¹¹

7.120 The European Communities asserts that the United States has not been able to substantiate its view that the word "investigation" in Article 2.4.2 has a limited meaning and that the United States ignores the plain language of Article 2.4.2 by not interpreting the word "investigation" in its ordinary meaning. The interpretation of "investigation" advanced by the United States in this dispute is contradicted by the fact that in the retrospective assessment proceedings at issue in this case the USDOC used the term "investigation" and by the use of the terms "Five-Year Review(Sunset) Investigations" and "review investigation" by the USITC.²¹² The interpretation advanced by the United States is also inconsistent with the Vienna Convention because it ignores the relevant context, such as Articles 2.2 and 6, and the object and purpose of the AD Agreement and of retrospective assessment proceedings.²¹³

7.121 The European Communities finds support for its interpretation of "investigation" in panel and Appellate Body reports that characterize investigations under Article 5.1 as "original investigations" and in panel reports that have used the words "investigated" and "investigation" in connection with annual administrative reviews and sunset reviews carried out by the United States.²¹⁴

7.122 The European Communities rejects what it characterizes as a monolithic conception of the term "investigation" according to which an investigation exists only when a careful examination is undertaken of all the three matters referred to in Article 5.1 (existence, degree and effect of any alleged dumping). With reference to Article 3, the European Communities argues that a careful and systematic examination of injury (the effect of any alleged dumping) is nevertheless an "investigation". Thus, it is the nature of the activity carried out by an authority, not the scope of the inquiry, that determines whether that activity is an investigation. Consequently, an assessment proceeding under Article 9.3.1, which involves a careful and systematic examination of the "degree" of dumping is an "investigation".²¹⁵ The European Communities finds further support for its view that an investigation under the AD Agreement is not necessarily an investigation to determine the existence degree and effect of any alleged dumping in Article 5.8 and in the provisions on on-the-spot investigations in Annex I.²¹⁶

7.123 The European Communities refers to the negotiating history of the AD Agreement as support for its position that the applicability of Article 2.4.2 is not limited to investigations within the meaning of Article 5 of the AD Agreement.²¹⁷ The European Communities asserts that there is "subsequent practice" within the meaning of Article 31(3)(b) of the Vienna Convention that supports its interpretation of Article 2.4.2 because a review of domestic anti-dumping legislation notified by 105 Members indicates that no other Member has adopted the position that Article 2.4.2 is limited to original investigations.²¹⁸

7.124 The European Communities submits that the term "margin of dumping" is defined in Article VI:2 of the GATT 1994, which definition is implemented and further elaborated in Article 2. The same defined term "margin of dumping" is also used in Articles 9.3, which cross-refers to all of Article 2, and in Articles 9.1 and 9.5. Article 2.4.2 states what it applies to, not what it does not apply to. Article 2.4.2 unlike the SAA, does not contain the words "(not reviews)". The United States is therefore using an a contrario reasoning in an attempt to defeat the definition of "margin of dumping" and the cross-reference to Article 2 in Article 9.3. If Members had really wanted to achieve what the United States asserts, it would have been a simple matter for them to have used express exclusionary language.²¹⁹

7.125 The European Communities submits that not applying Article 2.4.2 to duty assessment under Article 9.3.1 would entail unequal treatment between prospective and retrospective duty assessment systems. In a prospective system, the amount of anti-dumping duty collected is based on the margin of dumping established in the original investigation in which context zeroing is prohibited. By contrast, in a retrospective assessment system the result of the initial investigation, in which dumping is established without zeroing, would be eclipsed by the use of the zeroing methodology to determine the final amount of anti-dumping duty in a periodic administrative review. In the case of the United States, the results of such a review would apply retrospectively from the date provisional duties were first applied. The European Communities points out that this will necessarily imply that under the retrospective assessment method a higher amount of anti-dumping duties will be collected than under the prospective system and that, while in the prospective system it will be sufficient for the exporter to increase the export price by the margin of dumping to obtain a refund of all duties paid, in the retrospective system such an increase will not suffice to remove duty liability. The European Communities submits that retrospective and prospective duty assessment systems must lead to the same level of protection against dumped imports, which must correspond to the actual margin of dumping of the exporter.²²⁰

7.126 The European Communities points out that acceptance of its claims would mean that there would be only one margin of dumping, for each exporter, in all cases of retrospective duty assessment. Thus the practice of considering each export transaction would be prohibited at the final assessment stage when authorities must ensure that the total amount of duty collected with respect to exports made by one specific exporter during the assessment period does not exceed the relevant margin of dumping of the same exporter.²²¹

7.127 The United States argues that the text of the AD Agreement expressly limits the obligations of Article 2.4.2 to the investigation phase of an anti-dumping proceeding. Investigations and assessment proceedings constitute distinct phases of an AD proceeding and have different purposes. The distinction made between investigations and reviews in Article 18.3 of the AD Agreement mirrors the distinction between these terms in the substantive provisions of the Agreement. Panel and Appellate Body reports have recognized this distinction, consistently finding that provisions in the AD Agreement with express limitations to investigations are in fact limited to the investigation phase of a proceeding.²²² The argument of the European Communities that Article 2.4.2 applies to assessment proceedings under Article 9 ignores the clear distinctions made in the text of the AD Agreement between investigations and other proceedings. The Appellate Body has recognized, in US � Corrosion Resistant Steel Sunset Review, that investigations and other proceedings serve different purposes and have different functions and are therefore subject to different obligations. Article 9 assessment proceedings are not concerned with the question of whether injurious dumping "exists" above a de minimis level such that the imposition of anti-dumping measures is warranted but with the amount of duty to be assessed on particular entries, which is a separate exercise from the calculation of an overall margin of dumping during the threshold investigation phase of an anti-dumping proceeding.²²³

7.128 The United States submits that it is clear from the phrase "the existence of margins of dumping during the investigation phase" that Article 2.4.2 only applies in an Article 5 investigation. Articles 1 and 5.1 of the AD Agreement provide for the existence of a discrete investigation phase. The first sentence of Article 1 cross-references Article 5 to define "investigations initiated" under the AD Agreement. Only investigations under Article 5 are initiated under the AD Agreement. Article 5.1, in turn defines the scope of such investigations as "to determine the existence, degree and effect of any alleged dumping". An investigation under Article 5 is the only investigation phase under the AD Agreement that requires a determination of the "existence" of dumping. Article 2.4.2 establishes how the "existence" of dumping is to be determined. Thus Article 2.4.2 is expressly limited to an Article 5 investigation in that it refers to the "investigation phase" and in that it provides that its purpose is to establish the existence of dumping. The Appellate Body in EC � Bed Linen and the panel in Argentina � Poultry Anti-Dumping Duties have confirmed that the application of Article 2.4.2 is limited to the investigation phase.²²⁴

7.129 The United States argues that Article 9 does not incorporate the requirements of Article 2.4.2 of the AD Agreement. The general reference in Article 9.3 to Article 2 necessarily includes any limitations found in the text of Article 2. Therefore, the explicit limitation of Article 2.4.2 to the investigation phase means that Article 2.4.2 does not apply to Article 9.3.²²⁵

7.130 The United States submits that the limited application of Article 2.4.2 to the investigation phase is also consistent with the fact that the AD Agreement allows for the use of different assessment systems. Application of Article 2.4.2 to assessment proceedings would render this divergence of assessment systems impossible.²²⁶ Article 9.4(ii) explicitly allows for the calculation of anti-dumping duties on the basis of a comparison of a prospective normal value and individual export prices.²²⁷ There is no basis in the AD Agreement for a requirement that Members with prospective normal value systems provide credits for non-dumped entries when assessing duties on subsequent entries that are dumped.²²⁸ Prospective normal value systems inherently operate on an entry-by-entry basis. The assessment system used by the United States functions on a retrospective basis but in substance it operates much like a prospective normal value system, albeit with contemporaneous normal values.²²⁹

7.131 The United States considers that the concern of the European Communities that limiting the application of Article 2.4.2 to the investigation phase would place retrospective anti-dumping systems at a relative disadvantage as compared to prospective systems is misplaced because Article 9.3 places prospective and retrospective systems on the same footing. Under both systems, the AD Agreement permits the investigating authority to attach liability for antidumping duties on imports as they cross the border. The AD Agreement then provides for the Member, no matter what the system, to determine whether any refund is due. A prospective system does not inherently require the application of Article 2.4.2 to refund proceedings. Nothing in Article 9.3 requires Members to conduct such assessment proceedings so as to cover all imports from a particular exporter over any period of time. The AD Agreement permits Members to apply their duty assessment systems so as to focus any Article 9.3 proceedings on particular imports, particular importers or particular exporters as they deem most appropriate.²³⁰ The United States further asserts that the spectre of unequal treatment conjured up by the European Communities has no legal or factual basis because, first, the argument of the European Communities erroneously assumes that Article 5 investigation phase margins are the basis for duty collections in all prospective systems. In a prospective normal value system duties need not be limited by the margins of dumping calculated during the Article 5 investigation phase. Second, even Members with prospective ad valorem systems may conduct Article 11.2 reviews at any time to update those margins so that the amount of duty collected corresponds to the actual margin of dumping of the exporter. Article 2.4.2 is no more applicable to such Article 11.2 reviews than it is to Article 9.3 assessment proceedings.²³¹

7.132 The United States submits that while investigations determine whether anti-dumping measures may be applied by determining the existence, degree and effect of any alleged dumping, the collection and assessment of anti-dumping duties is a separate and distinct phase that occurs after an anti-dumping measure is imposed. Because importers will incur liability for duties, it is appropriate to determine that liability on an importer- and transaction-specific basis.²³² The United States further argues that the European Communities proposes an approach that divorces the amount of anti-dumping duty assessed with respect to an import from the dumping margin associated with that import transaction. The argument of the European Communities ignores the distinction between the Article 5 investigation phase referred to in Article 2.4.2 with its focus on the existence of margins of dumping and the assessment proceeding under Article 9.3.1 with its focus on duty liability, and is inconsistent with the Appellate Body finding in EC � Bed Linen. Moreover, there is no support in the text or context of Article 9.3.1 for the proposition that a Member must apply an assessment methodology whereby importers that pay relatively high non-dumped export prices could also be liable for anti-dumping duties because of low, dumped prices paid by other, unrelated importers.²³³ There is a basic logic and fairness that supports importer-specific assessment in that an importer will pay no anti-dumping duties when none of its imports are dumped. By contrast, the exporter-oriented assessment process required by the European Communities' interpretation would under any of the margin methodologies, even without, zeroing require some assessment of anti-dumping duties on the non-dumped imports of such an importer.²³⁴

(b) Arguments of third parties

7.133 Argentina submits that Article 2.4.2 of the AD Agreement does not apply outside the investigation phase and therefore disagrees with the assertion of the European Communities that insofar as the United States uses a retrospective duty assessment system for calculating anti-dumping duties to be collected, the authorities should calculate those duties in a manner consistent with Article 2.4.2. Article 2.4.2 is expressly limited to the investigation phase. The AD Agreement makes a clear distinction between the investigation phase and the imposition and collection of anti-dumping duties. Therefore, the requirement in Article 9.3 that anti-dumping duties shall not exceed the margin of dumping established pursuant to Article 2 does not mean that authorities conduct an on the spot recalculation of the margin of dumping in accordance with Article 2.4.2. As stated by the Panel in Argentina � Poultry Anti-Dumping Duties, Article 9.3 refers to the margin of dumping established in accordance with Article 2 as a whole.

7.134 Brazil submits that while it agrees with the positions taken by the European Communities in this dispute, it is largely academic to discuss whether Article 2.4.2 applies to the reviews, or whether reviews are considered to be part of investigations or whether average-to-transaction comparisons are permitted because zeroing is inconsistent with the "fair comparison" requirement of Article 2.4, which is not limited to certain types of proceedings or certain types of comparisons.

7.135 China submits that "during the investigation phase" does not limit Article 2.4.2 to original investigations because there exists no definition of the word "investigation" and there is no precedent in WTO dispute settlement practice for distinguishing between assessment proceedings and original investigations. While there may be differences between reviews and original investigations, assessment proceedings are closely related to original investigations in that they both pertain to the imposition of anti-dumping duties. Moreover, Article 2 of the AD Agreement applies for the entire purpose of the Agreement and Article 9.3 specifically refers to "the margin of dumping as established under Article 2", which includes Articles 2.4 and 2.4.2.

7.136 Hong Kong, China submits that Article 18.3 of the AD Agreement and the panel and Appellate Body reports cited by the United States do not provide any guidance with respect to how the term "investigation" is generally used in the AD Agreement. The word "investigation" in the AD Agreement does not necessarily mean original or initial investigation. In the context of Article 2.4.2, "investigation" and "investigation phase" should be interpreted to refer to any procedure undertaken by an investigating authority which conforms to the ordinary meaning of an "investigation" ("action or process of investigating", i.e. search[ing] or inquir[ing] into; examining (a matter) systematically or in detail") and which leads to the establishment of the existence of margins of dumping for the subject product. The steps taken by USDOC in a retrospective duty assessment proceeding conform to this meaning of "investigation". Article 9.3 is further support for the proposition that Article 2.4.2 applies to the calculation of margins of dumping in duty assessment proceedings. Hong Kong, China rejects the argument of the United States that Article 9 4(ii) of the AD Agreement provides for an average-to-transaction comparison method. Article 9.4 is applicable only in the special situation where authorities have used sampling to limit their examination and does not permit the calculation of a margin of dumping on the basis of an average-to-transaction comparison as such. Instead, it refers to the use of variable duties. In any case, the United States does not use the prospective normal value system envisaged in Article 9.4 (ii).

