Static welfare effects

Let us consider first the static welfare effects of the measures introduced by the EU, which have raised tariffs on the affected high-tech products from zero to up to 14%.

These measures clearly hurt foreign high-tech producers, by reducing their access to the European market. They are also clearly detrimental to European consumers, since they increase the price they have to pay for IT products. Starting from Krugman (Reference Krugman1979), ‘new trade theory’ suggests that trade barriers can also hurt consumers by reducing the number of varieties available to them. Recent work by Broda and Weinstein (Reference Broda and Weinstein2006) suggests that this effect could be sizeable: examining disaggregated US import data available for the period between 1972 and 2001, they show that goods from different countries are far from perfect substitutes, implying that increases in the number of varieties can generate potentially large welfare gains.Footnote ⁷ Indeed, they find that the four-fold increase in available global varieties arising since the early 1970s has produced large welfare gains for the United States, raising real income by about 3%.

The protectionist measures introduced by the EU also give rise to static welfare gains. First, they generate tariff revenues, estimated to be around €300 million per year.Footnote ⁸ Second, they benefit European producers by sheltering them from foreign competition. The tariff restrictions can benefit both producers of multifunctional high-tech goods (e.g., LCD computer monitors) and producers of more traditional consumer electronics that can be incorporated in new technologies (televisions).

Standard trade theory suggests that the overall effect of a tariff increase depends on whether or not it affects the terms of trade, leading to changes in the world prices of traded goods. In the absence of terms-of-trade effects, a tariff unambiguously lowers welfare, since the welfare loss in terms of consumer surplus more than offsets the welfare gains associated with increased tariff revenues and producer surplus. If instead the tariff increase improves the country's terms of trade, its overall (static) welfare effect will be ambiguous.

Dynamic welfare effects

We discuss next the dynamic welfare implications of the introduction of EU tariff restrictions on IT products.

In general, restricting trade is a bad idea for growth, as suggested by the recent literature exploring the link between countries’ trade policies and their firms’ productivity.Footnote ⁹ This literature asks whether firms achieve higher productivity growth by becoming exporters or by being forced to improve as a result of more intense competition with foreign rivals. Heterogeneous-firm models of international trade (e.g., Melitz, Reference Melitz2003; Bernard et al., Reference Bernard, Eaton, Bradford Jensen and Kortum2003) show that lowering trade barriers leads to productivity gains via the reallocation of economic activity across firms within industries.Footnote ¹⁰ Various empirical studies have examined the connection between changes in trade costs and firm-level outcomes, finding support for the predictions of heterogeneous-firm models.Footnote ¹¹ There is also some evidence that liberalizing trade leads to increases in productivity by inducing firms to innovate.Footnote ¹²

Imposing restrictions on international trade in general can thus be detrimental for firms’ productivity and innovation. Imposing tariff restrictions on IT products in particular is likely to be even more harmful. The vast literature on endogenous growth (see Aghion and Howitt (Reference Aghion and Howitt2009) for a review) stresses that investment in new technology and knowledge capital is key to fostering output growth. In line with this idea, the ITA was expressly designed to encourage growth through innovation in the IT sector. A study by Feenstra et al. (Reference Feenstra, Mandel, Reinsdorf and Slaughter2009) suggests that lowering tariffs on IT products has indeed led to significantly higher productivity growth,Footnote ¹³ implying that tariff restrictions on high-tech goods are likely to reduce innovation and be detrimental to growth.

We conclude the analysis of the dynamic welfare effects of the EU measures by considering the impact of trade-policy uncertainty on investment and growth. Unpredictable tariff treatment can discourage investment in new technology. This is because, as first pointed out by Arrow (Reference Arrow and Nelson1962), uncertainty poses problems to exploratory scientific research and investment in innovation, which are often compounded by the long-time horizon needed for research to bear fruit in practical applications.

