1. Background
The EC is bound in its tariff schedule not to apply a duty rate in excess of 15.4% to ‘Meat and edible meat offal, salted, in brine, dried, or smoked’ (Tariff item 02.10). The complainants, Brazil and Thailand, had been exporting frozen chicken cuts treated with salt to the EC, since, respectively, 1996 and 1998. Between 1996 and 2002, EC customs points generally, although not always and everywhere, applied the bound rate in tariff item 02.10 to these imports; beginning in 2002, however, the Commission issued regulations requiring that customs authorities classify imports of meat under 02.10 only where salt has been added to the meat for the purpose of long-term preservation. The result of this regulatory action was that the exports of the claimants were reclassified as falling under tariff item 02.07, which applies to ‘fresh, chilled or frozen’ poultry; the bound rate under 02.07 is higher than under 02.10 (102.4 Euros per 100 kgs, an effective ad valorem rate of between 40 and 60%). As well, products under 02.07 may face special safeguard action pursuant to the WTO Agreement on Agriculture. Because the exports of the complainants were not salted for purposes of long-term preservation, European customs authorities were thus required to apply the higher tariff rate under 02.07 to these products. The complainants argued that through its binding under 02.10, the EC was obligated to apply a duty that did not exceed 15.4% ad valorem to their exports, regardless of whether the salt was applied for purposes of long-term preservation. By applying a higher rate of duty since 2002, the EC and its member-state customs authorities were, according to the complainants, in violation of Art. II:1(a) and II:1(b) GATT, which prohibit WTO Members from applying duties and charges higher than the bound rates in their schedules for the product in question.
The disputing parties agreed that: (1) the exports in question were not salted for long-term preservation; (2) as explained above, the tariff rate applied to the products in question was, during the period of time that was the subject of the complaint, always higher than the bound rate under 02.10; (3) if the products in question did not fall within 02.10, the EC would not be in violation of Article II in applying the higher rate prescribed by 02.07; (4) in determining whether the products in question fell under 02.10, the EC could specify a minimum quantitative threshold in ascertaining whether the products were ‘salted’ (historically, in EC Regulations, the minimum threshold had been 1.2%).
Thus, the only issue in this dispute, as framed by the parties, was whether or not the term ‘salted’ in 02.10 means, exclusively, ‘salted for purposes of long-term preservation’.
The parties, the Panel, and the Appellate Body (AB) all dealt with this issue as a question of treaty interpretation: the central task was to decide the meaning of ‘salted’ in accordance with the canons of treaty interpretation to be found in Articles 31 and 32 of the Vienna Convention on the Law of Treaties (VCLT). Of course, the word ‘salted’ does not appear in any WTO treaty, and the document being interpreted here is a Harmonized System (HS) classification found in the defending member's schedule. The obligation to follow the classification practice of the HS is an obligation of WTO Members in their capacity as members of the World Customs Organization (WCO), which develops the HS. Although the Panel solicited advice from the WCO on the interpretation of 02.07 and 02.10, neither the Panel nor the AB suggested that the WTO dispute settlement organs must, as a matter of law, defer to WCO interpretations or practice with respect to HS classifications. This is consistent with earlier AB jurisprudence that considers Members' schedules as forming an integral part of the Single Undertaking: the terminology of schedules should be interpreted and applied as if one is interpreting and applying the GATT itself or any other treaty in the Single Undertaking (EC – Lan Equipment, Report of the AB, para. 84). The WCO Secretariat, however, took the view that it would be appropriate for the parties to resort to the WCO dispute settlement system and obtain a ruling on the meaning of ‘salted’ in 02.107 before having recourse to WTO procedures. The Panel concluded that there was no possibility of declining jurisdiction even temporarily under the DSU. The Panel therefore proceeded to determine whether, when read in accordance with VCLT Arts. 31 and 32, ‘salted’ in HS 02.10 means, exclusively, ‘salted for purposes of long-term preservation’.
The Panel interpreted the norms in Arts. 31 and 32 of the VCLT as intended to be applied sequentially: thus, it began with ‘ordinary meaning’ in Art. 31(1), and, having found that the ordinary meaning of ‘salted’ did not exclusively refer to salting for long-term preservation, then proceeded to examine whether, in light of the other interpretative considerations in Art. 31 VCLT (object, purpose, and context), this was the proper meaning to attribute to tariff item 02.10. While, as discussed below, it went on to correct some specific aspects of the Panel's interpretation and application to the VCLT, the AB upheld its general approach but also noted that it is appropriate to apply the VCLT in a holistic manner and that the factual considerations in question could be taken into account either in considering ‘ordinary meaning’ or in considering ‘context’ (para. 176): here the AB was responding to a claim that the Panel had violated its own sequential methodology in going beyond the dictionary and taking into account the factual context in its determination of ‘ordinary meaning’.
In what follows, we will discuss the adjudicating bodies' treatment of the burden of proof (Section 2), ‘ordinary meaning of words’ (Section 3), ‘context’ (Section 4), the WCO General Rules (Section 5), ‘object and purpose’ (Section 6), ‘subsequent practice’ (Section 7), and ‘supplementary means’ of treaty interpretation (Section 8). We will also discuss the choice of forum for the resolution of this dispute (Section 9). We conclude by summarizing our views concerning the extent to which the outcome of the dispute was correct (Section 10).
2. The burden of proof
According to the complainants in EC–Chicken Cuts, the EC violated Art. II:1(a) and II:1(b) GATT by according a less favorable treatment than that provided for in the EC schedule. A threshold question is what the complainant is required to prove in order to establish a violation of Article II based on the scheduling practices of the defendant. The Panel interprets the appropriate burden of proof as follows:
7.78 … In light of the rules governing the burden of proof, it is our view that the complainants bear the burden to prove that the products at issue are, in fact, covered by the concession contained in heading 02.10 of the EC Schedule.
In our view, there is a fundamental problem with this definition of the burden of proof, and this is a point we will return to repeatedly: we believe that this is a too-low standard for a finding of a violation, unless ‘covered by’ is read as ‘only covered by’. Due to the incomplete nature of the EC schedule (as of all such schedules), it may well be the case that several of its headings can cover the product at issue. In such a case, it would be inappropriate to require of the importing country to provide the most favorable treatment, without any rationale for why this should be the default case. On the contrary, we would argue that it should not be viewed as a violation of the agreement if the importing country applies a tariff heading that is shown to cover the product, even if there is another heading that would also fit the product.
In our view, in order for a classification under 02.07 to amount to a violation of the agreement, it would have to be incorrect to classify the product at issue under 02.07. The investigation by the adjudicating bodies does not address the appropriateness of classifying under this heading however; for instance, the VCLT canons of treaty interpretation are not employed to examine this heading. The focus in the reports is instead on heading 02.10, and the conclusion is only that this heading is not exclusively for chicken cuts preserved through salting. A leap is thus taken from the finding that the product is classifiable under heading 02.10 to the conclusion that it must not be classified under heading 02.07. We believe that there is, in this regard, insufficient evidence for a finding of violation.
3. Ordinary meaning of words
The Panel finds no indication in the ordinary meaning of the words in heading 02.10 that the product at issue would not be covered:
7.151 The Panel considers that there is nothing in the range of meanings comprising the ordinary meaning of the term ‘salted’ that indicates that chicken to which salt has been added is not covered by the concession contained in heading 02.10 of the EC Schedule.
The AB in essence agrees with the Panel's conclusion:
187 … we see no reason to disturb the Panel's conclusion concerning the ordinary meaning of the term ‘salted’.
We have little objection to the conclusion as such, that the words in heading 02.10 could be read so as to include the product at issue. It is notable, however, that – although at times the issue is framed as whether the products in question are correctly classified under 02.10 or 02.07 – the Panel's and AB's analysis proceeds not by analyzing under the VCLT the ‘ordinary meaning’ of both 02.10 and 02.07 but rather only bringing in 02.10 as a matter of context. This approach excludes from the start the possibility that a particular product could have characteristics such that it would fall within the ‘ordinary meaning’ of more than one HS classification. This possibility is, however, explicitly contemplated in WCO law, which contains rules to deal with just this situation (see Section 4). The Panel notes in para. 7.238 of its report that the disputing parties agreed that these rules did not apply to the case at hand because ‘the products at issue in this dispute are not prima facie classifiable under two or more headings’. This approach of only focusing on heading 02.10 is followed by the adjudicating bodies throughout their analyses.