7.137 Japan submits that the use by USDOC of simple zeroing in duty assessment proceedings is inconsistent with Articles 2.4, 2.4.2 and 9.3 of the AD Agreement. The ruling of the Appellate Body in US � Corrosion-Resistant Steel Sunset Review that Article 2 of the AD Agreement applies to dumping margins used in a sunset review under Article 11.3 must apply a fortiori to dumping margins calculated and used in periodic reviews under Article 9.3, which specifically refers to Article 2. It follows from this ruling that the "fair comparison" requirements in Article 2.4 apply equally to a dumping margin calculated or used for purposes of Article 9.3. The requirement in Article 2.4.2 for the comparison of normal value and export price to include "all comparable export transactions" is one of such "fair comparison" requirements. In any event, leaving aside whether Article 2.4.2 applies to margins of dumping calculated pursuant to Article 9.3, the use of simple zeroing in periodic reviews is inconsistent with Article 2.4, which plainly does apply.

7.138 Korea argues that a periodic review under Article 9.3 is part of the investigation phase under Article 2.4.2. The Appellate Body stated in US � Corrosion-Resistant Steel Sunset Review that reviews under Article 11 of the AD Agreement involve both investigatory and adjudicatory aspects and therefore concluded that the prohibition of zeroing implicit in Article 2.4.2 also applied to dumping margin calculations under Article 11.3. This suggests that "investigation phase" is properly understood in the context of Article 2.4.2 to mean the portion of the proceeding (original investigation or review) in which an authority "investigates" whether dumping has occurred. Korea submits that the same logic should apply to duty assessment under Article 9.3 of the AD Agreement. The reference in Article 9.3 to "the margin of dumping as established under Article 2" explicitly indicates that the requirements of Articles 2.4 and 2.4.2 must be applied to periodic reviews under Article 9.3.

7.139 Mexico submits that the use of simple zeroing in duty assessment proceedings is inconsistent with Article 2.4.2 of the AD Agreement because it fails to take into account all sales of the product under investigation. The argument that Article 2.4.2 only applies to original investigations is untenable because Article 9.3 refers to, and incorporates, Article 2 as a whole without limitation and therefore requires authorities to apply Articles 2.4 and 2.4.2 in assessment proceedings. In addition, the definition of dumping in Article 2.1 applies throughout the AD Agreement. Article 9.4 does not support the position of the United States because it applies only where authorities use sampling to limit the scope of their examination and where liability for payment of anti-dumping duties is calculated on a prospective basis, which is not the case of the United States.

7.140 Norway submits that the argument of the United States that the word "investigation" in Article 2.4.2 prevents application of Article 2.4.2 to periodic reviews and new shipper reviews is without merit because Article 9.3 explicitly refers to the calculation of dumping margin, an issue that falls squarely within the scope of Article 2.4.2. Second, the Appellate Body has based its conclusion that zeroing is prohibited also on Article 2.4, which includes Article 2.4.2. Third, the steps taken by USDOC in duty assessment proceedings and new shipper reviews effectively amount to a new investigation.

7.141 Turkey submits that Article 2.4.2 does not prohibit simple zeroing in all circumstances but that it may only be applied in connection with the exceptional, average-to-transaction method provided for in the second sentence of Article 2.4.2. By systematically applying in the periodic reviews at issue in this dispute a method that is permitted only in exceptional circumstances, the United States has violated Article 2.4.2. Turkey disagrees with the United States that Article 2.4.2 is not applicable to the duty assessment phase and refers in this regard to the Appellate Body ruling in US � Corrosion-Resistant Steel Sunset Review that whenever a Member calculates a margin of dumping it must abide by Article 2, and 2.4 in particular.

(c) Evaluation by the Panel

7.142 The measures at issue²³⁵ are certain administrative reviews²³⁶ of anti-dumping duty orders, in which USDOC determined percentage weighted average dumping margins and cash deposit rates for individual exporters/producers and assessment rates for individual importers.²³⁷

7.143 It is not disputed that these administrative reviews of existing anti-dumping duty orders fall within the scope of Article 9 of the AD Agreement ("Imposition and Collection of Anti-Dumping Duties") which provides:

"9.3 The amount of the anti-dumping duty shall not exceed the margin of dumping as established under Article 2.

9.3.1 When the amount of the anti-dumping duty is assessed on a retrospective basis, the determination of the final liability for payment of anti-dumping duties shall take place as soon as possible, normally within 12 months, and in no case more than 18 months, after the date on which a request for a final assessment of the amount of the anti-dumping duty has been made. Any refund shall be made promptly and normally in not more than 90 days following the determination of final liability made pursuant to this sub‑paragraph. In any case, where a refund is not made within 90 days, the authorities shall provide an explanation if so requested.

9.3.2 When the amount of the anti-dumping duty is assessed on a prospective basis, provision shall be made for a prompt refund, upon request, of any duty paid in excess of the margin of dumping. A refund of any such duty paid in excess of the actual margin of dumping shall normally take place within 12 months, and in no case more than 18 months, after the date on which a request for a refund, duly supported by evidence, has been made by an importer of the product subject to the anti-dumping duty. The refund authorized should normally be made within 90 days of the above‑noted decision.

9.3.3 In determining whether and to what extent a reimbursement should be made when the export price is constructed in accordance with paragraph 3 of Article 2, authorities should take account of any change in normal value, any change in costs incurred between importation and resale, and any movement in the resale price which is duly reflected in subsequent selling prices, and should calculate the export price with no deduction for the amount of anti-dumping duties paid when conclusive evidence of the above is provided."

Sub-paragraph 3.1 specifically addresses retrospective duty assessment procedures such as the administrative reviews at issue in this dispute.

7.144 In these administrative reviews, USDOC determined margins of dumping on the basis of "asymmetrical" comparisons between (monthly) weighted-average normal values and individual export transactions without regard to whether the conditions expressly set out in the second sentence of Article 2.4.2 of the AD Agreement were fulfilled. In aggregating the results of these comparisons to calculate an overall margin of dumping for each exporter, USDOC did not include in the numerator of that overall margin any amounts by which export prices exceeded normal value.²³⁸ The European Communities asserts that in these respects USDOC acted inconsistently with Article 2.4.2 of the AD Agreement. The United States submits that Article 2.4.2 does not apply in the context of these administrative reviews.

7.145 Therefore, the threshold legal question before us is whether the obligations contained in Article 2.4.2 apply to duty assessment proceedings provided for in Article 9.3.

7.146 As noted above, Article 9.3 provides that "[t]he amount of the anti-dumping duty shall not exceed the margin of dumping as established under Article 2."²³⁹ Thus, it is clear that as a general matter "Article 2" is relevant to duty assessment procedures²⁴⁰, and there is nothing in Article 9.3 itself that suggests that Article 2.4.2 specifically does not apply in the context of Article 9.3. That, however, is not the end of the matter. If Article 2 itself provides that Article 2.4.2 does not apply in the case of reviews under Article 9.3, that is not overridden by the fact that "Article 2" is specifically referred to in Article 9.3. Absent anything explicitly to the contrary, that reference to "Article 2" in Article 9.3 must be read as including any limitation that is expressed in Article 2 itself. We commence therefore with Article 2.4.2.

7.147 Article 2.4.2 of the AD Agreement provides that:

"Subject to the provisions governing fair comparison in paragraph 4, the existence of margins of dumping during the investigation phase shall normally be established on the basis of a comparison of a weighted average normal value with a weighted average of prices of all comparable export transactions or by a comparison of normal value and export prices on a transaction‑to‑transaction basis. A normal value established on a weighted average basis may be compared to prices of individual export transactions if the authorities find a pattern of export prices which differ significantly among different purchasers, regions or time periods, and if an explanation is provided as to why such differences cannot be taken into account appropriately by the use of a weighted average‑to‑weighted average or transaction‑to‑transaction comparison." (emphasis added).

7.148 The interpretive question raised by the disagreement between the parties is whether, as argued by the European Communities, Article 2.4.2 applies to any proceeding under the AD Agreement in which authorities determine margins of dumping or whether, as argued by the United States, Article 2.4.2 only applies to determinations of dumping in the context of investigations within the meaning of Article 5 of the AD Agreement.²⁴¹

7.149 In analyzing this issue we examine successively: (i) the interpretation of the phrase "the existence of margins of dumping during the investigation phase" in context; (ii) relevant panel and Appellate Body reports; (iii) the alternative interpretations of "during the investigation phase" suggested by the European Communities; (iv) arguments of the European Communities regarding object and purpose of Article 9.3 and the AD Agreement; (v) arguments of the European Communities regarding subsequent practice; and (vi) arguments of the European Communities regarding supplementary means of treaty interpretation.

(i) Interpretation of "the existence of margins of dumping during the investigation phase" in context

7.150 The disciplines of Article 2.4.2 apply in the context of the establishment of "the existence of margins of dumping during the investigation phase". We therefore consider that our analysis should focus on the meaning of the expression "the existence of margins of dumping during the investigation phase..." as a whole and not on the word "investigation" taken in isolation. Consistent with basic principles of treaty interpretation²⁴², we must proceed on the basis that this phrase � which is unique in Article 2 and in fact in the AD Agreement as a whole � has a meaning and function. Our task is to determine whether, as argued by the United States, that function is to limit the application of Article 2.4.2 to the "investigation phase" in the sense of the "original investigation" as opposed to subsequent phases of an anti-dumping proceeding such as duty assessment. Since the relevant "context" for purposes of treaty interpretation in principle comprises the text of the entire treaty, the requirement to interpret the terms of the treaty "in their context", means that we must not only consider the word "investigation" as part of the phrase "the existence of margins of dumping during the investigation phase" but that we must also consider whether other provisions of the AD Agreement illuminate the meaning of the phrase "the existence of margins of dumping during the investigation phase" in Article 2.4.2.

7.151 In light of these considerations, we disagree with the view of the European Communities that the decisive element regarding the interpretation of the scope of Article 2.4.2 is the word "investigation� which has not been defined in the AD Agreement and which must therefore be interpreted strictly by reference to a dictionary definition.²⁴³ Rather, we must determine the meaning of the entire phrase "the existence of margins of dumping during the investigation phase" in the context of the AD Agreement as a whole. The dictionary definition of "investigation" is only one element in this analysis and is not necessarily dispositive.²⁴⁴

7.152 In interpreting the phrase "the existence of margins of dumping during the investigation phase" in context we consider the following textual and contextual elements, taken together, to be particularly significant: (a) the reference to "the investigation phase"; (b) the textual similarity between the phrase "the existence of margins of dumping during the investigation phase" and the language of Article 5.1 of the AD Agreement; (c) the fact that where the terms "investigation" and "investigations" are used in the AD Agreement to refer to particular proceedings, it is in most cases clear from the substantive context that such proceedings are investigations within the meaning of Article 5 and that the AD Agreement does not use "investigation" and "investigations" in relation to proceedings that take place once an anti-dumping measure has been taken; (d) the fact that certain provisions governing investigations apply to provisions governing different proceedings only by virtue of express cross-references; and (e) the express distinction between "investigations" and "reviews" in Article 18 of the AD Agreement.

- Investigation phase

7.153 First, we consider that a key element in resolving the issue of the scope of Article 2.4.2 is the use of the phrase "the investigation phase". This phrase, in our view, reflects the notion that there exists in the AD Agreement a particular phase, characterized by the word "investigation" used as an adjective, that is distinct from other "phases". The yoking of these two words together must be given effect. It would be inherently flawed to act as if this word "phase" did not exist. The meaning of the term "investigation" in any abstract or general sense is not the issue. Rather the issue is the phrase "investigation phase", to wit, a determinate stage in a process of longer duration and can only be interpreted in that way. Therefore, what we must be concerned with is investigation as a particular stage in a process. We note in this connection that, as held by the Appellate Body in US � Softwood Lumber V, the phrase "during the investigation phase" operates to limit the applicability of Article 2.4.2.²⁴⁵

7.154 In considering how to distinguish this "investigation phase" from other phases, we see significance in the fact that Article 5 ("Initiation and Subsequent Investigation") is the only provision in the entire AD Agreement that uses the word "investigation" to define its subject matter. As the provision that is the most specific in the AD Agreement with respect to the concept of "investigation", it would be most natural in the context to read "the investigation phase" in Article 2.4.2 as referring to the concept of "investigation" as used in Article 5. To interpret "the investigation phase" as corresponding to an investigation within the meaning of Article 5 makes it possible to maintain a meaningful distinction within the AD Agreement between "the investigation phase" and subsequent phases: the "investigation phase" referred to in Article 2.4.2 is the phase that begins when an investigation is initiated under Article 5 and that ends when that investigation is concluded, which, under Article 5.10, must occur within one year, and in no case more than 18 months after the initiation of the investigation. This interpretation of "the investigation phase" in Article 2.4.2 is, moreover, consistent with the distinction drawn by the Appellate Body between "investigations" and other phases of countervailing duty and anti-dumping duty proceeding subsequent to the application of countervailing or anti-dumping measures.²⁴⁶

7.155 In contrast, according to the interpretation advanced by the European Communities, "the investigation phase" within the meaning of Article 2.4.2 would effectively include any kind of "systematic examination or inquiry or careful study", be it in the context of an investigation under Article 5, an assessment proceeding under Article 9.3, a review under Article 9.5 or a review under Articles 11.2 and 11.3. But this has the inevitable consequence of eroding any meaningful distinction between "the investigation phase" and subsequent phases. In fact, by the approach of the European Communities each and every phase of an anti-dumping proceeding is "the investigation phase". This would effectively read the word "phase" out of the text.