In the context of trade policy, uncertainty can arise because of the failure of the government to credibly commit to tariff cuts vis-à-vis private investors and firms.Footnote ¹⁴ Indeed, foreign IT firms complained that EC duties ‘create uncertainty in the marketplace’ and ‘stifle innovation’ on information-technology products.Footnote ¹⁵

It has been suggested that international agreements can be used to ‘tie the policymakers’ hands’, enhancing the credibility of trade liberalization. This idea was first put forward by Staiger and Tabellini (Reference Staiger and Tabellini1987) and has been later formalized by Maggi and Rodríguez-Clare (Reference Maggi and Rodríguez-Clare1998).Footnote ¹⁶ They suggest that entering into a binding trade agreement, such as the GATT/WTO, can be a solution to this problem. This argument, however, forgets that, absent a supranational authority with autonomous powers of enforcement, a country's international commitments are not directly binding on that country, but rather they must be sustainable in light of the dynamic incentives that the country faces vis-à-vis its trading partners as well as its domestic agents.Footnote ¹⁷

The literature on commitment problems in trade policy suggests that WTO rules can help to foster investment and innovation, but only if the private sector believes that they are credible. Uncertainty about the applicability of the rules can discourage firms from innovating. From this point of view, one of the fundamental problems with the ITA signed in 1996 is that it is based on a rigid list of products covered. The narrow range of products covered by the ITA implies that newly developed IT products face unpredictable tariff treatment: new products coming to the market, thanks to technological development, will have duties imposed on them when traded across borders.

The Panel's decision on EC–IT Products establishes the idea that all goods developed, starting from the high-tech products included in the ITA, should be traded freely among member countries. By effectively expanding the list of IT products covered by the original agreement, the Panel's ruling on this dispute should foster the credibility of trade-policy liberalization in the high-tech sector, thus stimulating innovation and growth.

The nature of state responsibility under GATT Article II and ‘as such’ claims

A preliminary objection of the EC to the complaint was that the regulatory actions in question did not engage state responsibility under the applicable provisions of the GATT, because they did not formally mandate or predetermine that any particular identifiable product would, necessarily, have duty applied to it. For example, one of the regulatory actions in question was the promulgation of a document providing interpretative guidance to customs officials in applying existing classifications, and the EC argued that, since it was not binding, it could not form part of the basis of an ‘as such’ complaint – that is, a challenge to laws or regulations on their face, as opposed to their case-by-case application to particular products at the border. In other instances, it appeared that the products in question were entering duty-free under a waiver, despite some of these regulatory actions, and there were issues also raised by the EC about the time frame or the formal validity of some of the actions.

Following the approach to state responsibility in the US–S.301 PanelFootnote ²⁸ and subsequent rulings such as the US–Corrosion-Resistant Steel Sunset Review Appellate Body Report,Footnote ²⁹ the Panel held that state responsibility under the GATT with respect to legal security of tariff bindings could be engaged by measures that did not formally predetermine or mandate that a particular traded product must be subject to duties in excess of the bindings in the member's schedule. It was enough that some features of the EC regulatory actions complained of would result in making some products otherwise covered by the duty-free concessions dutiable, if they were found to possess the particular characteristics identified in these challenged EC regulatory actions. In terms of state responsibility, one could express the principle as follows: the obligation in Article II of the GATT has to be understood as extending to a duty not to undermine the legal security afforded to the expectations of traders by the GATT scheduling architecture. Thus, Article II is violated by any regulatory action that diminishes the certainty provided by the concession. It is enough that the action in question gives rise to the possibility that the product will be subject to higher duties, due to characteristics or conditions not stated or implicit at the time the concession was made. Later on, we shall attempt to defend this apparently broad conception of state responsibility.

With respect to the non-guidance document, the Panel held that even though non-binding in domestic law, it nevertheless engaged state responsibility:

Flat-screen monitors

In ruling on the complaint with respect to flat-screen monitors, the Panel relied heavily on a textual analysis of the ‘ordinary meaning’ of the EC schedule, and in particular the headnote. It followed Appellate Body jurisprudence, indicating that, because schedules form an integral part of the GATT treaty, the rules in articles 31 and 32 of the Vienna Convention on the Law of Treaties (VCLT) apply to interpretation of Members’ schedules. On the other hand, even though the basis for the concessions at issue and the way they were structured in the EC's schedule was the ITA, the interpretation of the provisions of the ITA itself played a rather marginal role in the Panel's analysis. In the normative legal analysis, we will question certain aspects of the Panel's approach to the ITA. At this point, we simply note that the legal claims of the complainants did not include violations of the ITA itself, even though they might well have asserted such violations.