4. The meaning of ‘salted’ considered in its context
In their discussion of ‘context’, the adjudicating bodies deal with three main issues: what constitutes the context for interpreting ‘salted’ in heading 02.10, the meaning of ‘salted’ considered in its context, and the relevance of General Rule 3 for the interpretation of the HS system. We will in this Section discuss the second issue, and in the next Section the applicability of General Rule 3.
4.1 The adjudicating bodies' findings
The Panel starts its examination of context by considering the text of the EC schedule. What it finds, and this is the pattern throughout the analysis of context, is that the text of the schedule does not provide any information, and, in particular, that it does not confirm a long-term-preservation criterion:
7.180 In summary, the Panel concludes that the definitions of the terms in the concession contained in heading 02.10 of the EC Schedule other than ‘salted’ do not indicate any intrinsic notion that characterizes all the terms in that concession other than that they are in a state that is not simply fresh, chilled or frozen. Nor does the Panel consider that these terms can be defined as pertaining exclusively to ‘preparation’ or to ‘preservation’ processes. In addition, the Panel does not consider that any inferences that are useful for the purposes of our interpretation of the concession contained in heading 02.10 of the EC Schedule can be drawn from the structure of Chapter 2 of the EC Schedule nor from other parts of the EC Schedule other than the fact that they do not indicate that Chapter 2, including heading 02.10, is necessarily characterized by the notion of long-term preservation. In conclusion, it is the Panel's view that an examination of these various aspects of the EC Schedule does not clarify the ordinary meaning of the term ‘salted’ in the concession contained in heading 02.10 of the EC Schedule. (Underlining added)
The Panel then turns to the terms and structure of the HS, the Explanatory Notes to the HS, and other parties' schedules, with similar results:
7.205 … Therefore, we consider that the evolution of the terms and structure of Chapter 2 of the HS does not clarify the ordinary meaning of the concession contained in heading 02.10 of the EC Schedule. Further, it is the Panel's view that the terms and structure of the HS do not indicate that the concession contained in heading 02.10 is necessarily characterized by the notion of long-term preservation. (Underlining added)
7.223 … we consider that the Explanatory Notes to the HS do not clarify the ordinary meaning of the term ‘salted’ in the concession contained in heading 02.10 of the EC Schedule. Further, it is our view that the Explanatory Notes do not indicate that that concession is necessarily characterized by the notion of long-term preservation. (Footnote omitted, underlining added)
7.244 To the extent that the terms of the relevant concessions in other WTO Members' schedules are identical to the terms of the concession contained in heading 02.10 of the EC Schedule and of the HS, we do not consider that they can assist us any further in the analysis we have undertaken thus far. (Underlining added)
The AB frames the significance of ‘context’ in this dispute in the following terms:
209. Therefore, we need to determine whether the context of the term ‘salted’ – or other elements of the customary rules of treaty interpretation – require or permit a reading of the term ‘salted’ in heading 02.10 of the EC Schedule more narrowly than the ordinary meaning of that term suggests; that is to say, that the customary rules of treaty interpretation other than ‘ordinary meaning’ indicate that ‘salting’ under heading 02.10 contemplates exclusively the notion of ‘preservation’.
The AB first examines the words in heading 02.10 that describe treatments other than salting. Even if disagreeing with the Panel's view that ‘in brine’ excludes preservation, the AB basically agrees with the Panel's findings that preservation is not a requirement for a product to be ‘dried, in brine, and smoked’:
212 … We note that the dictionary meaning of the term ‘to dry’ is, in relevant part, ‘to remove the moisture from by wiping, evaporation, draining; preserve (food, etc.) by the removal of its natural moisture’; in turn, the dictionary meaning of the term ‘to smoke’ is to ‘dry or cure (meat, fish, etc.) by exposure to smoke’. The ordinary meanings of these terms suggest that the relevant processes can be applied to meat in various ways and degrees of intensity, thereby producing different effects on the meat, effects that may or may not place the meat in a state of ‘preservation’. Nor are we persuaded by the European Communities' argument that the terms ‘dried’ and ‘smoked’, in the present context, ‘concern [exclusively] means to preserve’. It is clear from the evidence on the record that, while the processes mentioned in heading 02.10 – ‘salted, dried, in brine and smoked’ – may include the notion of ‘preservation’, these processes are also used extensively to confer special characteristics on meat products. Similar reasoning may also be valid with respect to the term ‘smoked’. (Footnotes omitted)
The AB then goes on to consider whether the structure of Chapter 2 of the EC Schedule and the HS support a reading of heading 02.10 as referring exclusively to processes of preservation.
217. We note that heading 02.10 does not make reference to refrigeration. By contrast, other headings of Chapter 2 – that is, headings 02.01 to 02.09 – refer to freezing and chilling. The European Communities argues that (i) this circumstance implies that refrigeration is of ‘little or no importance’ for heading 02.10, and that (ii) the reason for this is that these products are ‘preserved’ by the processes mentioned in heading 02.10. The European Communities uses this argument to support its view that heading 02.10 covers exclusively meats that have been ‘preserved’ by the processes referred to in that heading. (Footnote omitted)
218 … In our view, whether a product has been frozen or not will not influence whether that product falls under heading 02.10 … [But] it does not follow from the absence of refrigeration in the text of heading 02.10 that the processes referred to in heading 02.10 must necessarily place the meat in a state of ‘preservation’. (Underlining added)
The AB also examines Chapter and Explanatory Notes to the HS, and while criticizing the Panel on certain scores basically arrives at the same overall conclusion:
229 … we conclude that the Harmonized System and the relevant Chapter and Explanatory Notes thereto do not support the view that heading 02.10 is characterised exclusively by the concept of preservation. Furthermore, the term ‘salted’ in heading 02.10, when considered in its context, suggests that meat to which salt has been added, so that its character has been altered, will be ‘salted’ within the meaning of heading 02.10, even if such salting does not place such meat in a state of ‘preservation’. Heading 02.10 of the Harmonized System, read in its context, suggests that it is neither limited to, nor excludes, meat that is ‘prepared’ by salting or that has been ‘preserved’ by salting. Specifically, for resolving this dispute, heading 02.10 does not contain a requirement that salting must, by itself, ensure ‘preservation’.
As a result, the AB upholds the Panel's findings that the context of the term ‘salted’ in heading 02.10 does not exclusively refer to long-term preservation.
4.2 Discussion
The adjudicating bodies discuss context separately from object and purpose. First, in the discussion of context, the adjudicating bodies look at the methods of treating meat other than salting under heading 02.10 – in brine, drying, smoking – to determine whether there is a common element of long-term preservation reflected in this choice of treatments of meat. The adjudicating bodies then examine whether the EC classification principle can be implemented in a manner that is consistent with the object and purpose of the treaty.
It can be noted that the Panel and the AB fail to see any logic in the grouping of ‘salted’ with the other criteria that are listed in heading 02.10. For instance, the Panel states:
7.180 … In conclusion, it is the Panel's view that an examination of these various aspects of the EC Schedule does not clarify the ordinary meaning of the term ‘salted’ in the concession contained in heading 02.10 of the EC Schedule.
We find it somewhat odd that the adjudicating bodies are content with this, having decided not to rely more heavily on advice by the WCO with regard to the classification issue. Should not their inability to detect logic to the grouping of these various kinds of treatments have impelled them to look deeper for the rationale behind the structure of the HS in this regard?
From an economic point of view, the customs classification of the product at issue should be determined in light of what the parties seek to achieve through the agreement. Such an analysis should in particular take account of the reason why the parties agreed to have different tariffs for headings 02.07 and 02.10, and how the classification affects the achievement of this purpose. Admittedly, there is no economic theory that directly deals with the principles of customs classification, at least to the best of our knowledge. But we believe that a discussion of the economic relationship between the product at issue and those falling under the two headings could still have shed some light on how to classify the chicken cuts. We are not able to undertake an analysis of the appropriate customs classification here, however, but we will just very briefly discuss some aspects that we believe would likely arise in a more thorough investigation.