- Textual similarity between Article 2.4.2 and Article 5.1 of the AD Agreement

7.156 Second, we note the textual similarity between establishing "the existence of margins of dumping during the investigation phase" in Article 2.4.2 and "an investigation to determine the existence of ...any alleged dumping" in Article 5.1 While there are various provisions of the AD Agreement requiring authorities to determine a margin of dumping, it is only in Article 5 that the "existence... of dumping" must be determined in the context of an "investigation". This textual similarity between "the existence of margins of dumping during the investigation phase" in Article 2.4.2 and the language used in Article 5 of the AD Agreement stands in clear contrast to the terminology used in other provisions of the AD Agreement that may require authorities to determine margins of dumping and which the European Communities considers to be "investigations� based on the dictionary definition of the meaning of "investigation". The AD Agreement characterizes the nature and purpose of proceedings under Article 9 and 11 in terms that are very different from establishing "the existence of margins dumping during the investigation phase".²⁴⁷

7.157 The European Communities has submitted various arguments in support of its view that the word "existence" in the first sentence of Article 2.4.2 cannot support the interpretation of Article 2.4.2 as not applicable to administrative reviews.²⁴⁸ We recall our view that the issue before us is the meaning of "the existence of margins of dumping during the investigation phase" in its entirety rather than the individual words used in this phrase isolated from their context. Contrary to what the European Communities suggests, proceedings other than investigations under Article 5 are not concerned with determining "the existence of margins of dumping during the investigation phase". In this respect, whether or not it is possible as an empirical matter to determine the "existence" of dumping without at the same time determining its "magnitude", this is not the legally relevant issue. Rather, what matters is that the AD Agreement does not express textually the nature and purpose of proceedings under provisions other than Article 5 in terms of establishing "the existence of margins of dumping during the investigation phase".²⁴⁹ Similarly, while it is true that the word "existence" in and of itself is not unique and is used in other provisions of the AD Agreement as well as in the SCM Agreement, none of the provisions cited by the European Communities²⁵⁰ refers to the establishment of the existence of dumping (or a subsidy) in the context of an investigation.²⁵¹

- Use of the term "investigation" in the AD Agreement to refer to (phases of) proceedings

7.158 Third, in order to interpret the phrase "the existence of margins of dumping during the investigation phase" in Article 2.4.2, it is also helpful to examine the general pattern of the usage of the term "investigation" in the AD Agreement.²⁵² That term is used both: (i) in a general sense (on-the-spot investigation, period of investigation); and (ii) to refer to a proceeding or phase of proceeding.²⁵³ Since in the context of Article 2.4.2 the word "investigation" in "the investigation phase" is used to refer to a particular phase of an anti-dumping proceeding, its use in the general sense is not germane to the issue at hand. If, however, an examination of the use of the term to refer to a (phase of a) proceeding consistently refers to Article 5 proceedings, that fact would strongly buttress the interpretation of "investigation phase" that we reached above. It is instructive in this respect to review the reference made to "investigations" in Article 1 of the AD Agreement as a prerequisite for the application of an anti-dumping measure and the usage of "investigation" in the provisions of Articles 3, 6, 7, 8 and 10 and 12 on evidence, provisional measures, price undertakings, retroactive application of anti-dumping measures and public notice and explanation of determinations.

7.159 Article 1 of the AD Agreement ("Principles") provides that the initiation and conduct of "investigations" in accordance with the Agreement is one of the two fundamental prerequisites for the application of an anti-dumping measure. The fact that anti-dumping measures may only be applied "pursuant to" "investigations" and the express link made to Article 5 in footnote 1 indicate that "investigations" within the meaning of Article 1 necessarily precede the application of an anti-dumping measure. This is also apparent from the fact that an "investigation" within the meaning of Article 1 is initiated by the procedural action by which a Member formally commences an investigation as provided in Article 5.²⁵⁴

7.160 With respect to Article 3 of the AD Agreement ("Determination of Injury"), we note that the Appellate Body has affirmed in US � Oil Country Tubular Goods Sunset Reviews that "anti-dumping investigations" in the first sentence of Article 3.3 refers to investigations within the meaning of Article 5 and thus not to sunset reviews under Article 11 of the AD Agreement.²⁵⁵ The European Communities does not contest that it refers to "investigations" within the meaning of Article 5.

7.161 The European Communities argues that Article 6 of the AD Agreement ("Evidence") employs the word "investigation" in a generic sense that corresponds to its dictionary definition and that this is important contextual support for the view that "investigation" in "investigation phase" has a generic meaning. We disagree. First, Article 6 begins with the phrase "[a]ll interested parties in an anti-dumping investigation....". The fact that Article 6 immediately follows the provisions on "Initiation and Subsequent Investigation" in Article 5 would more naturally imply that "an anti-dumping investigation" in the first sentence of Article 6.1 has the same meaning as the word "investigation" in Article 5. Second, Article 6.1.3 provides that as soon as "an investigation" has been initiated, the authorities shall provide the text of the application received under Article 5.1 to the known exporters, the authorities of the exporting Member and, upon request, to other interested parties. The direct link made in this sentence between "an investigation" and the application received under Article 5.1 confirms that "an investigation" within the meaning of this sub-paragraph is an investigation initiated under Article 5 i.e. an investigation "to determine the existence, degree and effect of any alleged dumping". This link between "anti-dumping investigation" and "investigation" in Article 6.1 and the word "investigation" as used in Article 5 logically implies that references to "the anti-dumping investigation", "an anti-dumping investigation" or "an investigation" in other provisions of Article 6 must be construed in the same manner. More generally, the fact that Article 6 is placed between an Article on "Initiation and Subsequent Investigation" and an Article on "Provisional Measures" gives further contextual support to the reading that this Article lays down rules applicable to investigations initiated under Article 5 to determine whether or not the conditions for the application of an anti-dumping measure are fulfilled.^{256 257}

7.162 Article 7 of the AD Agreement ("Provisional Measures") expressly refers in sub-paragraph 1(i) to an investigation initiated in accordance with Article 5. Given this express cross-reference to Article 5, the terms "the investigation" in Article 7.1(iii) and 7.3 and "an investigation" in Article 7.4 must also be understood to refer to investigations within the meaning of Article 5.

7.163 In Article 8 of the AD Agreement ("Price Undertakings") the phrase "the investigation of dumping and injury" in paragraph 4, read together with "a negative determination of dumping or injury" in the second sentence and with "an affirmative determination of dumping and injury" in the last sentence clearly means an investigation to determine whether dumping and injury exist i.e. an investigation within the meaning of Article 5.

7.164 In Article 10 of the AD Agreement ("Retroactivity"), the phrases "after initiating an investigation" in Article 10.7 and "the date of initiation of the investigation" read in their context must necessarily be interpreted to refer to investigations within the meaning of Article 5 of the AD Agreement.

7.165 In Article 12 of the AD Agreement ("Public Notice and Explanation of Determinations"), it logically follows from the reference to Article 5 in paragraph 1 that "initiation of an investigation" in Article 12.1.1 and 12.1.1(ii) means initiation of an investigation pursuant to Article 5. The concept of an investigation as the investigatory phase culminating in the decision whether or not to apply an anti-dumping measure is also apparent from the context in which the word "investigation� appears in the first sentence of Article 12.2.2 and in Article 12.2.3.²⁵⁸

7.166 Thus, our review of the use of "investigation" and "investigations" in Articles 1, 3, 6, 7, 8, 10 and 12 of the AD Agreement reveals that, where these words refer to a proceeding or a phase of a proceeding, they are limited to investigations within the meaning of Article 5.1 of the AD Agreement, i.e. investigations that aim to determine whether the conditions for the application of an anti-dumping measures are present. By contrast, the AD Agreement typically does not use the words "investigation" and "investigations" in relation to proceedings that take place after an anti-dumping measure has been imposed. In particular, as discussed above²⁵⁹, there is nothing in the text of Articles 9 and 11 of the AD Agreement to suggest that proceedings under those provisions are "investigations". The AD Agreement does not distinguish between "original investigations" and other types of "investigations" but between "investigations" and other types of "proceedings".

7.167 Given our view that we need to focus on how the word "investigation" is used in the AD Agreement in relation to (phases of) proceedings, we see no merit in the argument of the European Communities that the fact that other provisions in Article 2 use the word "investigation" in the general sense of "a systematic examination or inquiry or a careful study of a particular subject"²⁶⁰ because in these provisions²⁶¹ "investigation" is generally not used to refer to a particular (phase of a) proceeding. We also recall that the issue before us is whether the phrase "the existence of margins of dumping during the investigation phase" as a whole, and not the word "investigation" by itself, limits the scope of Article 2.4.2. That a provision in Article 2 that refers to "period of investigation" or to "exporters or producers under investigation" is not limited in its applicability to investigations within the meaning of Article 5 is of little relevance to whether Article 2.4.2 is so limited as a result of the phrase "the existence of margins of dumping during the investigation phase".

- Cross-references in Articles 11 and 12 of the AD Agreement

7.168 Fourth, the idea that "investigation", when used to refer to a (phase of a) proceeding, is generally used in the AD Agreement in relation to the formal process of inquiry that precedes the possible application of an anti-dumping measure finds further support in the fact that certain provisions governing proceedings that take place after the imposition of an anti-dumping measure contain express cross-references to provisions applicable to investigations. Article 11.4 provides that "[t]he provisions of Article 6 regarding evidence and procedure shall apply to any review carried out under this Article". Article 12.3 states that "the provisions of this Article shall apply mutatis mutandis to the initiation and completion of reviews pursuant to Article 11 ..." The need for such cross-references can only be explained by the fact that provisions applicable to "investigations" (Articles 6 and 12.1-2) are not automatically applicable to provisions dealing with other types of proceedings such as duty assessment proceedings and reviews.²⁶² That such provisions applicable to investigations do not ipso facto also apply to other kinds of proceedings confirms that the concept of "investigation" as used in the AD Agreement in relation to proceedings is limited to particular kinds of proceedings and that the mere fact that an authority engages in an intellectual activity that can be described as "investigation" within the dictionary meaning of that term is not determinative.

- Distinction between investigations and reviews in Article 18 of the AD Agreement

7.169 Finally, we note, in this respect, the distinction drawn in Article 18.3 of the AD Agreement between "investigations" and "reviews of existing measures". The juxtaposition of these two concepts in the same sentence makes it unambiguously clear that the drafters of the AD Agreement were aware of the difference between these two concepts. The concept of a "review investigation" used by the European Communities²⁶³ is in direct contradiction with this distinction and has no textual basis in the AD Agreement.

7.170 Based on our analysis of the textual and contextual elements discussed in the preceding paragraphs � the use of "investigation phase", the textual similarity between "the existence of margins of dumping during the investigation phase" in Article 2.4.2 and the wording in Article 5.1; the fact that the AD Agreement consistently uses the word "investigation" in relation to proceedings under Article 5 and uses different terminology in relation to proceedings under Articles 9 and 11; the express cross-references in Articles 11 and 12; and the express distinction between investigations and reviews in Article 18 � we conclude that Article 2.4.2 of the AD Agreement must be interpreted to apply only to determinations of dumping in the context of investigations pursuant to Article 5 of the AD Agreement.

(ii) Panel and Appellate Body Reports

7.171 Panel and Appellate Body reports support our interpretation that "the existence of margins of dumping during the investigation phase" in Article 2.4.2 restricts the applicability of that provision to investigations within the meaning of Article 5 of the AD Agreement. First, the Appellate Body has confirmed that the SCM Agreement and the AD Agreement distinguish between investigations and other "phases" or "stages" of countervailing duty and anti-dumping proceedings, such as duty assessment and review proceedings. Second, the Appellate Body has highlighted the difference in purpose and nature of the determination to be made between "investigations" and other proceedings under the SCM Agreement and the AD Agreement. Third, WTO dispute settlement practice has also made clear that provisions that apply to investigations are not ipso facto applicable to other proceedings under the SCM Agreement and the AD Agreement.

- Investigations and other "phases" or "stages" of countervailing duty and anti-dumping proceedings

7.172 The Appellate Body has made it clear that there are distinct phases or stages discernible in a countervailing duty or anti-dumping action and that the investigation phase, which precedes the imposition of a countervailing or anti-dumping duty, is distinct from phases or stages that follow the imposition of such a duty.

7.173 In US � Carbon Steel, the Appellate Body reversed the finding of the Panel in that case that the de minimis standard in Article 11.9 of the SCM Agreement also applies to sunset reviews under Article 21.3 of that Agreement.²⁶⁴ The Appellate Body noted in this connection that Article 11.9 of the SCM Agreement sets forth a de minimis standard "for original investigations" and that all of its provisions relate to the authorities' initiation and conduct of a countervailing duty investigation.²⁶⁵ According to the Appellate Body:

"Although the terms of Article 11.9 are detailed as regards the obligations imposed on authorities hereunder, none of the words in Article 11.9 suggests that the de minimis standard that it contains is applicable beyond the investigation phase of a countervailing duty proceeding. In particular, Article 11.9 does not refer to Article 21.3, nor to reviews that may follow the imposition of a countervailing duty."²⁶⁶

In this respect, the Appellate Body rejected an argument of Japan that the use of the word "cases" in the second sentence of Article 11.9 means that the de minimis provision must be applied "to all phases of countervailing duty proceedings not only in investigations". According to the Appellate Body, "[t] he use of the word 'cases' does not alter the fact that the terms of Article 11.9 apply the de minimis standard only to the investigation phase."²⁶⁷ The Appellate Body also observed later in its Report that:

"... the non-application of an express de minimis standard at the review stage, and limiting the application of such a standard to the investigation phase alone, does not lead to irrational or absurd results".²⁶⁸

Thus, the Appellate Body clearly used the term "the investigation phase" in the context of a countervailing duty proceeding to refer to an "investigation" within the meaning of Article 11 of the SCM Agreement²⁶⁹, as distinguished from other phases of countervailing duty proceedings, such as "the review stage".