As a matter of textual interpretation of the EC's schedule, the EC argued that, because a list of HS-based tariff lines for the products in the Attachment B narrative description followed the general concession stated in the headnote for Attachment B products, customs administration based on HS classifications still governed or controlled the implementation of the general Attachment B concession. In other words, the EC was free to interpret the scope of the general Attachment B concession in light of its approach to HS-based classification of the products in question. In other words, the EC was saying that it did not intend through the ITA and the headnote to make a legally binding concession autonomous from the HS framework itself. The Panel rejected this argument on the following basis:

In accordance with the methodology suggested in Appellate Body jurisprudence concerning the sequential application of Articles 31–32 of the VCLT, having made these findings with respect to the ‘ordinary meaning’ of the EC schedule, the Appellate Body went on to address other elements in VCLT 31 (context, object, and purpose) and, finally, VCLT 32.

In examining the context, object, and purpose, the Appellate Body was inevitably going to have to confront the significance of the ITA itself. The Panel concluded that, at a minimum, the ITA would qualify as ‘context’ by virtue of being within the scope of VCLT 31(2)(b):

It is notable here that, as we will explore at length in our legal analysis, the Panel felt it had to stretch the wording of 31(2)(b) of the VCLT (the ITA was, as a matter of crude chronology, not collateral to the ‘conclusion’ of the covered Agreements but a subsequent agreement in time), when it could easily have found that the norms of the ITA were other relevant rules of international law applicable between the parties, within the meaning of VCLT 31(3)(c). As we shall explore in Section 3.3 below, this may have been due to earlier panel rulings that interpreted the word parties in VCLT 31(3)(c) in such a manner to make that provision largely inutile. We also note that in its analysis, by focusing on the word instrument, the Panel seemed to be creating some level of doubt as to whether the ITA is a treaty within the definition provided in the VCLT. This may have consequences for any future dispute where a claim of violation of the ITA itself is central to the Panel's jurisdiction.

In any case, having found the ITA to be part of the ‘context’ within the meaning of VCLT 31(2)(b), the Panel nevertheless held that the object and purpose of this contextual instrument, as expressed in its preamble, was not relevant to the interpretative exercise before the Panel:

At the same time, the Panel did find that the structuring of concessions under the operative provisions of the ITA was relevant, and that it supported the Panel's view of the ‘ordinary meaning’ of the Annex to the EC schedule, including the headnote.

On the other hand, the Panel rejected the HS as ‘context’ for interpretation of the EC's Annex to its Schedule and the Headnote to the extent that the EC had explicitly scheduled the general concession in question on a basis other than HS classifications (paragraphs 7.443ff; 7.515–7.517). This is an important finding, which strongly reinforces the autonomy of WTO Members if they so choose to schedule additional concessions on a basis that is entirely independent of the HS, as a matter of legal commitment. In the normative legal analysis, we will discuss why this finding is correct and its desirable consequences.

Unlike that of the ITA, as disclosed in its Preamble, the Panel did consider the object and purpose of the WTO Agreement as relevant:

We would add here, again without preempting the normative legal analysis to follow, that this discussion of object and purpose in respect to primary obligations reinforces the Panel's view of state responsibility, especially the importance of security and predictability.

Having found that the ordinary meaning of the EC's Annex to its Schedule and the Headnote, in light of context, object, and purpose supported the complainant's claim, the Panel then went on to consider material that was presented as part of the negotiating history of the ITA. Here the relevant language in VCLT 32 that supports the use of such material is ‘in order to confirm the meaning resulting from the application of article 31’. However, in the end, the Panel concluded that the nature of the material in question was such that it did not qualify for consideration under Article 32, and thus, in effect, that such material could neither confirm nor indeed undermine the determination of meaning under Article 31.

Finally, the Panel responded to an argument of the EC that its schedule had to be interpreted dynamically, in light of developments in technology. The possibility that a product with the narrative description in Attachment B could have the functionality of a TV screen did not exist at the time that the EC modified its schedule, and therefore the EC had never contemplated that its concession would extend in effect to television, as opposed to computer components. However, according to the Panel:

Set-top boxes

As with the flat-screen monitors, the Panel held that the narrative description in Attachment B controlled the scope of the concession and thus focused on the ordinary meaning of the words in that description, rather than the HS classifications listed in conjunction with it in the Annex to the EC Schedule. The Panel's reasoning for this interpretative choice is essentially the same as that for flat-screen monitors.