In order to discuss the appropriate tariff classification of the chicken cuts, we first need to determine why the parties chose to have different tariffs under the two headings. This, in turn, requires an explanation of why the parties found it suitable to allow tariffs at all. There are, generally speaking, several reasons why an agreement may allow tariffs. For instance, for some countries tariffs may be important sources of government revenue. In the case of the EC, however, the most plausible reason seems to be that the tariffs under the two headings are meant to protect local production, and thereby indirectly the incomes and employment of certain groups of producers of competing chicken cuts. The negotiating stance of the EC suggests that the EC finds the benefit from this protection to outweigh the cost in terms of, for example, consumer welfare in the EC. Against this motive for positive tariffs on certain products stand, of course, the interests of exporters to the EC, who would prefer to compete with local EC producers without the disadvantage of tariffs. But despite the latter adverse effects, the parties agreed to allow for positive tariffs. This illustrates the more general point that it may be efficient to let an agreement feature positive tariffs on certain products, if the gain to the importing country is sufficiently large relative to the costs to the exporters.Footnote 1, Footnote 2
The issue of customs classification arises due to two fundamental circumstances. First, there is no perfect, one-to-one match between each imported product and a customs classification heading. There are probably several reasons why classification schemes such as the HS do not have such a fine distinction between products. One obvious reason is, of course, that it would be enormously costly to maintain such a system. To start with, it would require extremely elaborate product descriptions, including an uncountable number of tariff lines. Also, the negotiations would be tremendously complicated if one were to use the flexibility that such a system would permit. If for no other reasons, actual trade agreements use a much courser classification scheme.
Second, for a conflict over customs classification to arise, there needs to be a difference across tariff lines in the tariffs they stipulate (or in some other factor affecting the conditions for importation). Again, one can think of several reasons why this would be the case. It may be explained by the fact that imports under the two headings affect different domestic constituencies, and these differ in the extent to which they are harmed by the imports, or in their political clout. Another reason, which we perhaps find more plausible on the basis of our admittedly very limited understanding of the factual circumstances in the dispute, is that the tariffs under both headings are meant to protect essentially the same domestic interests, but that products with the two types of characteristics pose in different ways strong competitive threats to the interests the tariffs are to protect.Footnote 3
In the case of the product at issue in Chicken Cuts, we note first that it seems plausible that the cost of salting constitutes a very small proportion of total production costs, even when the salt suffices for ‘long-term’ preservation. Hence, such salting will not drive up exporters' costs enough to shield EC producers from competition to any considerable degree. Nor does the salting imply that the exporters have introduced a new, lower-cost production technique. If this were the case, it might have been efficient to let the exports enter under the lower tariff, despite the resulting damage to certain EC producers, since the exporters would then be able to compensate the EC for the damage, and still gain from the lower tariff treatment. There thus does not seem to be a cost-based reason for giving the salted chicken cuts a more favorable tariff treatment than the unsalted ones.
Turning to the demand side, if salt is added to improve the taste, the demand for the product would presumably be higher than it would be absent the salting. The product at issue would thus be a more severe competitive threat than the corresponding product under heading 02.07 due to the improved taste, and should, if anything, face a higher tariff in order to protect local producers.Footnote 4 This reasoning would hence suggest that to the extent the product at issue is only salted for seasoning purposes, there is a presumption that it should go under heading 02.07.
A case for a lower tariff on the product at issue than the one provided for under 02.07 would thus have to be based on it having a disadvantage on the demand side relative to the chicken cuts coming in under 02.07. We are not certain here about the factual aspects, but we assume that salting chicken enough so that it withstands transportation, storage, etc., will reduce the attractiveness of the product to buyers, and will thus tend to reduce the demand. If so, chicken cuts that are salted for preservation purposes would constitute less of a competitive threat to EC producers, thus providing one possible explanation of why a more favorable tariff treatment is appropriate. We can here note the statement by the EC, as summarized by the Panel, that:
7.310 … [The EC] satisfied itself with significantly lower tariffs on salted products because it was aware that few products were traded on an international basis under that heading … had it imagined that frozen chicken with added salt could have been classified under heading 02.10, it would never have set such a low tariff. (Footnote omitted)
It thus does not seem inconceivable that one could erect a plausible demand-based argument for treating chicken cuts that are salted for long-term preservation more favorably than chicken cuts that contain less salt. This would seem to fit with the difference in tariff levels between headings 02.07 and 02.10, and would thus tend to support the EC claim.
Can a similar logic be found with regard to the other characteristics under these headings – ‘in brine’, ‘dried’, and ‘smoked’? Here we have to be even more speculative. But it seems unlikely that putting meat ‘in brine’ improves the demand for the meat, if anything the opposite. Hence, by the same logic, a lower tariff may be warranted. The ‘dried’ and ‘smoked’ attributes are more complicated. On the one hand, both these types of products have a more limited range of use, and the production also tends to be costly. Both these factors suggest that products coming under this heading pose a more limited competitive threat compared to the products that are just frozen, chilled, or fresh, thus reinforcing the picture given above. On the other hand, the processes of smoking and drying may, in some cases, increase the value of the products to consumers, seemingly in contrast to the distinction described above.
Note that the point of the discussion here is not to advocate any particular interpretation of the EC schedule – this would require a much more thorough investigation than we have been able to undertake. The point is instead to explain why we believe that when examining the appropriate customs classification of this product, the adjudicating bodies should somewhere along the line have confronted the basic questions of (i) why did the parties agree to make the tariff distinctions with regard to the product characteristics listed under the two headings?, and (ii) how should the product at issue be seen from this perspective? After all, the agreement gives a form of legal protection or sanction to the use of tariffs, and this must be due to the fact that the Members have found the use of tariffs valuable enough to explicitly agree to allow for their use. In order to determine whether the product at issue should be classified under one or the other HS heading, it would therefore be natural to consider the role the tariffs play, why they differ across headings, and how the classification in this case may help achieve this objective while taking into account exporter interests. We see little consideration of this in the present dispute.
It should also be noted that a consequence of the adjudicating bodies' failure to take into consideration what objective the higher tariff on products entering under heading 02.07 was meant to achieve, is that the ‘salted’ criterion for the lower tariff treatment has been rendered essentially meaningless: The interpretation of ‘salted’ endorsed by the Appellate Body would, in theory at least, compel the European authorities to provide a significantly lower rate of duty to any meat that has been treated with salt, even minimally. Consequently, an exporter who wants to compete in the frozen-chicken market only has to find a way to treat the product with salt without affecting its suitability for that market in order to circumvent the higher protection afforded by Europe to domestic industries competing in that market. We are not suggesting that we have actual evidence of such behavior by Brazilian and Thai exporters, just that the possibility for such behavior now seems to exist.
Finally, we note that the limited number of questions the Panel asks the WCO concerning customs classification practices are not very helpful for resolving the issue. This is partly due to the way the questions are formulated, and partly to the answers provided. The question that most directly concerns the principle behind the customs classification system is the following:
What is the rationale behind the product coverage of the 4-digit heading and the 6-digit-heading levels in the Harmonized System?
The WCO responds by first giving a few details about earlier nomenclatures that affected the structure of the HS. The only information provided by the WCO that more directly answers the question is the following statement, which refers to these sources of the HS:
Consequently, the rationale behind the product coverage in the HS was to meet the needs of those involved in international trade by including goods or groups of goods with a significant volume of international trade, taking into consideration the structure of the nomenclatures consulted. (Panel Report p. C-135, underlining added and footnote omitted)
We cannot determine the correctness of this statement. But it does seem unlikely to exhaust the issue. Also, as a marginal note, the method of including products for which there is a large volume of trade would tend to exclude from the system products with very high tariffs (and consequently little trade). Was this really what happened?
5. WCO General Rule 3
We have already alluded to the fact that there are rules in the WCO to deal with certain problems of application and interpretation of HS classifications – the ‘General Rules for Interpretation of the Harmonized System’. In their discussion of context, the adjudicating bodies address the applicability of these rules in the instant dispute, and, in particular, General Rule 3, which addresses situations where goods are prima facie classifiable under two headings. The latter states in pertinent parts:
When by application of Rule 2(b) or for any other reason, goods are, prima facie, classifiable under two or more headings, classification shall be effected as follows:
(a) The heading which provides the most specific description shall be preferred to headings providing a more general description. However, when two or more headings each refer to part only of the materials or substances contained in mixed or composite goods or to part only of the items in a set put up for retail sale, those headings are to be regarded as equally specific in relation to those goods, even if one of them gives a more complete or precise description of the goods.
…
(c) When goods cannot be classified by reference to 3(a) or 3(b), they shall be classified under the heading which occurs last in numerical order among those which equally merit consideration.