7.174 The notion that the process of inquiry that culminates in the imposition of a countervailing duty or anti-dumping duty is a phase distinct from other proceedings that follow the imposition of such a duty is also apparent from the remark by the Appellate Body in EC � Bed Linen (Article 21.5) that:

"... the imposition and collection of anti-dumping duties under Article 9 is a separate and distinct phase of an anti-dumping action that necessarily occurs after the determination of dumping, injury, and causation under Articles 2 and 3 has been made".²⁷⁰

- The difference in purpose of the determination to be made between "investigations" and other proceedings

7.175 In connection with the temporal distinction between investigations and other phases or stages of countervailing duty and anti-dumping proceedings, the Appellate Body has pointed to the differences in purpose between these distinct phases or stages.

7.176 In US � Carbon Steel, the Appellate Body underlined the fact that under the SCM Agreement:

"...original investigations and sunset reviews are distinct processes with different purposes. The nature of the determination to be made in sunset review differs in certain essential respects from the nature of the determination to be made in an original investigation."²⁷¹

7.177 In US � Corrosion-Resistant Steel Sunset Review, the Appellate Body indicated that this observation also applies to the AD Agreement:

"In an original anti-dumping investigation, investigating authorities must determine whether dumping exists during the period of investigation. In contrast, in a sunset review of an antidumping duty, investigating authorities must determine whether the expiry of the duty that was imposed at the conclusion of an original investigation would be likely to lead to continuation or recurrence of dumping."²⁷²

7.178 In US � Oil Country Tubular Goods Sunset Reviews, the Appellate Body discussed this distinction as follows:

"Original investigations require an investigating authority, in order to impose an anti-dumping duty, to make a determination of the existence of dumping in accordance with Article 2, and subsequently, to determine, in accordance with Article 3, whether the domestic industry is facing injury or a threat thereof at the time of the original investigation. In contrast, Article 11.3 requires an investigating authority, in order to maintain an anti-dumping duty, to review an anti-dumping duty order that has already been established-following the prerequisite determinations of dumping and injury-so as to determine whether that order should be continued or revoked."²⁷³

7.179 In EC - Bed Linen, the Appellate Body distinguished between the determination to be made under Article 2.4.2 and the imposition and collection of duties under Article 9:

"... Article 2.4.2 is not concerned with the collection of anti-dumping duties, but rather with the determination of 'the existence of margins of dumping'. Rules relating to the 'prospective and 'retrospective' collection of anti-dumping duties are set forth in Article 9 of the AD Agreement. The European Communities has not shown how and to what extent these rules on the 'prospective' and 'retrospective' collection of anti-dumping duties bear on the issue of the establishment of 'the existence of dumping margins' under Article 2.4.2."^{274 275}

7.180 We also note in this connection the discussion in the panel reports in US - DRAMS and Mexico � Anti-Dumping Measures on Rice on the difference in purpose between Article 5.8 and Article 9.3 of the AD Agreement.²⁷⁶

- Provisions that apply to investigations are not ipso facto applicable to other proceedings under the SCM Agreement and the AD Agreement.

7.181 The Appellate Body has held that a provision that applies to a countervailing duty or anti-dumping investigation is not automatically applicable to provisions of the SCM Agreement and the AD Agreement that deal with proceedings that take place following the imposition of a countervailing or anti-dumping duty.

7.182 As affirmed by the Appellate Body in US � Carbon Steel and US � Corrosion-Resistant Steel Sunset Review, rules relating to due process aspects of investigations contained in Article 12 of the SCM Agreement and Article 6 of the AD Agreement apply to reviews under Article 21 of the SCM Agreement and Article 11 of the AD Agreement only because they are expressly referred to in those provisions.²⁷⁷

7.183 Similarly, in US � Oil Country Tubular Goods Sunset Reviews, the Appellate Body found that Article 3.3 of the AD Agreement, which allows for a cumulative assessment of the effects of imports where imports from more than one country "are simultaneously subject to anti-dumping investigations", does not apply in the context of sunset reviews under Article 11.3 of the AD Agreement:

"This provision plainly speaks to the situation '[w]here imports of a product from more than one country are simultaneously subject to anti-dumping investigations' (emphasis added) It makes no mention of injury analyses undertaken in any proceeding other than original investigations; nor do we find a cross-reference to Article 11, the provision governing review of anti-dumping duties, which itself makes no reference to cumulation. We therefore find Articles 3.3 and 11.3, on their own, not to be instructive on the question of the permissibility of cumulation in sunset reviews..."²⁷⁸

7.184 The Appellate Body thus made it clear that the expression "anti-dumping investigations" in Article 3.3 by itself has the effect of limiting the applicability of that provision to "original investigations". We note the argument of the European Communities that Article 3.3 has no bearing on the meaning of "investigation" in Article 2.4.2 because the reference in Article 3.3 to Article 5.8 implies that Article 3.3. is limited to investigations within the meaning of Article 5. However, we note that the Appellate Body's conclusion that Article 3.3 makes no mention of injury analyses "in any proceeding other than original investigations" immediately follows its reference to the phrase "anti-dumping investigations". There is nothing in this passage to suggest that the Appellate Body based this conclusion on the fact that Article 3.3 contains a textual link to Article 5. On the contrary, in a subsequent passage the Appellate Body expressly stated that "the opening text of Article 3.3 plainly limits its applicability to original investigations".²⁷⁹

7.185 The European Communities argues that the Appellate Body's use of the term "original investigation" supports its position because the word "original" would be redundant if an investigation were always an investigation within the meaning of Article 5. In our view, however, the above passage indicates that the Appellate Body treats "investigation" and "original investigation" as synonymous. We note that the Appellate Body very frequently refers to "investigation" as used in the SCM Agreement and AD Agreement as "original investigation". In many instances in which the Appellate Body uses "original investigation" in these reports it is clear from the context that the word "original" is used simply for the sake of clarity to highlight the distinction between "investigation" and "review".²⁸⁰ We see nothing in these reports to suggest that the Appellate Body's use of "original" before "investigation" should be interpreted to mean that apart from "original investigations" there are "other types of investigations" under the AD Agreement, such as reviews and duty assessment proceedings.

7.186 To conclude, relevant Appellate Body and panel reports reveal a clear pattern: the concept of an "investigation" in countervailing duty and anti-dumping proceedings, when used to refer to a proceeding or phase of a proceeding, has been consistently distinguished from duty assessment and reviews as a unique phase with a distinct purpose, and, as a consequence, rules applicable to investigations have been found not to be ipso facto applicable to other phases of countervailing duty and anti-dumping proceedings. No Appellate Body or panel report has found that the dictionary definition of the word "investigation" is determinative of the interpretation of that concept in the AD Agreement and that, as a consequence, the concept of "investigation" in the SCM Agreement and AD Agreement applies to any provision, including Articles 19 and 21 of the SCM Agreement and Articles 9 and 11 of the AD Agreement, that can be characterized as contemplating a systematic examination, inquiry or careful study.²⁸¹

7.187 Finally, we note specifically with respect to Article 2.4.2 of the AD Agreement, that the Panel in Argentina � Poultry Anti-Dumping Duties observed:

"We consider that Brazil's principal argument misinterprets the reference to "margin of dumping" in Article 9.3. Based on that language, Brazil focuses entirely on Article 2.4.2, and the reference to the "investigation phase" in that provision. However, Article 9.3 does not refer to the margin of dumping established "under Article 2.4.2", but to the margin of dumping established "under Article 2". In our view, this means simply that, when ensuring that the amount of the duty does not exceed the margin of dumping, a Member should have reference to the methodology set out in Article 2. This is entirely consistent with the introductory clause of Article 2, which sets forth a definition of dumping "for the purpose of this Agreement �". In fact, it would not be possible to establish a margin of dumping without reference to the various elements of Article 2. For example, it would not be possible to establish a margin of dumping without determining normal value, as provided in Article 2.2, or without making relevant adjustments to ensure a fair comparison, as provided in Article 2.4. Thus, the fact that Article 2.4.2, uniquely among the provisions of Article 2, relates to the establishment of the margin of dumping "during the investigation phase" is not determinative of the issue before us, since other provisions of Article 2 do not contain that limitation."²⁸²

7.188 That "during the investigation phase" constitutes a unique limitation of the scope of Article 2.4.2 contradicts the view that "during the investigation phase" has the general meaning that the European Communities ascribes to that phrase.

(iii) The Alternative Interpretations of "during the investigation phase" suggested by the European Communities

7.189 It is of course well accepted in public international law that a treaty must be interpreted in a manner that gives meaning to all its terms. If, as argued by the European Communities the word "investigation" in Article 2.4.2 must be interpreted in line with its dictionary definition, the inclusion of the phrase "the existence of margins of dumping during the investigation phase" means that Article 2.4.2 applies to the establishment of "the existence of margins of dumping during the phase in which authorities carry out a systematic examination or inquiry or a careful study of or research". Such an interpretation of "investigation" as simply referring to the conduct of an examination, inquiry or study would seem to render the phrase "the investigation phase" redundant given that the need to conduct an examination, inquiry or study is already implied by the words "established" in the same sentence of Article 2.4.2 and by the word "determination" in the title of Article 2.

7.190 The European Communities proposes three possible interpretations of "during the investigation phase" in an attempt to demonstrate that this provision can be given meaning without limiting its applicability to investigations within the meaning of Article 5.

7.191 The first of these interpretations is that "during the investigation phase", if associated with "the existence of margins of dumping" rather than with "establishment", means "during the investigation period." Specifically, the European Communities suggests that references to the "period of investigation" in Article 2.2.1, footnote 6 to Article 2.2.1.1 and Article 2.4.1, serve to ensure that investigating authorities rely only on data from the period of investigation, particularly in respect of normal value; the reference in Article 2.4.2 to the establishment of margins of dumping "during the investigation phase" ensures a similar limitation on the use of data from outside the "period of investigation" in respect of the comparison phase generally and in respect of export prices in particular, since there is not otherwise any restriction in regard to them.²⁸³ We note, however, the striking difference in this regard between, on the one hand, the repeated use of "period of investigation" in Articles 2.2.1, 2.2.1.1, footnote 6 and Article 2.4.1²⁸⁴ and, on the other, the use of "investigation phase" which appears only in Article 2.4.2. Although not defined in the AD Agreement, the meaning of the concept of a "period of investigation" in anti-dumping proceedings is well understood.²⁸⁵ We note that a Recommendation Concerning the Periods of Data Collection for Anti-Dumping Investigations adopted by the Committee on Anti-Dumping Practices in May 2000 states that the AD Agreement:

"refers to the period for data collection for dumping investigations when it refers to the period of investigation".²⁸⁶

If the drafters of the AD Agreement had intended to refer in Article 2.4.2 to "the period of data collection" used by investigating authorities as the basis for their findings, they could have used the concept of "period of investigation", which they consistently used in four other places in the same Article, including in the last sentence of Article 2.4.1, which immediately precedes the sentence in Article 2.4.2 that contains the phrase "during the investigation phase".²⁸⁷ We note in this respect the very clear difference in the French and Spanish texts of the AD Agreement between "la p�riode couverte par l'enqu�te" and "la phase d'enqu�te" and between "el per�odo objeto de investigaci�n" and "la etapa de investigaci�n". The argument that "the investigation phase" can be understood to mean "the period of investigation" therefore lacks support in the text of the AD Agreement.

7.192 Further, the overall characterization by the European Communities of how Article 2 addresses the issue of "period of investigation" and the role played by Article 2.4.2 in that regard is unconvincing. The provisions in Article 2 that employ the term "period of investigation" do not set out a general rule restricting data collection in the context of normal value to the period of investigation such that a counterpart in the context of comparison and export prices is required, but rather address very specific issues relating to the period of investigation in particular contexts.²⁸⁸ Thus, the proposition that Article 2.4.2's reference to "during the investigation phase" was intended to ensure that such a data limitation requirement extended to the comparison phase and to export prices is difficult to sustain. The question thus remains why Article 2.4.2 would contain an explicit limitation on the use of data outside the period of investigation in one particular context while no general requirement to that effect may be identified in Article 2.

7.193 More generally, it is hard to imagine that if the drafters had wanted to express the rule that data collection must be restricted to a period of investigation they would have chosen to express themselves in the manner in which Article 2.4.2 is written. While we acknowledge that the WTO Agreement is not always drafted in the clearest terms, we would have expected that a general rule to that effect would have been expressed as a distinct thought somewhere in Article 2, and would have in some manner indicated that as a general rule Members should only rely on data from within that period of investigation.

7.194 That "investigation phase" in Article 2.4.2 cannot be equated with "period of investigation" as used elsewhere in Article 2 (and in Article 9.5) is confirmed by the statement of the Appellate Body that:

"Article 2.4.2 of the Anti-Dumping Agreement permits the use of three methodologies, applicable during the investigation phase, for establishing the existence of 'margins of dumping'."²⁸⁹

Therefore, it is clear that the Appellate Body interprets "during the investigation phase" as pertaining to the issue of when the three methodologies contained in Article 2.4.2 are applicable, rather than to the period of time with respect to which the existence of margins of dumping is established.