In considering the ‘ordinary meaning’ of the narrative description, the Panel rejected the EC argument that the language ‘which have a communication function’ allowed the EC to exclude from the concession set-top boxes with other functions, namely in this case recording:

The Panel rejected the further EC argument that, because the narrative description referred to communication through a ‘modem’, the EC was not required to extend the concession to boxes that used other means than the telephone modem typical of the technology current at the time of the concession to connect to the Internet, including ISDN, WLAN, and Ethernet technology. In the Panel's view, the use of the word ‘modem’ was intended to describe the functionality of Internet connection and was not a static description of one particular kind of device used for Internet connection. The Panel thus rejected, as with flat-screen monitors, that the scope of the concession should somehow be limited in terms of the state of technology existing at the time the concession was made:

These last observations of the Panel are of considerable importance. They suggest that the narrative description is only controlling with respect to EC classification up to a point: even if the product meets the criteria in the narrative description, and thus prima facie fits within the concession, there is still a holistic judgment to be made as to whether given its overall characteristics and functions, the product is better classified in a dutiable HS classification. Thus, what engages state responsibility is that the EC measures give rise to the possibility that certain products may be excluded from duty free treatment merely because of the technology used for connection to the Internet or merely because they contain a recording device.

With respect to context, the Panel, in addition to its rejection of the notion that the state of technology at the time was relevant context, also rejected the EC argument with respect to ‘subsequent practice’ within the meaning of the VCLT, namely that there was evidence that the overwhelming majority of WTO Members were classifying in the 1997–1999 period set-top boxes in HS classifications that supported the EC interpretation of the limited nature of the concession based on the narrative description. According to the Panel:

Examining the evidence presented by the EC with respect to Members’ classification practice, the Panel found that the standard of ‘consistent, common and concordant’ classification practice was not met, noting considerable divergence in the HS classifications that various Members associated with their narrative description of set-top boxes.

Transparency

The complainants argued that the EC had violated certain provisions of Article X of the GATT, in failing to publish some of the regulatory actions at issue in a timely fashion and by allowing for the possible (and actual in some cases) application of those changes prior to their being published. The EC argued, along the lines of its submissions on state responsibility, that because a number of the regulatory actions did not of themselves and at the time in question have a formal and binding character, they were not ‘laws, regulations, judicial decisions and administrative rulings’ within the meaning of Article X and thus that the operative provisions of that Article invoked by the complainants did not apply. The EC also argued that since some of the material had been put on a website in the form of committee minutes and draft documents, this sufficed in any case to meet the requirement of timely publication. The Panel disagreed with the EC on both points, and found for the claimants:

Printers with fax and photocopying capacity

In the case of these products, the relevant EC concession pursuant to the ITA was based on an HS classification, and therefore interpretation of the HS system was the central issue with respect to this part of the complaint. The complainants argued that the correct HS classification was subheading 8471 60, ‘Automatic data-processing machines and units thereof’; ‘Input or output units, whether or not containing storage units in the same housing’. The European Communities insisted that in some cases, that is where the photocopying function of the device met certain performance standards, it was justified in classifying it instead under another concession under the EC Schedule, that is the HS1996 subheading 9009 12 (‘indirect process electrostatic photocopying apparatus’).

As context for its interpretation of these classifications, the Panel resorted to the kind of material considered relevant to context by the Appellate Body in EC–Chicken Cuts,Footnote ³⁰ including HS interpretative notes. Of particular significance was one particular note that qualified the products that would fall under subheading 8471 60 as ‘of a kind solely or principally used in an automatic data-processing system’. The EC argued that a ‘unit’ may not be classified under HS1996 subheading 8471 60 unless it is used ‘solely or principally’ with an ADP system. This would exclude any unit that was principally used as a photocopier.

The Panel disagreed with the EC reading of the note:

Again following the relevant Appellate Body jurisprudence in for example EC–Computer Equipment, the Panel considered, under ‘context’, whether there was consistent customs classification practice of WTO Members that could be considered ‘subsequent practice.’