5.1 The adjudicating bodies' findings
According to the Panel, this rule is inapplicable in the instant dispute, despite the contradictory claims by the parties concerning the correct classification. The Panel here relies on the arguments of the parties:
7.227 With respect to General Rule 3, Brazil, Thailand and the European Communities concur that the condition for its application – namely, that the products at issue are prima facie classifiable under two or more headings – has not been fulfilled in this case. The complainants consider that the products at issue are classifiable under heading 02.10 whereas the European Communities considers that the products at issue are classifiable under heading 02.07. (Footnote omitted)
As a consequence, General Rule 3 should not be applied:
7.238 As regards the question of whether or not General Rule 3 applies, all the parties appear to be in agreement that a textual and contextual analysis of the relevant headings indicates that the products at issue in this dispute are not prima facie classifiable under two or more headings. Accordingly, we will proceed on the same assumption with the result that we will not apply General Rule 3.…
The AB's entire treatment of General Rule 3 is the following:
231. We now turn to consider the General Rules for the Interpretation of the Harmonized System (the ‘General Rules’). We recall that the Panel found that:
… all the parties appear to be in agreement that a textual and contextual analysis of the relevant headings indicates that the products at issue in this dispute are not prima facie classifiable under two or more headings. Accordingly, we will proceed on the same assumption with the result that we will not apply General Rule 3. Given our conclusion that General Rule 3 is inapplicable, we do not consider it necessary to address the various arguments that have been advanced by the parties regarding that Rule. (Footnote omitted)
232. Brazil and Thailand appeal this finding. Both argue that the Panel incorrectly found that the parties were in agreement that General Rule 3 was inapplicable in this case. Brazil also requests that the Appellate Body complete the analysis and find that the products at issue are classifiable under heading 02.10 by virtue of General Rule 3(a) or by virtue of General Rule 3(c). (Footnote omitted)
233. We note that the General Rules are, by their very name, rules for the interpretation of the Harmonized System. Specifically, General Rule 3 deals with the question of classification in circumstances in which goods are prima facie ‘classifiable’ under two or more headings.
234. We recall that the task of the Panel, as well as of the Appellate Body upon appeal, is to determine whether the European Communities has acted consistently with Article II:1(a) and with Article II:1(b) of the GATT 1994 with respect to the products at issue. Therefore, in our view, the primary task of the Panel, as well as of the Appellate Body, is to determine the meaning and scope of the concession contained in heading 02.10 of the EC Schedule. In our view, it is only after properly determining the meaning and scope of the tariff commitment in heading 02.10 that the question whether the products at issue are prima facie classifiable under two or more headings can arise. General Rule 3 is relevant in this case only for the second step, namely, under which heading a product is properly classified. It is therefore not necessary for us to consider, at this stage, General Rule 3. (Footnote omitted, underlining added)
Having stated this, the AB does not find it necessary to get back to this issue in its report.
5.2 Discussion
There are several aspects of the adjudicating bodies' reasoning in this context that we find puzzling.
A first difficulty we have is the fact that the Panel effectively reads into the term prima facie a requirement of a shared perception of the situation among the parties. If this was the intention behind the rule, presumably this would have been made explicit by the drafters of the HS. Also, such a reading significantly reduces the ambit of the rule, since there will presumably be relatively few instances where there is such an agreement. Moreover, it removes from the arsenal a main instrument for resolving a conflict over classification.
This is not to say that we believe that General Rule 3 provides a good way of resolving this kind of dispute, only that it seems applicable, given the situation. It is true that not all Members of the WTO are bound by the WCO treaty or are members of the World Customs Organization (WCO); however, all WTO Members have used the HS of the WCO in tariff negotiations and the resultant schedules, which are bound under the GATT. Would it not be the common expectation of all WTO Members, therefore, that the meaning of the schedules in question be decided through the application of rules governing the HS itself and the practice of the international organization that is the guardian of the HS?Footnote 5 The Panel accepts that the HS rules and related WCO practice are relevant to resolving the dispute but for some reason refuses to provide a clear legal theory as to why they are relevant, refusing to indicate whether the HS is ‘context’ within the meaning of 31(2)(b) of the VCLT or rather a relevant rule of international law ‘applicable between the parties’ within the meaning of 31(3)(c).
Second, the dismissal by the Panel of the General Rule is, in some sense, internally consistent, even if, in our view, based on erroneous reasoning. We find the approach by the AB more peculiar, however. The AB does not exclude the applicability of the rule, but instead says that it first has to address whether the product fits under heading 02.10. On this logic, the Panel having found the products at issue indeed to be classifiable under 02.10 would have had to go on and examine whether they were also classifiable under 02.07, and if the answer were also positive, at that point the Panel would need to apply the General Rule. Oddly, although accepting (unlike the Panel) the possibility that the Rule may be applicable and upholding the Panel's finding that the products are classifiable under 02.10, the AB does not fault the Panel for not proceeding to the next logical step in ascertaining whether to apply the rule, namely considering whether the products are also classifiable under 02.07. We simply cannot understand how the Panel could have made a proper determination of whether to apply the rule unless it proceeded to that second step. In sum, while we agree with the AB's finding that the rule may be applicable, we see incoherence or contradiction in the AB's failure to correct the Panel for not taking all necessary steps to determine whether the conditions for applying the rule existed in the case at hand.
The AB does not explain why it does not evaluate the possibility that the product at issue is classifiable under heading 02.07. The AB does argue, in footnote 443, that the interpretation of the concept ‘prima facie’ is unresolved:
We note that General Rule 3 refers to circumstances in which a product is ‘prima facie’ classifiable under two or more headings. In this respect, nothing on the Panel record indicates how the term ‘prima facie’ has been interpreted by the WCO's Harmonized System Committee, or the WCO itself.
But are there not sufficient indications in the dispute that the appropriate classification of the product at issue is unclear, and that therefore one need consider whether, prima facie, both headings could apply? For instance, the AB continues the same footnote by providing evidence suggesting that the classification of the product at hand is an unresolved issue:
However, according to information provided by the WCO Secretariat to the Panel, the Harmonized System Committee has not, so far, considered classification issues related to headings 02.07 and 02.10. (WCO's response to Question 8 posed by the Panel, Panel Reports, p. C-139) Furthermore, the WCO did not provide specific guidance to the Panel as to the meaning of the term ‘specific’, so as to determine the ‘most specific description’ within the meaning of General Rule 3(a) for purposes of the present dispute; according to the WCO, heading 02.07 could be considered more ‘specific’ due to the term ‘poultry’, but, at the same time, heading 02.10 could be considered more ‘specific’ by virtue of the term ‘salted’. (Panel Reports, para. 7.235) We also note that the participants do not argue that General Rule 3(b) is applicable to the dispute at hand. (Underlining added)
Furthermore, the WCO argues in the above-quoted footnote that the classification is not self-evident, at least not as it concerns the question of the degree of specificity in the two headings. To quote the Panel report:
7.235 The WCO states that, for the purposes of General Rule 3(a), it could be argued that heading 02.07 of the HS provides the more ‘specific’ description since it refers to ‘meat of poultry, frozen’, whereas heading 02.10 of the HS refers to ‘meat, salted’, in general. However, the WCO states that it could also be argued that the reference to the specific type of meat (poultry) in heading 02.07 should not be taken into consideration since it is the processing (freezing and salting) that matters when determining the classification of the products at issue in this case, giving rise to a possibility of heading 02.10 providing a more specific description. (Footnote omitted)
But could it not be argued that if a main tie-breaker for customs classification issues – General Rule 3(a) – cannot resolve the question concerning classification, there is at least prima facie a possibility that the product could be classified under two separate headings?
This goes to a basic question of the nature of state responsibility under Article II of the GATT, where the defending Member's customs classification practices are at issue. It is arguable that where more than one possible classification is reasonable, and WCO rules and practice do not dictate the use of one or the other, a WTO Member has some margin of discretion for determining which classification to apply to a particular product. In other words, the nature of the obligation in Article II, inasmuch as it applies to customs classification practices by individual Members, might well be to classify products in a reasonable, objective, and nondiscriminatory manner, consistent with WCO rules and practice. The Panel and AB approach to Article II suggests that there must be a single, correct way in which to classify every product in terms of the HS classifications and that the Article II obligation implies that a WTO Member must follow this single, correct way. This implicit view of state responsibility – a crucial issue that is not addressed explicitly by either the Panel or the AB – seems at odds with the very nature of the classification problem pointed out in Section 4.2 above: namely that for transaction-cost reasons it is impossible to have a separate customs classification number for each conceivable product that countries may want to identify. Instead, classification schemes will always leave ample room for groups of different products to fall under the same classification number – the classification scheme will, in this sense, always be ‘coarse’. This inevitable coarseness implies that any classification scheme will be plagued by many ambiguities in terms of whether a particular ‘variant’ is to be slotted into one (relatively crude or fine) classification or another.