7.195 The second possible interpretation advanced by the European Communities is that "during the investigation phase" in Article 2.4.2 refers to the period during which investigating authorities have to make their determination.²⁹⁰ However, Article 2 deals with the methodology for determining the existence of dumping and does not address procedural aspects, such as the timing of determinations. Articles 5.10, 9.3 and 11 contain specific requirements regarding the time-limits applicable to investigations, assessment proceedings and reviews. Thus, to read into "during the investigation phase" in Article 2.4.2 a requirement that investigating authorities abide by the time-limits specified in Articles 5, 9 and 11 of the Agreement would render this phrase redundant. Moreover, it is not clear why the need to state expressly that authorities must abide by those time-limits arises only with respect to this particular aspect of a determination, and not with respect to others such as the injury determination.

7.196 The third possible interpretation advanced by the European Communities of the meaning "during the investigation phase" is that it clarifies that Article 2.4.2 does not apply to the pre-initiation phase. This interpretation poses the same kind of interpretive problems as the other two suggested interpretations. The issue of the requirement to be met by an application for initiation of an anti-dumping investigation are dealt with in some detail in Article 5.2. It is clear from Article 5.2(iii) that an applicant is required to simply provide information on domestic and export prices and not to perform the calculations foreseen in Article 2.4.2. Therefore, there was no need for the drafters of the Agreement to clarify that Article 2.4.2 does not apply to the pre-establishment stage. Moreover, there is no logical explanation why the drafters would have considered it necessary to provide such a clarification specifically with respect to Article 2.4.2.

7.197 In sum, the possible interpretations offered by the European Communities in an attempt to demonstrate that "during the investigation phase" does not limit the applicability of the requirements of Article 2.4.2 to investigations within the meaning of Article 5 are not persuasive. We note that the European Communities itself does not actually endorse any of these interpretations.²⁹¹

(iv) Arguments of the European Communities regarding object and purpose of Article 9.3 and the AD Agreement

7.198 Based on an analysis of text and context of Article 2.4.2 of the AD Agreement we have found that the application of this provision is limited to investigations within the meaning of Article 5. However, the European Communities argues that an interpretation of Article 2.4.2 as being limited to original investigations cannot be correct because it would be contrary to the object and purpose of Article 9.3 and the AD Agreement. While we are not convinced that these arguments are properly characterized in terms of "object and purpose"²⁹², we consider them important and therefore consider them on their merits.

7.199 The European Communities asserts that since the purpose of a duty assessment proceeding under Article 9.3.1 is simply to update the temporal frame for the normal value, there is no rational justification for differentiating between the comparison method applied in an assessment proceeding and the comparison method applied in an investigation. The European Communities also argues that the effect of an interpretation that restricts the application of Article 2.4.2 to investigations is that in the same factual situation a greater amount of duties will be collected in a retrospective system than in a prospective system. By contrast, the United States argues that because the purpose of a duty assessment proceeding is fundamentally different from the purpose of an investigation to determine whether dumping exists and given the explicit recognition in Article 9 of various duty assessment systems, the application of the comparison methods set out in Article 2.4.2 necessarily had to be limited to investigations within the meaning of Article 5.

7.200 We recall that in US � Carbon Steel the Appellate Body found, with respect to the lack of an explicit reference in Article 21.3 of the SCM Agreement to a de minimis standard, that the non-application of the de minimis standard to sunset reviews was not an irrational or absurd result because of the "qualitative differences" between the purpose of investigations under Article 11 of the SCM Agreement and the purpose of sunset reviews under Article 21.3. It follows from the reasoning of the Appellate Body in that case that it is perfectly possible that the fact that a particular aspect of the methodology provided for in Article 2 of the AD Agreement only applies in the context of an investigation within the meaning of Article 5 may be explained by qualitative differences between the purpose of an investigation under Article 5 and the purposes of subsequent phases of an anti-dumping proceeding.

7.201 In this regard, we note that Article 9 of the AD Agreement is concerned with the "Imposition and Collection of Anti-Dumping Duties" and that, as stated by the Appellate Body, rules on the determination of the margin of dumping are "distinct and separate" from rules on the imposition and collection of anti-dumping duties. It follows from Article 9.2, which provides that anti-dumping duties are collected on imports of the product in respect of which an anti-dumping duty has been imposed, that the amount of duty to be assessed in a proceeding under Article 9.3 is the amount of duty payable on imports of the subject product. In our view, the fact that in an assessment proceeding in Article 9.3 the margin of dumping must be related to the liability incurred in respect of particular import transactions is an important element that distinguishes Article 9.3 proceedings from investigations within the meaning of Article 5. In the context of Article 5, when the question is whether dumping "exists" so as to determine whether or not the imposition of an anti-dumping measure is justified, the focus is on the overall pricing behaviour of the exporter under investigation. By contrast, in an Article 9.3 context the extent of dumping found with respect to a particular exporter must be translated into an amount of liability for payment of anti-dumping duties by importers in respect of specific import transactions. Considerations that are relevant to determining the design of an appropriate comparison methodology in the context of investigations to determine whether dumping exists may not apply with equal force to the design of a methodology for determining the final liability for payment of anti-dumping duties. Therefore, we do not agree with the view of the European Communities that there could be no rational basis for differentiating between the comparison method applied in an assessment proceeding and the comparison method applied in an investigation.

7.202 Similarly, the qualitative differences between determining the existence of margins of dumping in an investigation and determining margins of dumping as part of the imposition and collection of an anti-dumping duty mean that the results of an administrative review cannot "eclipse" the results of the "original investigation". The comparison methodology used in the investigation is one element in establishing whether or not the imposition of an anti-dumping duty is justified. To the extent that the methodology used in the investigation has established the legal basis for the imposition of an anti-dumping duty, it cannot be superseded by the methodology subsequently used to determine the amount of duty to be collected. In light of this difference in purpose between the collection of anti-dumping duties and the determination of the existence of margins of dumping, it does not follow that the fact that a particular provision of Article 2 is not applicable to Article 9.3 is necessarily illogical.

7.203 We must now proceed to consider whether the non-application to Article 9.3 of Article 2.4.2 in particular is comprehensible in terms of the differences in purpose or function between investigations and duty assessment. We note the argument of the European Communities that Article 9.3 must be interpreted in light of Article 2.4.2 and therefore prohibits duty assessment on a transaction-specific and importer-specific basis. The European Communities submits that all export transactions that fall within the time period covered by a review must be treated as a whole rather than being treated individually. Accordingly, the position of the European Communities is that in normal circumstances, if an authority establishes an average normal value based on a particular time period, the amount of anti-dumping duty must be established by comparing that average normal value with the average price of all export transactions that have occurred during the same period. In this situation an importer will incur liability for payment of anti-dumping duties on an individual import transaction if the exporter's average export price is below the average normal value, regardless of the export price of that particular transaction. The assessment of liability based on a comparison of an average normal value with individual export prices would be permitted only in the circumstances set out in the second sentence of Article 2.4.2.

7.204 We see no textual support in Article 9.3 for the view that the AD Agreement requires an exporter-oriented assessment of anti-dumping duties, whereby, if an average normal value is calculated for a particular review period, the amount of anti-dumping duty payable on a particular transaction is determined by whether the overall average of the export prices of all sales made by an exporter during that period is below the average normal value. While the present AD Agreement is more detailed in its provisions on duty assessment than the Tokyo Round Anti-Dumping Code in that it explicitly reflects the existence of retrospective and prospective duty assessment systems, Articles 9.3 and 9.4 provide little detail as regards the substantive methodology to be followed by an authority to determine the basis for attributing liability in respect of a particular transaction.²⁹³ In our view, if the drafters of the AD Agreement had wanted to impose a uniform requirement to adopt an exporter oriented-method of duty assessment, which would have entailed a significant change to the practice and legislation of some participants in the negotiations, they might have been expected to have indicated this more clearly.

7.205 If anything, the plain text of Article 9 contradicts the view that the AD Agreement requires an exporter-oriented approach to duty assessment based on the application of the comparison methods set out in Article 2.4.2. Article 9.4 provides inter alia that an anti-dumping duty applied to imports of exporters or producers who have not been included in an examination that has been limited in accordance with Article 6.10, shall not exceed:

"... where the liability for payment of anti-dumping duties is calculated on the basis of a prospective normal value, the difference between the weighted average normal value of the selected exporters or producers and the export prices of exporters or producers not individually examined."

7.206 This provision necessarily means that duty assessment based upon a prospective normal value is compatible with Article 9.3. To calculate liability on the basis of a prospective normal value means that prices of individual export transactions are compared with a minimum reference price. As discussed in the panel report in Argentina � Poultry Anti-Dumping Duties, Article 9.4(ii) thus clearly envisions that a Member may operate a system of variable duties.²⁹⁴ We see nothing in Article 9.4 to suggest that if duty is initially collected on the basis of the extent by which the export price in a particular transaction is below the prospective normal value, a Member must subsequently, in a refund procedure under Article 9.3.2, determine the entitlement to a refund of anti-dumping duties by comparing an average normal value with an average export price charged by a given exporter to all importers. We can see no textual support for such an interpretation.

7.207 In addition, we consider that the European Communities has not presented a convincing explanation as to why as a matter of logic or fairness Article 2.4.2 must be interpreted as applying to duty assessment under Article 9.3. In particular, we fail to see why the AD Agreement must be interpreted to require that when an authority calculates an average normal value for a review period, an importer can be liable to pay anti-dumping duties on an import of the subject product even where the export price in that transaction actually exceeds the average normal value. While we are of course aware that the Appellate Body has ruled that dumping can be found to exist only for the product under investigation as a whole, we note that the Appellate Body expressed that view in cases involving investigations within the meaning of Article 5 and that the Appellate Body has not yet had the occasion to consider in detail the question of how the amount of anti-dumping duty must be assessed for purposes of Article 9.3. Given that the Appellate Body has also repeatedly stressed the need to have regard to distinct purposes of provisions of the AD Agreement and the SCM Agreement, and has in particular pointed to the distinction between rules on the existence of margins of dumping and rules on the imposition and collection of anti-dumping duties, we do not believe that the general notion that dumping can be found only for a product can sustain the view that Article 9.3 must be read to require an exporter-oriented approach to duty assessment.

7.208 We next turn to the argument of the European Communities that the non-application of Article 2.4.2 to assessment proceedings under Article 9.3 would entail unequal treatment of prospective and retrospective duty assessment systems. The essence of this argument is that while in a prospective system the amount of duty collected cannot exceed the margin of dumping established without the use of the zeroing methodology in the initial investigation, in a retrospective system, such as the one applied by the United States, the amount of duty that is ultimately collected can exceed the original margin of dumping because it will be determined in periodic reviews in which a zeroing methodology will be used.

7.209 First, we note that it is not necessarily the case that in a prospective duty assessment system the amount of duty collected is always limited by the magnitude of the margin of dumping found in the initial investigation If a duty is collected based on a prospective normal value, the amount of that duty is determined by the extent to which the export price is below the prospective normal value.

7.210 Second, our interpretation of Article 2.4.2 as not applicable beyond "the investigation phase" in Article 5 applies to any type of duty assessment system, whether retrospective or prospective. Thus the consequence of our interpretation is that a Member applying a prospective duty assessment system is not required to apply Article 2.4.2 when it decides whether or not to grant a refund under Article 9.3.2. As discussed above, we see no support in Article 9.3 for a view that the average price behaviour of an exporter must determine the amount of liability incurred in relation to a particular import transaction. Consequently, we reject the view that the AD Agreement should be interpreted in such a manner as to ensure that both prospective and retrospective duty assessment systems provide an identical level of protection against dumped imports corresponding to the actual margin of dumping of the exporter in question.

7.211 We realize of course that in a prospective duty system the amount by which the duty exceeds the actual margin of dumping will be reimbursed in a refund procedure but that there is no possibility to collect the amount by which the duty is less than the actual margin of dumping. By contrast, in a retrospective system it is possible to collect the difference between the actual margin of dumping and the deposit paid. Thus, regardless of whether or not Article 2.4.2 applies to duty assessment proceedings, there is an inherent difference between retrospective and prospective duty assessment systems which, in this respect at least, constitutes an "advantage" for users of retrospective systems. The European Communities does not appear to dispute that the AD Agreement, in explicitly recognizing the appropriateness of both retrospective and prospective duty assessment systems, enshrines this advantage. We do not think that this "inequality" in the amount of duty collected is of any legal relevance to the interpretation of Article 2.4.2 of the AD Agreement. A variety of factors determine the choice by a Member of an appropriate duty assessment system. The fact that a prospective system may not in a given situation lead to the same amount of duty that it would be possible to collect in a retrospective system may well be outweighed by other considerations. Thus, a prospective system can be considered to have certain advantages over a retrospective system, for example in terms of predictability and ease of administration.²⁹⁵ Therefore, we disagree with the assumption underlying the argument of the European Communities that the AD Agreement must be interpreted to mean that in an identical situation retrospective and prospective duty assessment systems must necessarily lead to the same level of protection against dumped imports.

7.212 Grosso modo, the "object and purpose" arguments of the European Communities amount to the assertion that it would not make sense to interpret Article 2.4.2 to be limited to original investigations.  In light of the foregoing elements, however, we do not think that it would have been irrational for negotiators to agree to limit Article 2.4.2 to the investigation phase.  In any event, we cannot preclude that the explanation for such a differentiation might be the result of negotiating dynamics.  Specifically, it could reflect a compromise bridging different interests, even if it was not the initial negotiating position of any party.  From the perspective of those seeking to limit the use of antidumping duties, the limitation would make it more difficult to impose an anti-dumping duty in the first instance, while from the perspective of those defending the use of anti-dumping duties, it would have allowed them to continue their then-current practices once a measure was imposed.