Having found that, prima facie, printers with fax and photocopying capacity were classifiable under 8471 60, the Panel went on to examine whether they might also (at least in some instances) be classifiable under 9009 12, as the EC argued. Here the examination of the structure of the HS classification and certain related documents (the HSEN 1996) as ‘context’ influenced the Panel's conclusion that 9009 12 applied exclusively to copiers that performed analogue photocopying using an optical system thus excluding any multifunctional machine with a digital copying capacity. As the Panel noted, the HS material in question contained relatively precise descriptions of the components of the copiers covered by 9009 12 (e.g. lenses, mirrors, etc.), which suggested a common understanding that 9009 12 applied to analogue copying technology. The Panel rejected the EC argument that these descriptions merely reflected the understanding of the characteristic means of photocopying based on the technology available at the time, and should not be taken to limit the application of 9009 12 to photocopiers using newer (e.g. digital) technologies.

State responsibility and ‘as such’ violation claims

We agree with the Panel's findings that the regulatory actions of the EC in dispute engage state responsibility in that they affect the predictability and legal certainty of its tariff bindings pursuant to Article II of the GATT.

The first conceptually rigorous and self-conscious statement of the problem of state responsibility in WTO law is to be found in the US–S.301 Panel ReportFootnote ³¹:

As the Panel went on to hold, the nature of the state responsibility engaged by the general statement in Article XVI:4, that is what measures in the nature of ‘laws, regulations and administrative procedures’ might violate the requirement to ensure conformity with WTO obligations can only be determined by analyzing the primary obligations that are the basis of the complainant's claim. In that case, the Panel was dealing with DSU Article 23.

According to the Panel, some primary obligations engage state responsibility when the measure creates uncertainty or undermines the legal security of the expectations of private economic actors:

These observations would seem to apply very strongly to tariff concessions. Member's schedules provide one of the most important kinds of information on the basis of which individual economic actors make investment and other economic decisions to exploit competitive opportunities in global markets. Thus, as the US–S. 301 Panel held, the provisions of the Preamble of the WTO Agreement that refer to predictability are significant for understanding the nature of state responsibility. As we have seen, these preambular provisions were rightly appreciated by the instant Panel in its elaboration of primary obligations.

Regardless of the formalistic pleas of the EC, it is clear from the facts that the regulatory actions complained of in this case introduced an element of uncertainty, specifically with respect to whether certain market-driven technological innovations, if fully exploited by private economic actors, would undermine the possibility of relying on duty-free access. Such uncertainty is costly – as explained in the economic analysis – and thus the value of the concession is undermined in fact regardless of whether, in any given case, the EC were in the future to apply the regulatory actions to impose tariffs on any specific product that would otherwise be subject to duty-free treatment pursuant to the ITA and the Headnote of the EC's schedule.

Ordinary meaning, context, object, and purpose

The Panel followed the methodology of first ascertaining ordinary meaning (using dictionaries and the like) and then considering context, object, and purpose. As has been very strongly argued in a previous ALI report by Horn and Weiler (Reference Horn, Weiler, Horn and Mavroidis2005), this may amount to hermeneutical absurdity: the proposition that one could ever have an ordinary meaning of a treaty term that is non-contextual. However, in fairness to the Panel, it was here applying the same methodology as the Appellate Body.

The ITA Agreement and VCLT 31

The Panel appears to have rather bent the language in Vienna Convention 31(2)(b) in characterizing the ITA as ‘(b) any instrument which was made by one or more parties in connection with the conclusion of the treaty and accepted by the other parties as an instrument related to the treaty’. The use of ‘subsequent’ in the next subparagraph of VCLT 31 is a clear indication that the VCLT drafters did not mean ‘in connection with the conclusion of the treaty’ to refer to instruments that, temporarily, were negotiated and concluded at a later time than the treaty itself. This is obviously the case with the ITA. This is very clearly confirmed by the French text, which reads ‘a l'occasion de la conclusion du traite’ (emphasis added). This makes it obvious that the instruments referred to are ones concluded simultaneously with the conclusion of the main treaty. There were other, more logical options available to the Panel to characterize the ITA as ‘context’. First, VCLT 31(3)(a): ‘any subsequent agreement between the parties regarding the interpretation of the treaty or the application of its provisions’; second, VCLT 31(3)(c): ‘any relevant rules of international law applicable in the relations between the parties’. There are two reasons why the Panel may not have used one of these characterizations. The first is alluded to explicitly.