In deciding what is the prima facie case that the complainant must make in a case such as this in order to discharge its burden of proof, it is necessary to take some view as to the nature of state responsibility. Our analysis of the basic nature of the problem of customs classification – the inevitable coarseness of such schemes – would suggest that, in this instance, the prima facie case must be that the EC choice between 02.10 and 02.07 is unreasonable, discriminatory, contrary to WCO rules and practice, or otherwise not in accord with the common intentions of the parties and/or the WTO Membership in agreeing to schedules. If General Rule 3 could be applied so as to determine the clear answer as to which of the two headings is to be applied, then arguably the EC was required to so apply the rule; thus, one way of the complainant discharging its burden of proof would be to show that a proper application of General Rule 3 would yield a choice of classification different from the one adopted by the EC.
6. Object and purpose
Having concluded that neither the ordinary meanings of words as such, nor the interpretation of these words in light of their context, supports the EC claim that salted chicken cuts have to contain enough salt to be preserved in order to qualify for the tariff treatment under heading 02.10, the Panel and the AB turn to the question of whether the ‘object and purpose’ of the agreement sheds light on the issue.
6.1 The adjudicating bodies' findings
According to the Panel, the object and purpose of the WTO Agreement:
7.320 Taken together, the relevant aspects of the WTO Agreement and the GATT 1994 indicate that concessions made by WTO Members should be interpreted so as to further the general objective of the expansion of trade in goods and the substantial reduction of tariffs … In other words, the terms of a concession should not be interpreted in such a way that would disrupt the balance of concessions negotiated by the parties. Finally, the interpretation must ensure the security and predictability of the reciprocal and mutually advantageous arrangements manifested in the form of concessions.
7.321 … the question has arisen in this case as to whether the interpretation of the concession contained in heading 02.10 of the EC Schedule to include a long-term preservation criterion could undermine this objective. In this regard … it is necessary to look exclusively at the ‘objective characteristics’ of the product in question when presented for classification at the border … As regards the present case, the WCO surmises that laboratory analyses might be required to determine whether a product can be regarded as ‘salted’ within the meaning of heading 02.10 of the HS. (Footnote omitted)
7.322 … despite questioning by the Panel and by Brazil, the European Communities has not provided the Panel with any clear idea of what is meant by ‘long-term preservation’ in practice … [I]f a product arrives at the EC border that is salted and then frozen, (as in the case of the products at issue), the Panel questions how a customs officer at the border tasked with identifying the appropriate heading under which the product should be classified is to know whether: (a) the product has been preserved for the long-term; and (b) if so, whether the long-term preservation is the result of salting, freezing or a combination of the two. While the first question may be addressed through laboratory analyses, which the European Communities states that it conducts when necessary, it is far from clear to us how the answer to the second question will be determined. Yet, without a means to determine the answer to the second question, the customs officer will not be in a position to know whether the product in question should be classified under heading 02.10 (i.e., because the long-term preservation is attributable to the salting) or under heading 02.07 (because the long-term preservation is attributable to the freezing). (Underlining added, footnotes omitted)
The Panel concludes that the difficulty for customs officers to classify products means that the outcome is hard to predict and that this makes it violate GATT:
7.323 In the Panel's view, the lack of certainty associated with the application of the criterion of long-term preservation with respect to the concession contained in heading 02.10 of the EC Schedule runs counter to one of the objects and purposes of both the WTO Agreement and the GATT 1994, namely that the security and predictability of the reciprocal and mutually advantageous arrangements must be preserved. Therefore, the Panel concludes that an interpretation of the term ‘salted’ in the concession contained in heading 0.210 of the EC Schedule to include the criterion of long-term preservation could undermine the object and purpose of security and predictability, which lies at the heart of the WTO Agreement and the GATT 1994.
Turning to the AB, it notes a potentially important objection by the EC to the Panel's reasoning:
245. The European Communities contends that the Panel misrepresented the criterion of preservation, because EC Regulation 1223/2002 and EC Decision 2003/97/EC do not, in fact, apply a ‘criterion of long-term preservation’ but, instead, treat chicken cuts with a salt content of up to 3 per cent as falling under heading 02.07 rather than heading 02.10. (Footnote omitted)
The AB notes that there is nothing wrong in principle with a preservation criterion in the HS:
246 … we consider that the Harmonized System does not preclude the use of a criterion of preservation, as such … the application of such a criterion would not necessarily be in conflict with the objectives of security and predictability of the WTO Agreement and the GATT 1994 (including Schedules of tariff commitments). (Underlining added)
The AB recalls one of the problems the Panel found with the preservation criterion:
247 … However, it was unclear to the Panel whether preservation for the long-term had to be the result of salting, or freezing, or a combination of the two. (Footnote omitted)
The AB notes that this issue is effectively resolved by the EC. But for the AB there remains a concern about the second problem, in that the EC has not explained how the separate preservation effect of salting could be determined. There is consequently uncertainty concerning the criterion that is actually applied to evaluate the preservation effect:
248. Although the European Communities clarifies that, for purposes of heading 02.10 of the EC Schedule, preservation has to be the result of the processes mentioned in that heading and not of the processes listed under heading 02.07 (namely, chilling, freezing), it does not explain how, in respect of frozen and salted meat, the preservation effect of the processes listed in heading 02.10 could be distinguished from the processes listed in heading 02.07. Therefore, we share the Panel's concern about the lack of certainty in the application of the preservation criterion used by the European Communities regarding the tariff commitment under heading 02.10 of the EC Schedule. (Underlining added, footnote omitted)
This uncertainty is detrimental to the object and purpose of the agreement:
249. In the light of these considerations, we see no reason to disturb the Panel's finding, in paragraph 7.328 of the Panel Reports, that ‘the lack of certainty associated with the application of the criterion of long-term preservation with respect to the concession contained in heading 02.10 of the EC Schedule … could undermine the object and purpose of security and predictability, which [underlie] both the WTO Agreement and the GATT 1994’.
6.2 Discussion
There are two aspects of the adjudicating bodies' object and purpose analysis that we would like to highlight.
The first issue concerns the sense in which the EC measure is found to contradict the object and purpose of the WTO Agreement. It appears as if the Panel (and the AB seems to follow the Panel in this regard) oscillates between two different interpretations of the notions of predictability and security for the object and purpose of the agreement. In its analysis under object and purpose, summarized above, the focus is on the alleged uncertainty of the outcome of the new customs classification procedure. However, the Panel in its concluding Section jumps to another problem involving the object and purpose of the agreement. The two final paragraphs of the Panel Report subsection ‘Conclusions regarding the application of Article II of the GATT 1994 in this case’, where the Panel summarizes its findings, state that:
7.426 It is the Panel's view that the products at issue are covered by the concession contained in heading 02.10 of the EC Schedule. Therefore, such products are entitled to treatment provided for by that concession. Since the products at issue are not being accorded such treatment, the European Communities is in violation of Article II:1(a) and Article II:1(b) of the GATT 1994.
7.427 In reaching this conclusion, the Panel recalls that a fundamental object and purpose of the WTO Agreement and the GATT 1994 is that the security and predictability of reciprocal and mutually advantageous arrangements must be preserved. In the Panel's view, a Member's unilateral intention regarding the meaning to be ascribed to a concession that Member has made in the context of WTO multilateral trade negotiations cannot prevail over the common intentions of all WTO Members as determined through an analysis undertaken pursuant to Articles 31 and 32 of the Vienna Convention. (Underlining added)
As can be seen here, in finding that the EC violates Article II, the Panel heavily leans on the reduced security and predictability that the measure is found to cause. Note, however, that the Panel now seems to refer to the unilateral nature of the EC's imposition of its interpretation of the schedule, rather than to the uncertainty as to the customs classification outcome. However, this reference to unilateralism begs the question of whether one can establish that the interpretation in question is somehow contrary to the common intentions of the parties. The fact that, when one applies WCO rules and practice, the EC choice of the one classification over the other does not obviously appear to be impermissible or unreasonable, indicates that while ‘unilateral’ in the trivial sense that all customs administration is normally ‘unilateral’, the EC classification decision does not violate any established common expectations or intentions of the WTO Membership or the parties to the dispute.