7.213 In light of the foregoing considerations, we conclude that the European Communities has not advanced any arguments with respect to what it characterizes as object and purpose of Article 9.3 that would cause us to conclude that negotiators cannot have intended to limit the scope of Article 2.4.2 to investigations within the meaning of Article 5.²⁹⁶

(v) Arguments of the European Communities regarding subsequent practice

7.214 The European Communities asserts that its interpretation of Article 2.4.2 of the AD Agreement is supported by the fact that a review of notifications provided by 105 Members²⁹⁷ reveals that none of these Members has in its domestic legislation limited the application of Article 2.4.2 to investigations under Article 5. The European Communities claims that this is "subsequent practice" within the meaning of Article 31(3)(b) of the Vienna Convention.²⁹⁸

7.215 Article 31(3)(b) of the Vienna Convention provides, as one element of the general rule of interpretation, that:

"there shall be taken into account, together with the context, ... any subsequent practice in the application of the treaty which establishes the agreement of the parties regarding its interpretation."

7.216 The Appellate Body has interpreted the term "subsequent practice" as used in this provision to mean:

"...a 'concordant, common and consistent' sequence of acts or pronouncements which is sufficient to establish a discernible pattern implying the agreement of the parties [to a treaty] regarding its interpretation.

...a discernible pattern of acts or pronouncements implying an agreement among WTO Members on the interpretation of [the relevant provision]".²⁹⁹

7.217 We note that the argument of the European Communities with respect to subsequent practice is based on the fact that the particular provisions in domestic laws and/or regulations that have been notified to the Committee on Anti-Dumping Practices³⁰⁰ that in the view of the European Communities correspond to Article 2.4.2 of the AD Agreement either do not include the phrase "during the investigation phase" or use the phrase "during the investigation period."³⁰¹ In our view, this is simply not a sufficient basis to draw a conclusion as to whether or not it is the practice of a particular Member to apply Article 2.4.2 to the assessment of the amount of anti-dumping duties within the meaning of Article 9.3. Many Members have notified legislation that predates the WTO Agreement. We also note that in many cases, the legislation identified by the European Communities makes no or little mention of duty assessment at all, and it is therefore difficult to draw any conclusion regarding the specific methodology applied. We note, however, that many Members provide in their domestic anti-dumping legislation for the collection of anti-dumping duties through a system of variable duties. The transaction-specific character of such a duty assessment system would appear to suggest that those Members do not apply the symmetrical comparison methods foreseen in Article 2.4.2 to determine the amount of liability for payment of anti-dumping duties.

7.218 In any event, even if it were established conclusively that all the 76 Members referred to by the European Communities have adopted a practice of applying Article 2.4.2 to duty assessment, this would only mean that a considerable number of WTO Members have adopted an approach different from that of the United States. We fail to see how one can conclude on this basis that there exists "a discernible pattern of acts or pronouncements implying an agreement among WTO Members on the interpretation of" Article 2.4.2. We note that one third party in this proceeding submitted arguments contesting the view of the European Communities that Article 2.4.2 applies to the imposition and collection of anti-dumping duties. Therefore, we conclude that even if the documentation provided by the European Communities were relevant as evidence of "practice" within the meaning of Article 31(3)(b) of the Vienna Convention, that practice is not a practice "which establishes the agreement between the parties regarding the interpretation" of Article 2.4.2. Consequently, the reference by the European Communities to "subsequent practice" does not undermine the conclusion we have reached based on an interpretation of Article 2.4.2 in accordance with Article 31(1) of the Vienna Convention.

(vi) Reference made by the European Communities to supplementary means of treaty interpretation

7.219 We do not consider that an interpretation of Article 2.4.2 of the AD Agreement in accordance with Article 31 of the Vienna Convention leaves the meaning of that provision "ambiguous or obscure" or "leads to a result which is manifestly absurd or unreasonable".³⁰² Therefore, we see no need to have recourse to any of the supplementary means of treaty interpretation within the meaning of Article 32 of the Vienna Convention. In any event, the documentation submitted by the European Communities relating to the negotiating history of the Uruguay Round AD Agreement³⁰³ does not support its argument on the interpretation of Article 2.4.2 of the AD Agreement. First, none of the documents provided by the European Communities indicates that there was an understanding among negotiators that rules in the draft AD Agreement referring to "investigations" should apply generally to any provision that requires a process of systematic examination, inquiry or careful study. Second, we note that the European Communities relies on this negotiating history inter alia for the proposition that there never was a suggestion that a distinction should be made between investigations and retrospective assessment proceedings as regards the application of Article 2.4.2.³⁰⁴ However, most of the documents submitted by the European Communities pertain to a period of the Uruguay Round negotiations prior to the intense phase of work that led to the release in November 1990 of a draft ("New Zealand I") that introduced the phrase "when establishing the existence of dumping margins during the investigation phase" in what is now Article 2.4.2. It is therefore not surprising that the documentation provided by the European Communities is silent on the background to the insertion of this phrase. What is clear is that the insertion of this phrase marked a significant difference with the corresponding provisions in two previous draft texts (Article 2.4.3 in "Carlisle I" and "Carlisle II"). It is also clear from the fact that the second sentence of Article 2.4.2 of the "New Zealand I" text used "period of investigation" that "investigation phase" in the first sentence must have been understood to have a different meaning.

(vii) General Conclusion

7.220 We recapitulate the conclusions of our analysis. First, the phrase "the existence of margins of dumping during the investigation phase" in Article 2.4.2 read in its ordinary meaning in context of the AD Agreement as a whole means that Article 2.4.2 applies to the phase of the "original investigation" i.e. the investigation within the meaning of Article 5 of the AD Agreement, as opposed to subsequent phases of duty assessment and review. Second, our interpretation of the meaning of this phrase as limiting the applicability of Article 2.4.2 to investigations within the meaning of Article 5 is also consistent with the distinction made between investigations and subsequent proceedings in various Appellate Body decisions. Third, alternative meanings suggested by the European Communities are implausible at best and deny this phrase any real function, in contradiction with principles of interpretation. Fourth, this interpretation is entirely consistent with the different functions played by "original investigations" and duty assessments proceedings. Finally, the references made by the European Communities to subsequent practice and preparatory work do not undermine this interpretation.

7.221 We recall that the standard that we must apply in interpreting the AD Agreement is set forth in Article 17.6(ii) of the AD Agreement:

"the panel shall interpret the relevant provisions of the Agreement in accordance with customary rules of interpretation of public international law. Where the panel finds that a relevant provision of the Agreement admits of more than one permissible interpretation, the panel shall find the authorities' measure to be in conformity with the Agreement if it rests upon one of those permissible interpretations."

7.222 Based on an analysis in accordance with customary rules of treaty interpretation, we conclude that the scope of Article 2.4.2 of the AD Agreement is limited to the establishment of the existence of margins of dumping in the context of investigations under Article 5 of the AD Agreement. In any event, we cannot find that an interpretation that restricts the scope of application of Article 2.4.2 to investigations within the meaning of Article 5, thereby excluding its application to Article 9.3, is not permissible within the meaning of Article 17.6(ii)

7.223 The Panel therefore finds that the United States did not act inconsistently with Article 2.4.2 of the AD Agreement when, in the administrative reviews listed in Exhibits EC-16 to EC-31³⁰⁵, USDOC used a methodology that involved asymmetrical comparisons between export price and normal value and in which no account was taken of any amount by which export prices exceeded normal value.  
 

To continue with  4. Article 2.4 of the AD Agreement

Return to Table of Contents

176 EC-First Written Submission, paras. 16, 21-22, 37-38; EC-Response to Question from the United States following the First Meeting, para. 5; EC-Response to Panel Question 51 (para. 166) and EC-Rebuttal Submission, para. 69.

177 EC-Response to Panel Question 51.

178 EC-First Written Submission, para. 125.

179 EC-First Written Submission, paras. 104-124.

180 EC-First Written Submission, paras. 126-129.

181 EC-Response to Panel Question 53.

182 EC-Rebuttal Submission, paras. 71-82; EC-Response to Panel Question 72.

183 EC-Response to Question from the United States following the First Meeting.

184 EC-Second Oral Statement, paras. 55-56.

185 In its First Submission (paras. 82-83), the United States notes in this respect the lack of precision of the European Communities as to what it means by the term "Standard Zeroing Procedures" and assumes that the European Communities means the Anti-Dumping Manual and the Anti-dumping Margin Programme.

186 US-First Written Submission, paras. 81-92.

187 US-First Written Submission, paras. 93-97.

188 US-Response to Panel Questions 57, 59-60, 61, 63-65 and 73.

189 US-Second Written Submission, paras. 60-64.

190 Appellate Body Report, US � Corrosion-Resistant Steel Sunset Review, paras. 81-87. (Italics added)

191 Appellate Body Report, US � Corrosion-Resistant Steel Sunset Review, para. 88.

192 Appellate Body Report, US � Corrosion-Resistant Steel Sunset Review, paras. 158 and 190.

193 Appellate Body Report, US � Oil Country Tubular Goods Sunset Review, para. 187.

194 Appellate Body Report, US � Oil Country Tubular Goods Sunset Review, para.187.

195 We also note that the Appellate Body has explicitly stated in a recent case that it "... has not, to date, pronounced upon the issue of whether "practice" may be challenged as such, as a measure in WTO dispute settlement." Appellate Body Report, US � Gambling, para. 132. See also Appellate Body Report, Oil Country Tubular Goods Sunset Reviews, para. 220 ("Therefore, even assuming argued that a practice may be challenged as a measure in WTO dispute settlement- an issue on which we express no view here...").

196 Appellate Body Report, US � Corrosion-Resistant-Steel Sunset Review, para. 82.

197 In US � Countervailing Measures on Certain EC Products, the Panel and the Appellate Body examined and made findings with respect to the "same person method" applied by USDOC for the purposes of determining the continued existence of a "benefit" within the meaning of the SCM Agreement following a change in ownership.

198 Appellate Body Report, US � Oil Country Tubular Goods Sunset Reviews, para. 172.

199 We note more generally that GATT and WTO panels have sometimes relied on evidence of an informal nature to establish the existence of a measure or policy. We find instructive in this regard the discussion by the Panel in Japan � Film of the term "measures" as it applies to "administrative guidance". Panel Report, Japan � Film, paras. 10.43-10.51.

200 US-Response to Panel Question 73. We also note, in this respect, the argument of the United States that it was well known by Uruguay Round negotiators that zeroing was utilised by the United States and other signatories to the Tokyo Round Anti-Dumping Code. US-Second Written Submission, para. 89.

201 In this context, we note that the United States has argued that USDOC has the discretion to zero or not and that, under the mandatory/discretionary distinction and "WTO/GATT" practice, a WTO Member must be presumed to implement its WTO obligations in good faith. This rationale, however, only applies where a Member's legislation provides discretion to the executive branch to avoid acting WTO-inconsistently and it is unknown how that discretion will be used. (E.g., GATT Panel report, US � Superfund, para. 5.2.9 and GATT Panel Report, US � Tobacco, para. 123) In such a situation, the rationale for applying a mandatory/discretionary distinction is that it should not be presumed that the Member will not act in good faith and implement its legislation in a WTO-inconsistent manner. The same rationale, based on an assumption of implementation in good faith, cannot apply when a Member has adopted a norm that systematically leads to WTO-inconsistent conduct.

202 Ball Bearings and Parts Thereof from France, Germany, Italy, Japan, and the United Kingdom; Final Results of Antidumping Duty Administrative Reviews, 67 Fed. Reg. 55780 (30 August 2002) (the European Communities challenges this determination with respect to France, Italy and the United Kingdom); Stainless Steel Plate in Coils from Belgium; Final Results of Antidumping Duty Administrative Review, 67 Fed. Reg. 64352 (18 October 2002); Stainless Steel Sheet and Strip in Coils From Italy; Final Results of Antidumping Duty Administrative Review, 67 Fed. Reg. 1715 (14 January 2002); Stainless Steel Sheet and Strip in Coils from Italy: Final Results of Antidumping Administrative Review, 68 Fed. Reg. 6719 (10 February 2003); Notice of Final Results of Antidumping Duty Administrative Review: Stainless Steel Sheet and Strip in Coils from France, 67 Fed. Reg. 6493 (12 February 2002); Notice of Final Results of Antidumping Duty Administrative Review: Stainless Steel Sheet and Strip in Coils from France, 67 Fed. Reg. 78773 (26 December 2002); Stainless Steel Sheet and Strip in Coils From Germany; Notice of Final Results of Antidumping Duty Administrative Review, 67 Fed. Reg. 7668 (20 February 2002); Notice of Final Results of Antidumping Duty Administrative Review: Stainless Steel Sheet and Strip in Coils from Germany, 68 Fed. Reg. 6716 (10 February 2003); Notice of Final Results of Antidumping Duty Administrative Review and Determination Not to Revoke in Part: Certain Pasta from Italy, 68 Fed. Reg. 6882 (11 February 2003); Notice of Final Results of Antidumping Duty Administrative Review: Granular Polytetrafluoroethylene Resin from Italy, 67 Fed. Reg. 1960 (15 January 2002); Notice of Final Results of Antidumping Duty Administrative Review: Granular Polytetrafluoroethylene Resin from Italy, 68 Fed. Reg. 2007 (15 January 2003); Notice of Final Results of Antidumping Duty Administrative Review, Partial Rescission of Antidumping Duty Administrative Review and Revocation of Antidumping Duty Order in Part: Certain Pasta From Italy, 67 Fed. Reg. 300 (3 January 2002); Industrial Nitrocellulose From France: Final Results of Antidumping Duty Administrative Review, 66 Fed. Reg. 54213 (26 October 2001); Industrial Nitrocellulose From the United Kingdom; Final Results of Antidumping Duty Administrative Review, 67 Fed. Reg. 77747 (19 December 2002). The European Communities treats the administrative review in Ball Bearings from Italy as a sample case. The European Communities indicates that in addition to the Final Results of the administrative review, the measures at issue include the amendments to these Final Results and that the Final Results of the administrative review refers to the accompanying Issues and Decision Memorandum, which in turn refers to the Margin Calculations, i.e. the Final Margin Program Log and Outputs for the firms investigated, and to the assessment instructions. EC-First Written Submission, para. 148.