The Panel seems reticent to come to the conclusion that the ITA is a treaty or an ‘agreement’ within the meaning of VCLT. The VCLT definition of a treaty is as follows: ‘an international agreement concluded between States in written form and governed by international law, whether embodied in a single instrument or in two or more related instruments and whatever its particular designation’. The Panel might have been led to a margin of doubt as to whether the ITA constitutes an agreement or treaty because it is called a Ministerial Declaration. But the VCLT, as just noted, clearly specifies that the ‘particular designation’ of the instrument does not matter. Moreover, the ITA article 6 specifies that Article XXIII of the GATT, that is the WTO dispute-settlement system, applies to address nullification and impairment of benefits under the ITA, regardless of whether there is any claim of a GATT violation itself. In other words, the dispute-settlement system is applicable to the ITA autonomously. Given such applicability, it would seem absurd to think that the ITA is something other than ‘an international agreement in written form and governed by international law’.

The second reason why the Panel might not have had resort to VCLT Article 31(3)(c) is due to an erroneous reading by an earlier Panel, EC–Approval and Marketing of Biotech Products,Footnote ³² of the expression ‘parties’ in 31(3)(c) as meaning that the other legal norm would have to be binding not just between the parties in the dispute, that is the parties engaged by the immediate exercise of treaty in treaty interpretation but all WTO Members. Under this reading, a plurilateral agreement such as the ITA would per se be excluded from consideration under VCLT 31(3)(c). A working group of the International Law Commission has persuasively criticized the EC–Approval and Marketing of Biotech Products interpretation of ‘parties’:

Citing extensively the concerns of the ILC about ‘systemic integration’ in international law, the Appellate Body of the WTO, in a ruling subsequent to the Panel that is the subject of this report, EC and Certain Member States–Large Civil Aircraft,Footnote ³³ has taken a different approach to the meaning of ‘parties’ in 31(3)(c), which allows, in some circumstances, for taking into account provisions of treaties to which not all WTO Members are parties:

Here the Appellate Body suggests a case-by-case balancing approach, taking into account the need for an interpretation of WTO law that applies to all Members not only those that are bound by another treaty, as well as the broader interest of international law in general in ‘systemic integration’, a concern that goes beyond the WTO as a specialized regime to the coherence and legitimacy of the international legal system as a whole – the importance of an interpretation of rights and obligations in one regime not undermining those undertaken in other multilateral fora.

In EC and Certain Member States–Large Civil Aircraft, the Appellate Body did not need to go on to apply the approach it outlined to the specific dispute at issue, since it also found that the rules of the other treaty that was being invoked were not ‘relevant’ in the sense required by the VCLT. In the case of the ITA, while it is obviously a treaty to which only a subset of WTO Members adhere, the ITA creates expectations as to how the parties will schedule concessions which they are then bound to extend to all WTO Members. In this respect, considering the ITA would, we argue, have been consistent with the Appellate Body's concern in EC and Certain Member States–Large Civil Aircraft about relevance to the common intentions of all WTO Members, when the agreement in question only binds a subset of them. Having found (probably wrongly) that the ITA fell under 31(2)(b), the Panel came to the conclusion that only the binding or operative provisions of the ITA would be relevant as context to the interpretation of the EC's Annex to its schedule and Headnote. It is difficult to discern the logic of this distinction. In fact, the preamble of an Agreement might well be the part of the text that discloses most fully the parties’ intended relation of that Agreement to the main treaty instrument being interpreted, and therefore be highly relevant as ‘context’.

The fact that an instrument is being invoked as context hardly suspends the interpretative canons of VCLT 31(3)(c) in respect of that instrument. Invoking an instrument as context obviously requires that the instrument in question be interpreted: and where to go other than the VCLT for those rules of interpretation (at least if the instrument in question is a treaty). The VCLT rules, in turn, justify considering the preamble in interpretation.

The autonomy of members to schedule non-HS-based concessions

A major dimension of the EC defense was, as noted, the idea that the non-HS-based narrative description that made up the general concession in the Headnote was still subject to the proper classification of each individual product under the EC's HS classifications. In effect, the EC was saying that it should not have been understood as making a concession, or perhaps even that it could not, within the WTO tariff architecture, make a concession, that bypasses the HS entirely. The Panel, consistent with the Appellate Body ruling in Chile–Price Band System,Footnote ³⁴ is clear and indeed adamant that WTO Members have the autonomy to create concessions using an entirely different or alternative nomenclature, conditions, or qualifications than those found in the HS system. What Members cannot do (as was clear in Chile–Price Band System) is to restructure their tariff bindings with the result that existing HS-based concessions are not honored, or the products to which they apply are not correctly classified based on HS nomenclature to six digits. Superimposition of additional, further liberalizing concessions based on a different methodology is entirely acceptable, provided it is consistent with MFN and transparency obligations in the GATT. There is a fine distinction but an important one: under the WCO system, there is a requirement to employ to six digits the HS classifications in the creation of domestic nomenclature; but the WTO permits additional liberalization commitments on other grounds than the characteristics of products as denoted in the nomenclature.