The second aspect of the finding under object and purpose that we would like to point to is the uncertainty that the adjudicating bodies are concerned with. As is clear from para. 248 of the AB report, the alleged source of the uncertainty is the difficulty in distinguishing the effect of freezing from the effect of salting on long-term preservation. What is not sufficiently explained, however, is the reason why this would be difficult. Is the problem that one cannot tell whether a defrosted chicken cut is salted enough for long-term preservation? Or is it that the notion of ‘long-term’ is left undefined? Or is there some other problem?
The determination invites several questions on this score. For instance, what if the EC had used a minimum quantitative threshold as a surrogate for the criterion of preservation? In other words, the EC might have established a new minimum salt requirement, which reflected the minimum quantity of salt that would be needed, as a general rule, in order for a product to be effectively preserved in the long term through treatment by salt. Although such a criterion might be somewhat crude – as such a level would, in any given case, likely not always perfectly reflect the exact quantity needed for preservation – it would provide a bright-line standard that supported predictability and certainty in the application of tariffs. The AB is unclear as to whether a bright-line quantitative threshold to limit the ambit of 02.10 to products salted for preservation would still violate Article II GATT inasmuch as there is no intrinsic reason to consider 02.10 as limited by the concept of preservation. In other words, the legal weight attached to the question of administrability remains unclear.
Or, to consider another modification, what if the procedure is altered such that it always results in the higher tariff under heading 02.07? This would presumably take care of the problem with uncertainty, since the outcome is now completely foreseeable. Would this modification mean that the measure no longer threatens the ‘object and purpose’ of the agreement?
In any case, we note that customs regulations are subject to the transparency requirements in Art. X GATT, which have recently been the subject of an important AB judgment (EC–Customs Practices). Arguably, the issue of the administrability of the ‘preservation’ standard is more appropriately addressed under Art. X:(3)(b) GATT, which requires that, ‘Each contracting party shall administer in a uniform, impartial and reasonable manner all its laws, regulations, decisions and rulings’ pertaining inter alia to customs classification.
7. Subsequent practice
In the EC – LAN Equipment case, the AB held, pursuant to Art. 31:(3)(b) VCLT, that subsequent practice in tariff classification by WTO Members was relevant to the interpretation of a tariff schedule but only to the extent that it revealed the common intention of the parties (para. 90). Thus, the subsequent practice of only one WTO Member would be of limited relevance. In EC – Chicken Cuts, the EC was the only importing WTO Member engaged in classifying the products in question for customs purposes. Nevertheless, the Panel found that the consistent practice of the EC in classifying the products in question under 2.10 prior to 2002 constituted relevant ‘subsequent practice’ establishing the common intent of the parties. The Panel based its conclusion on the notion that the acquiescence of other WTO Members established that the practice of the EC, during this period, reflected the common intentions of WTO Members. The AB saw this logic as at odds with its understanding of common intentions in EC– LAN Equipment and reversed the Panel:
259 … To our mind, it would be difficult to establish a ‘concordant, common and discernible pattern’ on the basis of acts or pronouncements of one, or very few parties to a multilateral treaty, such as the WTO Agreement. We acknowledge, however, that, if only some WTO Members have actually traded or classified products under a given heading, this circumstance may reduce the availability of such ‘acts and pronouncements’ for purposes of determining the existence of ‘subsequent practice’ within the meaning of Article 31(3)(b).
262 … although this dispute concerns the scope of a tariff commitment contained in the WTO Schedule specific to the European Communities, the relevant headings are common to all WTO Members.
266 … we fail to see how the Panel's finding that it was ‘reasonable to rely upon EC classification practice alone in determining whether or not there is ‘subsequent practice’ that ‘establishes the agreement’ of WTO Members within the meaning of Article 31(3)(b)’ can be reconciled with these statements of the Appellate Body in EC–Computer Equipment. (Footnote omitted)
270. In our view, as the Panel examined only a subset of salted meat products classifiable under heading 02.10, and it did not examine classification practice with respect to alternative headings such as heading 02.07, it could not draw valid conclusions as to the existence of ‘subsequent practice’ establishing the agreement of the parties within the meaning of Article 31(3)(b) with respect to all salted meat products potentially covered by the tariff commitment under heading 02.10 of the EC Schedule. (Underlining added)
272 … We, therefore, disagree with the Panel that ‘subsequent practice’ under Article 31(3)(b) has been established by virtue of the fact that the Panel ‘[had] not been provided any evidence to indicate that WTO Members protested against the EC classification practice in question from 1996–2002’. (Footnote omitted)
273 … ‘lack of reaction’ should not lightly, without further inquiry into attendant circumstances of a case, be read to imply agreement with an interpretation by treaty parties that have not themselves engaged in a particular practice followed by other parties in the application of the treaty.
276. For the reasons set out above, we reverse the Panel's interpretation and application of the concept of ‘subsequent practice’ within the meaning of Article 31(3)(b) of the Vienna Convention; consequently, the Panel's conclusions, in paragraphs 7.289–7.290 and 7.303 of the Panel Reports, that the European Communities' practice of classifying, between 1996 and 2002, the products at issue under heading 02.10 of the EC Schedule ‘amounts to subsequent practice’ within the meaning of Article 31(3)(b) of the Vienna Convention cannot stand. (Underlining added)
We believe the AB is correct in stating that one cannot, without very careful analysis of the entire context, come to the conclusion that silence concerning one Member's tariff classification practice constitutes assent to that practice. The product in question was imported only by the EC and exported only by complainants. It would be odd if other WTO Members, in these circumstances, were to pay any attention to the customs classification of these products. The logic of the Panel's approach is that one or a few WTO Members who have an original import or export interest in a product may establish an approach to classification that then becomes binding on other WTO Members who may have an import or export interest in the future; to avoid this outcome the other Members in question would have to speak out, even before they were aware that they had any import or export interest!
8. Supplementary means of interpretation
In the EC – LAN Equipment case, the AB found that Art. 32 VCLT may be relevant to interpretation of a WTO Member's schedule (para. 86). In particular, the AB referred to the possibility for a treaty interpreter, under Art. 32, to take into account ‘circumstances of [the] conclusion of the treaty’, which the AB understood to entail ‘an examination of the historical background against which the treaty was negotiation’ (para. 86). Art. 32 may be used either to confirm an interpretation of a treaty based on Art. 31 VCLT, to resolve ambiguity that remains after the use of Art. 31, or where the interpretation resulting from Art. 31 results in absurdity.
The historical elements considered by the Panel as ‘circumstances of [the] conclusion of the treaty’ included EC law and classification practice prior to and during the Uruguay Round negotiations. First of all, the Panel considered EC Regulation 535/94, which came into effect shortly prior to the conclusion of the EC Uruguay Round schedule. The Panel found: ‘[t]he effect of that Regulation … was that if meat had been deeply and homogeneously impregnated with salt, with a minimum salt content of 1.2% by weight, it would meet the requirements of that Regulation and would qualify as “salted” meat under heading 02.10 of the CN’. Second, the Panel had recourse to minutes of the EC Customs Code Committee, which the Panel held ‘provide compelling evidence that the principle of long-term preservation was not included in the definition of “salted” in EC Regulation No. 535/94’ (Paras. 7.365–370). Third, the Panel considered the Dinter and Gausepohl judgments by the European Court of Justice. The Panel concluded that these were of little or no assistance, since they focused on a different tariff classification (Dinter), or were ambiguous (Gausepohl) and superseded by Regulation 535/94 prior to the conclusion of the Uruguay Round. The Panel rejected the relevance of certain other historical materials, such as evidence of US tariff classification practice prior to the Uruguay Round. As for the Explanatory Notes to the EC customs code, the Panel assumed these to have been superseded by the Additional Notes to Regulation 535/94 prior to the conclusion of the Uruguay Round and hence not relevant.