203 EC-First Written Submission, paras. 164-170.

204 EC- Oral Statement at the First Substantive Meeting of the Panel, paras. 24 and 31-37.

205 EC-Response to Panel Question 14.

206 EC-Response to Panel Question 52, paras. 55-71.

207 See also EC-Response to Panel Question 52, paras. 72-82

208 EC-Second Oral Statement, para. 28.

209 EC-Second Written Submission, paras. 118 and 156-160; EC-Oral Statement at the Second Substantive Meeting of the Panel, paras. 19-27; EC-Closing Oral Statement at the Second Substantive Meeting of the Panel, para. 11.

210 EC-Second Written Submission, paras. 154-155; EC-Closing Oral Statement at the Second Substantive Meeting of the Panel, para. 12.

211 EC-Second Written Submission, para. 28.

212 EC-Second Written Submission, paras. 123-136.

213 EC-Oral Statement at the Second Substantive Meeting of the Panel, paras. 8-9; EC-Closing Oral Statement at the Second Substantive Meeting of the Panel, paras. 9 and 15-16.

214 EC-Second Written Submission, paras. 168-170.

215 EC-Second Written Submission, paras. 179-187.

216 EC-Second Written Submission, paras. 188-197.

217 EC-Second Written Submission, paras. 199-217.

218 EC-Second Written Submission, para. 218; EC-Oral Statement at the Second Substantive Meeting of the Panel, para. 37.

219 EC-Closing Oral Statement at the Second Substantive Meeting of the Panel, paras. 5-6.

220 EC-Rebuttal Submission, paras. 30-37; EC-Oral Statement at the Second Substantive Meeting of the Panel, paras. 65-66.

221 EC-Response to Panel Question 39.

222 US-First Written Submission, paras. 36-39.

223 US-First Written Submission, paras. 44-46.

224 US-Response to Panel Questions 10, 13 and 14; US-Second Written Submission, paras. 5-13.

225 US-First Written Submission, paras. 50-54.

226 US-First Written Submission, para. 47.

227 US-First Written Submission, paras. 55-57.

228 US-First Written Submission, para. 31.

229 US-Opening Statement at the First Substantive Meeting of the Panel, para. 18.

230 US-Opening Statement at the First Substantive Meeting of the Panel, paras. 8-10.

231 US-Closing Statement at the Second Substantive Meeting of the Panel, paras. 16-17.

232 US-Response to Panel Question 11.

233 US-Second Written Submission, paras. 31-35.

234 US-Closing Statement at the Second Substantive Meeting of the Panel, paras. 12-14 and Exhibit US-3.

235 Supra, footnote 202.

236 The term "administrative review" in this context means "periodic review of the amount of antidumping duty" as required by Section 751(a) of the Tariff Act. This provision requires the administering authority to review and determine the amount of any anti-dumping duty at least once during each 12-month period beginning on the anniversary of the date of publication of an anti-dumping duty order if a request for such a review has been received. The United States uses a retrospective assessment system under which final liability for anti-dumping duties is determined after merchandise is imported. USDOC's Regulations (�351.414(c)(2) and (e)) provide that in such a periodic review of the amount of anti-dumping duty USDOC will normally use the average-to-transaction method and that, in applying that method, when normal value is based on the weighted average of sales of the foreign like product, the USDOC will limit the averaging of such prices to sales incurred during the contemporaneous month.

237 �351.212 of USDOC's Regulations provide that if the Secretary has conducted an administrative review, a new shipper review or an expedited antidumping review, "the Secretary normally will calculate an assessment rate for each importer of subject merchandise covered by the review. The Secretary normally will calculate the assessment rate by dividing the dumping margin found on the subject merchandise examined by the entered value of such merchandise for normal customs duty purposes. The Secretary then will instruct the Customs Service to assess antidumping duties by applying the assessment rate to the entered value of the merchandise."

238 It is not in dispute that USDOC included the values of all export transactions in the denominator of the aggregate margin calculation.

239 In addition, Article 2.1 contains a definition of dumping that applies to the entire AD Agreement. Appellate Body Report, US � Corrosion Resistant Steel Sunset Review, paras. 109 and 126.

240 We emphasize that the issue before us is limited to whether Article 2.4.2 applies to duty assessment proceedings under Article 9.3. The applicability of other provisions of Article 2 to such proceedings is not at issue in this dispute.

241 We note in this regard that the argument of the European Communities that the word "investigation" in Article 2.4.2 applies to any provision of the AD Agreement that obligates an authority to conduct "a systematic examination or inquiry or a careful study" logically implies that all proceedings under the AD Agreement � e.g., duty assessment proceedings under Article 9.3, new shipper reviews under Article 9.5, "interim" or "changed circumstances" reviews under Article 11.2 and sunset reviews under Article 11.3 � would be "during the investigation phase" and thus be subject to Article 2.4.2. Thus, while the issue before us with respect to this "as applied" claim is limited to the application of Article 2.4.2 to duty assessment proceedings, we must arrive at a more general understanding of the scope of application of Article 2.4.2, both in order to reach a reasoned conclusion regarding this claim, and in order to resolve this dispute in respect of the other as such claims. We also note that the European Communities has indicated that the implication of its position that Article 2.4.2 applies to duty assessment proceedings under Article 9.3 is that both the "import-specific" and the "importer-specific" aspects of the United States' duty assessment system are inconsistent with the AD Agreement. Thus, whether the AD Agreement must be interpreted to proscribe import-specific and importer-specific methods of assessment of anti-dumping duties is an unavoidable and central aspect of our analysis of the claim of the European Communities under Article 2.4.2. We discuss the issue of whether duty assessment must be based on the overall behaviour of an exporter or whether it is permissible to focus on individual transactions and individual importers below in paras. 7.203-7.207.

242 "One of the corollaries of the 'general rule of interpretation' in the Vienna Convention is that interpretation must give meaning and effect to all the terms of the treaty. An interpreter is not free to adopt a reading that would result in reducing whole clauses or paragraphs of a treaty to redundancy or inutility". Appellate Body Report, US �- Gasoline, p. 23.

243 The position of the European Communities that Article 2.4.2 of the AD Agreement applies to the assessment of the amount of an anti-dumping duty under Article 9.3 is based essentially on the following propositions: (1) The key interpretive issue before the Panel is the meaning of the word "investigation". (2) In the absence of a generic definition limiting the meaning of "investigation", its ordinary meaning must be established on the basis of dictionary definitions. (3) The dictionary definition of "investigation" is "a systematic examination or inquiry or a careful study of a particular subject". (4) The word "investigation" in Article 2.4.2 therefore applies to any provision of the AD Agreement that obligates an authority to conduct "a systematic examination or inquiry or a careful study". (5) A duty assessment proceeding under Article 9.3 is an investigation in this sense.

244 The fundamental rule of treaty interpretation that governs our analysis is that: A treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose (Vienna Convention, Art. 31(1)) In our view, it is not consistent with this rule to interpret Article 2.4.2 by focusing on the word "investigation" in isolation as determinative and to equate the ordinary meaning of this word with its dictionary definition. We note in the latter regard that the Appellate Body has on several occasions cautioned against placing too much emphasis on dictionary definitions. In US � Gambling, the Appellate Body observed that: "In order to identify the ordinary meaning, a panel may start with the dictionary definition of the terms to be interpreted. But dictionaries, alone, are not necessarily capable of resolving complex questions of interpretation, as they typically aim to catalogue all meanings of words� be those meanings common or rare, universal or specialized". Appellate Body Report, US- Gambling, para. 164 (footnotes omitted, emphasis in original). In the same Report, the Appellate Body stated that: "... to the extent that the Panel's reasoning simply equates the 'ordinary meaning' with the meaning of words as defined in dictionaries, this is, in our view, too mechanical an approach". Appellate Body Report, US � Gambling, para. 166. In US � Offset Act (Byrd Amendment), the Appellate Body observed that: "It should be remembered that dictionaries are important guides to, not dispositive statements of, definitions of words appearing in agreements and legal documents". Appellate Body Report, US � Offset Act (Byrd Amendment) para. 248.

245 "Article 2.4.2 of the AD Agreement permits the use of three methodologies, applicable during the investigation phase, for establishing the existence of 'margins of dumping'". Appellate Body Report, US � Softwood Lumber V, para. 76 (footnote omitted, emphasis added). It is implied by this statement that "during the investigation phase" must be read to be associated with the word "established" in the first sentence of Article 2.4.2 rather than with "the existence of margins of dumping". Contrary to what is suggested in the dissenting opinion, we do not believe that any clear grammar rules compel the result the dissenting opinion argues for in any of the three authentic languages, namely that "investigation phase" must be linked to the existence of margins of dumping. We also note in this regard that to read 'investigation phase' as qualifying "the existence of margins of dumping" would effectively amount to equating "investigation phase" with "period of investigation" as used elsewhere in Article 2. However, as discussed infra in paras. 7.190-7.194, "period of investigation" is clearly a concept different from "investigation phase". This difference in meaning between "investigation phase" and "period of investigation" thus supports a reading of "investigation phase" as being associated with the word "established".

246 Infra, Section VII.E.3(c)(ii).

247 Thus, Article 9.3 is part of an article on "Imposition and Collection of Anti-Dumping Duties". Article 9.3.1, the provision that applies to the measures at issue, addresses the "determination of the final liability for payment of anti-dumping duties" "when the amount of the anti-dumping duty is assessed on a retrospective basis. " Footnote 21 refers to "[a] determination of final liability for payment of anti-dumping duties, as provided for in paragraph 3 of Article 9". Article 18 refers to "refund procedures" under Article 9.3. If in a proceeding under Article 9.3 it is found that the margin of dumping in a particular period is zero, this does not amount to a finding, resulting from an "investigation", that dumping did not "exist". Footnote 22 of the AD Agreement provides that "[w]hen the amount of the anti-dumping duty is assessed on a retrospective basis, a finding in the most recent assessment proceeding under subparagraph 3.1 of Article 9 that no duty is to be levied shall not by itself require the authorities to terminate the definitive duty." This footnote refers to "a finding that no duty is to be levied" and not to "a finding that dumping does not exist". Similarly, Article 9.5 of the AD Agreement provides for "a review for the purpose of determining individual margins of dumping for any exporters or producers in the exporting country in question who have not exported the product to the importing Member during the period of investigation..." and does not use the words "investigation" or "investigate" to characterize the nature of this proceeding. A review under this provision is to be carried out on accelerated basis "compared to normal duty assessment and review proceedings" of the importing Member. Article 11 of the AD Agreement ("Duration and Review of Anti-Dumping Duties and Price Undertakings") uses the words "review" and "determine" rather than "investigate". Articles 9.5 and 11 also contain no references to "the existence of margins of dumping".

248 The European Communities argues that it is not possible to dissociate the "existence" of dumping from the "degree" of dumping; the existence of dumping must be established not only in an original investigation but also in other proceedings such as changed circumstances reviews or sunset reviews; the word "existence" is used in other provisions in the AD Agreement without having the meaning attributed to it by the United States; the word "existence" is also part of the title of Article 2 of the French and Spanish texts of the AD Agreement; and the United States does not apply certain provisions of Article 2.4.1 in administrative reviews, even though Article 2.4.1 does not contain the word "existence". EC-Rebuttal Submission, paras. 118 and 156-160; Oral Statement at the Second Substantive Meeting of the Panel by the European Communities, paras. 19-26.

249 We note, in this regard, that the Appellate Body has expressly relied on the phrase "the existence of (margins of) dumping" to distinguish original investigations under the AD Agreement from proceedings under Articles 9 and 11. Infra, paras. 7.177-7.179.

250 The European Communities refers to the word "existence" or "existing" in Articles 3.7, 8.4, 4.1(ii), 18.3, 18.3.2 of the AD Agreement and in Articles 1.1, 2.1(b), 4.2, 6.1, 6.5, 6.7(d), 6.8, 7.2, 7.9, 8.2(c), 9.4, 15.7, 16.2, 17.1(b), 18.4, 19.1 and 19.4, 22.4, 24.3, 27.6, 27.9, 32.3 Annex III, paras. 1-2 and 4, Annex IV, para. 4 and Annex V, para. 2 of the SCM Agreement. Oral Statement at the Second Substantive Meeting of the Panel by the European Communities, paras. 22-23. We do not need to express a view on whether, as argued by the European Communities, all these provisions are in fact relevant to the imposition and collection of anti-dumping and countervailing duties.

251 With regard to the argument that the French and Spanish texts of the AD Agreement refer to "existence" of dumping in the title of Article 2 ("Determination de l'existence d'un dumping" and "determinaci�n de la existencia de dumping"), we consider that this does not alter the fact that Article 2.4.2 is the only provision in Article 2 that expressly relates "the existence of margins of dumping" to "the investigation phase". Since the issue of the application of Article 2.4.1 is not before us, we need not express a view on the scope of application of that provision.

252 As noted above, we consider that the entire AD Agreement can be relevant context for the interpretation of Article 2.4.2.