On appeal, the EC challenged the Panel's use of historical background on a number of grounds. First of all, the EC argued that to be relevant under Art. 32 VCLT, an instrument or document must have had a direct influence on the treaty negotiators. The AB responded that this is too narrow a conception: the instrument or document may be relevant simply by virtue of illuminating what the negotiators' thinking or assumptions were at the time, without necessarily having influenced any actual positions or conduct in the negotiations. Second, the EC argued that to be considered as part of the ‘circumstances of [the] conclusion of the treaty’, the instrument or document must have a close temporal proximity to the conclusion of the negotiations. The AB considered that temporal proximity went to probative weight or relevance and that there was no specific time period beyond which a document or instrument would necessarily not qualify as part of the ‘circumstances’. Third, the EC argued that in order for an instrument or document to be considered as part of the ‘circumstances’, it must be demonstrated that the negotiators had actual knowledge of that instrument or document. The AB rejected this argument, holding that where an instrument or document was publicly available to negotiators they could be deemed to have knowledge of it; however, the AB also noted that where actual knowledge could be shown to exist, probative weight or relevance might be greater. Fourth, the EC argued that the Panel erred in ignoring the classification practice of the United States prior to the Uruguay Round. The AB appeared to assume that, as a matter of law, practice of other WTO members would be relevant (and indeed this had been emphasized by the AB in EC–LAN Equipment), but deferred to the Panel judgment as a trier of fact that different products were at issue in the case of the United States and thus that the practice in question was not relevant to the dispute at hand. Fifth, the EC claimed that the Panel had erred in finding that the Gauspohl judgment had been superseded by Regulation 535/94; according to the EC, as a matter of internal Community law a Regulation could not supersede a ruling of the ECJ but only implement it. The AB found that it did not have to decide this question of internal Community law, because a proper reading of Gauspohl revealed that Gauspohl did not stand for the proposition that only meat salted for purposes of long-term preservation was ‘salted’ within the meaning of 2.10, and thus Gauspohl did not provide support for the EC position in any case.
It is our impression that the challenges of the EC with respect to the Panel's use of historical background as ‘circumstances’ are largely formalistic. They imply, generally speaking, that the exercise under Art. 32 is an extremely mechanical one, involving the a priori inclusion or exclusion of particular categories of instruments or documents using narrow criteria such as temporal proximity. But the AB is surely right that Art. 32, which deals with supplementary means of interpretation used at the discretion of the treaty interpreter (unlike the sources in Art. 31 that are obligatory), and does not claim even to exhaust what is a ‘supplementary means’, should not be read in this inflexible or rigid manner; the aim is to try and get a comprehensive and persuasive picture of the common assumptions and intentions of the negotiators. What kinds of instruments and documents are most relevant will be a matter of judgment and finesse, and depends on the issue and the circumstances, as indeed any actual historian would know.
If we were to take issue with any element of the Panel's and the AB's analysis, it would be with the overall conclusion that the instruments and documents in question confirm the interpretation that salting is not limited to salting for long-term preservation, merely because none of these documents sets out clearly or explicitly the limitation in question. For none of the instruments or documents explicitly rejects that limitation either, nor is in direct contradiction with such an implied assumption. The Panel and AB assumed that the EC must show explicit indications of the limitation that it claims on the meaning of salting and not that the complainants need to show evidence that such a limitation was rejected. It is as if, once the adjudicator had determined that the ‘ordinary meaning’ of ‘salted’ in light of context, object, and purpose is not restricted to salting for long-term preservation, the EC bears the burden of showing the existence of a special understanding to the contrary. This seems at odds with the AB's own notion of the application of the VCLT rules as a holistic exercise. However, in fairness to the AB, while Art. 32 can be used to confirm an interpretation under Art. 31, it is far from clear that Art. 32 could be used to displace an interpretation based on the use of the primary and obligatory means of interpretation in Art. 31.
9. Where should the dispute have been adjudicated?
EC – Chicken Cuts concerns conflicting claims about the correct tariff classification of a product that was not traded at the time of the conclusion of the Uruguay Round, but that subsequently was exported to the EC by firms in the complaining countries. At an abstract level, it is the result of contractual incompleteness. Products in general have a large number of different features: they come in different colors, weights, shape; they contain different materials; are produced through different production processes, etc. As we have discussed above, it would be extremely costly to write a tariff classification schedule with such a degree of precision that it would be crystal clear where each and every product should be classified. Instead, product descriptions are more or less vague – and thus incomplete – just like the trade agreement is incomplete in many other respects (such as the ambiguity it leaves concerning domestic instruments). The particular incompleteness at issue in EC – Chicken Cuts is the fact that the agreement does not explicitly state what tariff treatment the salted chicken cuts should receive. Instead, the parties to the agreement have to interpret imprecise concepts such as ‘salted’ in order to determine this.
When an agreement does not regulate every conceivable situation the parties may end up in, the parties have a degree of discretion with regard to their actions at least in some circumstances. The parties' expectations concerning the behavior of other parties will then tend to play an important role, and may importantly affect the willingness of parties to make concessions in the first place. This is where the notions of good- and bad-faith behavior obtain meaning. Interestingly, in the instant dispute, neither the Panel nor the AB refers to these notions at all, despite the fact that the measure must be said to concern events that are not fully accounted for in the agreement, perhaps being restricted by the arguments advanced by the parties. Of course, we have no knowledge of what provoked this particular dispute, and whether either side can be said to have acted in bad faith. But it is still interesting to contemplate more generally the different scenarios that may give rise to a dispute of this type.
One possibility is that there was a mutual understanding at the time of the completion of the Uruguay Round negotiations concerning the appropriate classification, but that one party subsequently opportunistically sought to exploit an ambiguity in the text to its advantage. For instance, perhaps there was an understanding among the parties that the imports of chicken cuts from Brazil and Thailand were to be classified under heading 02.07, but producers in these countries saw the possibility of adding some salt with the sole purpose of getting a more favorable tariff treatment. Or alternatively, maybe the understanding was to classify the product under heading 02.10, but the EC saw the opportunity to effectively increase the tariff on the product by adding a ‘long-term preservation’ criterion that was not part of the understanding. In these cases, it would be natural to view the opportunistically acting party as behaving in bad faith, undermining the cooperative spirit of the agreement. The natural forum for adjudicating such a dispute would be the DSU, albeit with the input of some expert advice from the WCO.
A second and related possibility is that the parties did not think at all about the classification of this particular type of product that is both frozen and salted, since there was no trade at the time of the completion of the Uruguay Round. It can be noted that it is established in this case that Thailand started exporting the product in 1996 and Brazil in 1998, hence in both cases after the completion of the Uruguay Round negotiations.Footnote 6 This may have been caused by market developments, such as increased demand in the EC – indeed, the complainants claim that their firms have responded to explicit requests from EC buyers – or expanded export capacities in the complaining countries. Here there is a genuine uncertainty concerning the classification, and it seems as if the WCO could have a more prominent role to play, to the extent the matter can be resolved on technical grounds. But it may also call for negotiations between the parties.
A third possibility is that there was a genuine misunderstanding during the negotiations concerning the classification of the product. Maybe the complainants sincerely expected the chicken cuts to come under heading 02.10, while the EC sincerely believed that the parties shared the view that ‘salted’ referred to preservation. This is clearly an issue to be resolved through the DSU, and perhaps also through negotiation between the parties.
Yet another possibility is that there was a mutual understanding that each WTO Member's customs authority has the right to interpret their custom classification schedules, subject to general WTO norms such as nondiscrimination and transparency and subject to the possibility of resort to the WCO in the case of a disagreement concerning a Member's interpretation.
As is stated in the Panel report, the WCO Secretariat took the position that since the WTO claim of Brazil and Thailand entailed an underlying dispute over the meaning of HS classifications, that underlying dispute should first be brought before the WCO, pursuant to the compulsory dispute settlement clause in the WCO Convention. One interpretation of the existence of such a provision is that the WTO members bound their tariff schedules based on the understanding that each WTO Member's customs authorities have the right to interpret these schedules, subject to general WTO norms such as nondiscrimination and transparency and subject to the possibility of resort to the WCO in the case of a disagreement concerning a Member's interpretation. The parties certainly did not, in the Uruguay Round instruments related to customs matters, seek to limit or circumscribe the extent to which disputes about the meaning of HS classifications would be matters for the WCO. Customs nomenclature is a highly specialized technical vocabulary for which canons of interpretation exist in the WCO. These canons are different from those of the VCLT.
While the AB in EC – LAN Equipment held that ‘the only rules which may be applied in interpreting the meaning of a concession are the general rules of treaty interpretation set out in the VCLT’ (para. 84), the VCLT itself contemplates that some treaty terms have been given a specialized meaning by the parties, and therefore the methods in 31.1–3 of the VCLT may be preempted by an inquiry into this specialized meaning. Thus, Art. 31.4 VCLT provides: ‘A special meaning shall be given to a term if it is established that the parties so intended.’ Art. 31.4's ‘special meaning’ clearly contrasts with ‘ordinary meaning’, which is to be seen from the sources set out in Art. 31.1–3. In our view, the Panel and the Appellate Body ought to have considered the possibility that the WCO Convention, including its compulsory dispute settlement clause to which almost all WTO Members are parties (and certainly all parties to the dispute), indicates a common intention that the meaning attributed to HS classifications be that which has been established in the WCO.
If so, then the AB may have failed to grasp the underlying nature of the legal obligation in Art. II:(1) GATT in its relation to the WCO system. A WTO Member may have a right to apply a tariff up to the bound rate for a particular classification to a particular product, provided that product falls within the classification as understood in the WCO system. Thus, if the burden of proof is on the complainant to make a prima facie case, arguably Brazil and Thailand should have had to show that the interpretation by the EC of the HS classification in question was inconsistent with the specialized meaning of that classification established in the WCO.
The issue of ‘fragmentation’ is one of the largest challenges facing international law today – the possibility of more than one forum and/or international legal regime applying to a given problem or dispute: there are few meta-rules in international law to resolve such situations.Footnote 7 The response of the Panel to the challenge, as it emerges in this dispute, is formalistic and preemptory:
7.56 We understand that, once seized of a matter, Article 11 prevents a panel from abdicating its responsibility to the DSB. In other words, in the context of the present case, we lack the authority to refer the dispute before us to the WCO or to any other body.
This statement concerning the nature of the Panel's jurisdiction under the DSU may be technically or literally true,Footnote 8 but, as we have just explained, a plausible interpretation of the nature of the underlying substantive obligation of WTO Members concerning customs classification is that they are required to follow WCO practice with respect to the treatment of particular products under HS classifications of the WCO. On this view, a WTO Member would discharge its obligations under Art. II GATT if its decision to apply an HS classification to a particular product is consistent with WCO practice. In a case like this one, where no previous clear WCO ruling exists on the matter at hand, the practical consequence of this view is that, in order to make their case before the WTO Panel, the complainants will need to go first to the WCO to establish that the defendant has interpreted the classification in question in a manner that is impermissible in the WCO system. We emphasize that it is far from clear that the WCO would have provided a more satisfactory resolution of the dispute in substantive terms (i.e. a more efficient outcome); it would most likely have decided the matter by applying a set of mechanical rules for dealing with situations where, in principle, more than one HS classification might apply to a given product. The question is whether the adhesion of the parties to the WCO Convention and the arguably related absence of any code on customs classification in the WTO itself indicate the expectation or assumption that problems of interpretation of the HS would be dealt with in the WCO.
One kind of argument for not requiring, in effect, that the complaining Member have first gone to the WCO is that not all WTO Members are WCO Members, and thus in the case of a dispute where the complaining WTO Member is not a WCO Member, the complaining Member in question would be faced with discharging its burden of proof without the benefit of a prior WCO ruling. This would arguably result in a nonequitable or at least nonuniform application of Article II of the GATT to different WCO Members. However, as the International Law Commission Study Group has pointed out, it is an inevitable feature of ‘fragmentation’ that states are bound by different multilateral and bilateral treaties, overlapping in subject matter and partly but not entirely overlapping in membership. The rights and obligations of parties to agreement A may affect how agreement B applies to them; agreement B may thus inevitably have a somewhat different effect on parties to B who are not parties to A. There is no structural principle of international law that suggests the rights and obligations in agreement A should be rendered ineffective so as to make interpretation of agreement B uniform across states party to A and states nonparty to A. Indeed, there have been decisions of WTO dispute settlement organs precisely to the contrary. Thus, for example, in EC – Poultry, the AB held that a bilateral agreement between the EC and Brazil was relevant as a supplementary means of interpretation:
to the extent relevant to the determination of the EC's obligations under the WTO agreements vis-à-vis Brazil. (§ 202, italics in the original).
This statement contains an obvious recognition of the possibility that WTO rights and obligations may apply differently vis-à-vis particular WTO Members depending on the meaning of other agreements between those particular Members to which the entire WTO Membership is not bound. However, since very few WTO Members are not WCO Members, the degree of asymmetry is not likely to be significant.
On the other hand, there are important respects in which requiring the complaining Member to rely on a previous WCO ruling, where that Member is a party to the WCO and its compulsory dispute settlement arrangements, risks increasing uncertainty in the meaning of WTO obligations. What if, subsequent to a WTO ruling interpreting a customs classification, a dispute based on similar facts were brought to the WCO? The WTO findings would not be binding on the WCO adjudicator as res judicata; the weight given them would depend on the WCO's view of their persuasiveness.Footnote 9 There is more than a mere possibility that the WCO might make a finding concerning the meaning of HS classifications that is inconsistent in whole or in part with the result and/or reasoning of the WTO adjudicator. Now remember that the vast majority of WCO Members are also WTO Members. In such a situation, that vast a majority would be faced with conflicting obligations: as WCO Members, they would be required to follow WCO decisions concerning the proper interpretation of the HS; as WTO Members, they would have to follow the approach to interpretation of the HS laid out by the WTO dispute settlement organs. Thus, far from contributing to uniformity and consistency, the development of a WTO case law on the principles of HS classification, largely autonomous from decisions of the WCO, is a recipe for conflict and uncertainty.
10. Conclusion
At numerous points in this paper, we have expressed serious misgivings at the overall framing of this dispute by the parties, the Panel, and the AB:
1. The definition of the burden of proof for establishing a violation is too onerous for the importing country. In a case where two headings can be argued with equal force as the appropriate classifications, we see no reason to choose the one the exporter prefers to be the correct heading.
2. The argument that this is not a customs classification case is at best only partially correct. The core of the adjudicating bodies' determination concerns appropriate interpretation of ‘salted’ with regard to heading 02.10, and this is, as far as we can see, a customs classification issue. For this determination, the Panel should have asked for expert opinion by the WCO.
3. We believe that the adjudicators need to reflect more deeply on the nature of the obligation in Art. II, as it relates to the interpretation of the HS in domestic customs administration and to the jurisdiction and rules of the WCO. We are not persuaded that the intent or expectation of the WTO Membership was that a correctness standard of review under the VCLT would apply to the interpretation by domestic agencies of customs schedules. It is at least as plausible, if not more, that the background assumption was that disagreements about customs administration would be settled through the compulsory dispute settlement mechanism of the WCO, to which the vast majority of WTO Members are parties. Indeed, this might go a long way to explaining why, although customs classification issues have arisen often in international trade, they have not given rise to negotiation of specialized rules or practices in the WTO itself.
4. For the above reasons, we do not think that the conclusion that there was an EC violation of Art. II of the GATT could be justified on the basis of the evidence presented in this dispute.
5. This being said, we do think the actual outcome of the case might well be justified on other grounds. Quite apart from whether the WCO would find that the EC interpretation of salted was correct or not, Art. X of the GATT obliges each WTO Member to ‘administer in a uniform, impartial and reasonable manner all its laws, regulations, decisions and rulings’ concerning, inter alia, customs classification. An unexplained, unreasoned determination by the EC that, after years of almost all its customs offices classifying the products in question as salted, this practice is an error, and the resultant imposition of duties effectively double or even treble or more on the products in question may well not amount to a ‘reasonable manner’ of customs administration. Despite numerous questions asked to the EC by the Panel, and hundreds of pages of submissions in this litigation, the EC never provided a clear reasoned account of how this error was allowed to continue for years, from the perspective of what policy concerns it was an error, and how exporters were supposed to grasp, from EC practice, that even though its customs offices were routinely classifying these products erroneously, all along EC law and regulation imposed the more restrictive interpretation of when a product was to be considered ‘salted’. This relates to another provision of GATT Art. X, which requires that ‘Laws, regulations, judicial decisions and administrative rulings of general application, made effective by any contracting party, pertaining to the classification … of products’ be published promptly and in such a manner ‘to enable governments and traders to become acquainted with them’. It is the position of the EC that the long-term preservation criterion was always part of EC law and regulation, and that its own authorities were simply erroneously applying the law. If so, it is hard to imagine from any of the materials how a trader would have become acquainted with such a requirement – the EC's own officials obviously weren't even acquainted with it. In sum, the lack of reasoned explanation by the EC of its course of conduct in the administration of customs duties in regard to these products makes it difficult to argue for a different result in this case.
To conclude, we believe that the outcome of the dispute may possibly be justified, albeit with a different legal basis than that in the dispute, but we have severe reservations concerning the legal grounds for finding a violation of Art. II on the basis of the evidence presented.
We are grateful for very helpful discussions with Petros C. Mavroidis, Chad Bown, Tom Prusa, and Don Regan.