253 To illustrate this concept: we consider that the word "investigation" is used to refer to (a phase of a) proceeding when, for instance, a provision defines rules that apply to "an investigation" "an anti-dumping investigation" or "the investigation". The question before us is whether in such cases the proceedings referred to are only proceedings under Article 5 or whether these terms are also used in the AD Agreement in connection with proceedings under Articles 9 and 11. Therefore, in our view the relevant question is narrower in scope than whether the word "investigation" per se is always used in the AD Agreement to refer to investigations in the sense of Article 5 of the AD Agreement.

254 We note the argument of the European Communities that footnote 1 does not limit the meaning of "investigations" but explains the meaning of "initiated" in the context of investigations, both under Article 5 and in the context of other types of investigation. However, "investigations" within the meaning of the first sentence of Article 1 are "initiated" by "the procedural action by which a Member formally commences an investigation as provided in Article 5." Therefore, the only "investigations initiated" referred to in Article 1 of the AD Agreement are investigations within the meaning of Article 5.

255 Infra, para. 7.183.

256 The relationship between investigation and the application of measures is also explicit in the text of Articles 6.9 and 6.14. Thus, Article 6.9 provides: "The authorities shall, before a final determination is made, inform all interested parties of the essential facts under consideration which form the basis for the decision whether to apply definitive measures. Such disclosure should take place in sufficient time for the parties to defend their interests". Article 6.14 provides: "The procedures set out above are not intended to prevent the authorities of a Member from proceeding expeditiously with regard to initiating an investigation, reaching preliminary or final determinations, whether affirmative or negative, or from applying provisional or final measures, in accordance with relevant provisions of this Agreement."

257 Article 6.7 provides for "investigations" in the territories of other Members. The fact that in this provision investigation clearly does not refer to an investigation within the meaning of Article 5 is of no particular relevance because it is evident that this provision uses "investigation" to refer to a particular aspect of the data gathering process and not to investigation as a "proceeding".

258 Article 12.2.2 sets forth certain requirements regarding "[a] public notice of conclusion or suspension of an investigation in the case of an affirmative determination providing for the imposition of a definitive duty or the acceptance of a price undertaking". Article 12.2.3 deals with "[a] public notice of the termination or suspension of an investigation following the acceptance of an undertaking."

259 Supra, para. 7.156 and footnote 247.

260 E.g. EC-First Written Submission, para. 178 and Exhibit EC-45.

261 "...exporter or producer under investigation" (Article 2.2.11, first sentence and Article 2.2.2, first sentence); "in the course of the investigation" (Article 2.2.1.1, second sentence); "during the investigation" (footnote 6); "exporters or producers subject to investigation" (Article 2.2.2(ii)) and "period of investigation" (Article 2.2.1, last sentence, Article 2.2.1.1, last sentence, footnote 6 and Article 2.4.1).

262 The European Communities appears to argue that the reference to Article 6 in Article 11.4 of the AD Agreement means that a "review" within the meaning of that provision necessarily entails an "investigation". The more logical conclusion, however, is that the fact that it was necessary to make express reference in Article 11 to Article 6 to make it applicable within the context of Article 11 confirms that a provision applicable to investigations is not per se applicable to reviews. We note that unlike Article 11, Article 9 does not contain a cross-reference to Article 6. We consider that this absence of a cross-reference may well reflect an understanding among negotiators that, in view of the particular nature of duty assessment proceedings, Article 6 cannot apply in its entirety to Article 9. It is not clear, for example that negotiators would necessarily have envisaged that, if an individual importer requests a refund of anti-dumping duties under Article 9.3.2, all "interested parties", as defined in Article 6.11, should be able to participate in the ensuing refund procedure and exercise the procedural rights provided for in the various provisions of Article 6.

263 EC-First Written Submission, para. 176; EC-Closing Oral Statement at the Second Substantive Meeting of the Panel, para. 8.

264 Appellate Body Report, US � Carbon Steel, paras. 58-97.

265 Appellate Body Report, US � Carbon Steel, paras. 66-67.

266 Appellate Body Report, US � Carbon Steel, para. 68 (footnote omitted, emphasis in original)

267 Appellate Body Report, US � Carbon Steel, footnote 58.

268 Appellate Body Report, US � Carbon Steel, para. 89.

269 Article 11 of the SCM Agreement corresponds to Article 5 of the AD Agreement.

270 Appellate Body Report, EC � Bed Linen (Article 21.5), para. 123 (italics in original).

271 Appellate Body Report, US � Carbon Steel, para. 87.

272 Appellate Body Report, US � Corrosion-Resistant Steel Sunset Review, para. 107 (emphasis in original). See also para. 124. The Panel in US � Anti-Dumping Measures on Oil Country Tubular Goods considered that the Appellate Body's reasoning regarding the differences in purpose between original investigations and sunset reviews "applies with equal force to the question of injury". Panel Report, US � Anti-Dumping Measures on Oil Country Tubular Goods, paras. 7.116-7.117.

273 Appellate Body Report, US � Oil Country Tubular Goods Sunset Reviews, para. 279.

274 Appellate Body Report, EC � Bed Linen, para. 62, footnote 30.

275 See also Appellate Body Report, EC � Bed Linen (Article 21.5), para. 124.

276 Panel Report, US � DRAMS, para. 6.90; Panel Report, Mexico � Anti-Dumping Measures on Rice, para. 7.144.

277 Appellate Body Report, US � Carbon Steel, para. 72; Appellate Body Report, US � Corrosion Resistant Steel Sunset Review, para. 152.

278 Appellate Body Report, US � Oil Country Tubular Goods Sunset Reviews, para. 294 (emphasis in original).

279 Appellate Body Report, US � Oil Country Tubular Goods Sunset Reviews, para. 301.

280 E.g., Appellate Body Report, US � Carbon Steel, paras. 66, 83 and 86-87; US � Corrosion Resistant Steel Sunset Review, paras107, 112, 124, 135 and 154; Appellate Body Report, US � Oil Country Tubular Goods Sunset Reviews, paras. 279, 284, 287-288, 290, 294, 296-301, 327-328 and 359.

281 In this respect, the fact that a number of WTO panel and Appellate Body reports and GATT panel reports have used the words "investigation", "investigatory" and "investigate" in connection with proceedings under Article 9.3 and Article 11 (EC-Rebuttal Submission, paras. 168-170; EC-Second Oral Statement, para. 36) is of no particular significance. The mere fact that a panel or the Appellate Body characterizes a review under Article 11 as a "review investigation" does not imply that it thereby expresses a view that provisions in the AD Agreement that apply to "investigations" are also applicable to "reviews". Similarly, we fail to see the relevance of the documents submitted by the European Communities that show that USDOC and USITC have sometimes characterized reviews as "investigations". (EC-Rebuttal Submission, paras. 132-133) The issue before us is how certain terms are used in the AD Agreement.

282 Panel Report, Argentina � Poultry Anti-Dumping Duties, para. 7.357 (emphasis added)

283 EC-Response to Panel Question 15, paras. 55-64.

284 The concept of "period of investigation" also appears in Article 9.5 of the AD Agreement.

285 See Panel Report, EC � Tube or Pipe Fittings, paras. 7.100-7.101; Panel Report, Mexico � Anti-Dumping Measures on Rice, paras. 7.50-7.65; Panel Report, Argentina � Poultry Anti-Dumping Duties, para. 7.287.

286 G/ADP/6 (16 May 2000).

287 The fact that drafts of the second sentence of what is now Article 2.4.2 used the concept of "period of investigation" when the first sentence already used "the investigation phase" confirms that the drafters understood these concepts to have different meanings. See e.g. Draft Working Paper on AD dated 26 November 1991, referred to as "New Zealand III, Ramsauer Text" in Exhibit EC-51.

288 The first three uses of the term (Article 2.2.1, and 2.2.1.1 and its footnote 6) relate specifically and exclusively to the determination of cost of production. Article 2.2.1 provides that prices that are below cost at the time of sale still allow for recovery of costs within a reasonable period of time if above weighted average unit costs for the POI. Footnote 6 to Article 2.2.1.1 may actually require reference to data subsequent to and thus outside the POI. The fourth use of the term relates to the issue of exchange rate movements in the context of export prices.

289 Appellate Body Report, US � Softwood Lumber V, para. 76, (footnote omitted, emphasis added).

290 See, e.g. EC-Response to Panel Question 15, paras. 72-82.

291 We note, in particular, the statement of the European Communities, in response to Panel Question 15, at para. 54: "Furthermore, before entering into a discussion of what the meaning of the words 'during the investigation phase' in Article 2.4.2 of the Anti-Dumping Agreement might be, the European Communities would like to stress that it has no particularly strong views on this matter in these proceedings, and that the outcome of these proceedings does not depend in any way whatsoever on the meaning that might eventually be attributed to those words. On the contrary, since it is the United States that has repeatedly asserted that the word 'investigation' in Article 2.4.2 has a special and limited or defined meaning, namely 'an investigation to determine the existence, degree and effect of any alleged dumping' within the meaning of Article 5.1 of the Anti-Dumping Agreement, it is for the United States to substantiate that assertion. Insofar as the United States has failed to substantiate that assertion � and the European Communities believes that to be incontrovertibly what has happened in these proceedings � that is an end of the matter. The European Communities claim under Article 9.3 and 2.4.2 must prevail. The meaning of the phrase 'during the investigation phase' does not need to be decided by this Panel." (Emphasis added).

292 We note that Article 31 of the Vienna Convention provides that the object and purpose of a treaty must be taken into account in establishing the ordinary meaning of the terms used therein. Since Article 31 refers to "the object and purpose" of the treaty and not of its individual provisions, the argument of the European Communities regarding Article 9.3 might be better characterized as a further contextual argument rather than an argument relating to object and purpose. We further note that since the AD Agreement contains no discrete statement of objectives, one can only derive or deduce its objectives from the operational provisions of the Agreement. While it is perhaps possible at a very high level of generality to deduce from the operational provisions of the AD Agreement as a whole that for instance, one of the "objectives" of the AD Agreement is to provide a multilaterally agreed framework of rules governing actions against injurious dumping, claims of more specific objectives are difficult to discern with any facility or compelling force due to the lack of anything that could properly be described as constituting a clear statement of the objectives of the AD Agreement. In this regard, we note that the European Communities refers to "the object and purpose of the AD Agreement regarding the consistent application of basic economic concepts" in the measurement of international price discrimination between two markets. The precise meaning the European Communities ascribes to these concepts and the manner in which the European Communities derived them from the text of the AD Agreement are unclear.

293 We note in this respect the discussion of Article 9.3 in para. 7.355 of the panel report in Argentina � Poultry Anti-Dumping Duties.

294 Panel Report, Argentina - Poultry Anti-Dumping Duties, para. 7.359: "In our view, Article 9.4(ii) is describing the use of variable anti-dumping duties, which are calculated by comparing actual (i.e., at the time of collection) export price with a prospective normal value. Since Article 9.4(ii) expressly envisages the imposition of variable AD duties, there is no basis for us to find that Argentina's recourse to variable duties (calculated on the basis of "minimum export prices" used as prospective normal values) is necessarily inconsistent with Article 9.3 of the AD Agreement." The European Communities has explicitly recognized the permissibility of such a system of variable anti-dumping duties. E.g., EC-Response to Panel Question 4, para. 29 ("Variable duties are foreseen in Article 9.4.")

295 One aspect of this is that refund procedures in prospective duty assessment systems often occur much less frequently than administrative reviews in retrospective systems.

296 We are not thereby suggesting that it would necessarily have been illogical, in light of the object and purpose of Article 9.3, for the negotiators of the AD Agreement to have decided to extend the application of Article 2.4.2 to duty assessment proceedings under Article 9.3.

297 Exhibit EC-55.

298 EC-Rebuttal Submission, para. 218.

299 Appellate Body Report, US � Gambling, para. 191-192.

300 Of the 105 Members that have made notifications under Article 18.5 of the AD Agreement, 29 Members have indicated that they have no anti-dumping legislation. Report (2004) of the Committee on AD Practices, G/L/707 (4 November 2004), Annex A.

301 EC-Rebuttal Submission, para. 218. The European Communities argues that "if the words 'during the investigation phase' are not simply omitted from such implementing legislation, there are numerous instances where they are rendered as 'during the investigation period.'" We note that legislation notified by India and Tunisia does use "during the investigation phase". G/ADP/N/1/IND/1, p. 22 and G/ADP/N/1/TUN/2, p. 20. We also note that in some of the cases in respect of which Exhibit EC-55 notes that no reference is made to "during the investigation phase", the provision in question is of a rather general character. For example the European Communities notes that Article 4 of Indonesia's' Law Number 10, 1995 does not contain the phrase "during the investigation phase". Article 4 simply states that "(1) the amount of anti-dumping duties referred to in Article 2 shall be as high as that of dumping margins". G/ADP/N/1/IDN/2, p. 12.

302 Vienna Convention, Article 32(a) and (b).

303 The documentation contained in Exhibits EC-48-50 consists mainly of formal documents circulated in the Uruguay Round Negotiating Group on MTN Agreements and Arrangements in the period 1987- July 1990, including proposals by delegations, Secretariat notes, notes on meetings and the first draft Chairman's text circulated in July 1990. Exhibit EC-51 reproduces the successive versions of what is now Article 2.4.2 in what the European Communities refers to as" Carlisle I Text" , "Carlisle II Text", "New Zealand I Text", "New Zealand II Text", "New Zealand III Ramsauer Text" and "Dunkel Draft".

304 EC-Rebuttal Submission, para. 209.

305 Supra, footnote 202.