Howse and van Bork (Reference Howse and van Bork2006) have made this point, and developed its significance, in the context of the WTO negotiations on Environmental Goods and Services. ‘WTO negotiators should regard themselves as the clients or “masters” of the HS; this classification system is there to serve their needs, not to impose disciplines and obstacles on trade liberalisation efforts … While the Harmonized System does not provide classifications that correspond to the “environmental” properties of products, WTO customs classification practice is flexible enough to accommodate “ex-outs” and national nomenclature below the six-digit classification level. The WTO Information Technology Agreement “B” list is an example where tariff liberalisation commitments have been made on products, without prejudice to their classification within the HS.’ The ruling of the instant Panel affords important certainty to WTO members that they may proceed to give preferential tariff treatment to products based on environmental or social criteria for example, and bind additional concessions on that basis, despite the lack of such a basis for distinguishing products in the existing HS system. Here we note that the autonomy from the HS, according to the Panel, goes so far as to preclude its use even as ‘context’ for the interpretation of concessions scheduled on alternative bases.

Transparency

The core of the EC's argument on transparency was that in order to be subject to transparency obligations in GATT Article X, regulatory activity has to have certain formal legal or administrative characteristics such that it can be said to come within the expression ‘laws, regulations, judicial decisions and administrative rulings’. Had the Panel accepted this interpretation, the perverse consequence would have arisen that, by resorting to the most informal and indirect methods of achieving regulatory outcomes, that is the least transparent, Members would be able to circumvent the Article X transparency obligations in question. One finds a parallel here to the EC argument in the EC–Asbestos disputeFootnote ³⁵ that outright bans on products fall outside the discipline of the TBT Agreement; had the Appellate Body not in that case reversed the Panel's acceptance of the EC argument, Members would be able to avoid TBT disciplines altogether − including ironically the requirement of least-trade restrictiveness − by adopting the most restrictive form of trade measure possible.

In sum, we commend the instant Panel's purposive interpretation that the expression ‘laws, regulations, judicial decisions and administrative rulings’ in Article X is intended to reach a very wide range of regulatory activity, regardless of the exact legal or administrative form that the activity takes in a particular domestic administrative and constitutional-law system.

Article II of the GATT and ambiguity in classification

As noted above in the summary of the Panel's findings concerning printers, the relevant EC concession pursuant to the ITA was based on an HS classification, and therefore interpretation of the HS system was the central issue with respect to this part of the complaint. As with the earlier EC–Computer Equipment (LAN) and EC–Chicken Cuts cases, the question was whether the product fell under one of two possible subheadings, and implicitly the degree of discretion of the member in question to resolve any ambiguity one way or the other. In the instant dispute, the complainants argued that the correct classification for the printers in question was subheading 8471 60, ‘Automatic data-processing machines and units thereof’; ‘Input or output units, whether or not containing storage units in the same housing’. The European Communities insisted that in some cases, that is where the photocopying function of the device met certain performance standards, it was justified in classifying it instead under another concession under the EC Schedule, that is that of HS1996 subheading 9009 12 (‘indirect process electrostatic photocopying apparatus’).

In EC–Chicken Cuts, the Appellate Body had been faced with a situation where, arguably the product in question, chicken that was both frozen and salted, could have been classified either under an HS subheading that applied to frozen chicken or one that applied to salted chicken. The Appellate Body held that in order to prove a violation of Article II of the GATT, the complainant had only to prove that the product could be classifiable under the subheading that provides the lower rate of duty (in this case the salted category); it did not have to show that the product was incapable of being classified under the alternative classification that carries with it a higher rate of duty. In effect, the Appellate Body refused to consider seriously the possibility that certain products might have a combination of characteristics such that they could be classified plausibly or reasonably in either of the two classifications. On this point, Horn and Howse (Reference Horn, Howse, Horn and Mavroidis2008) criticized the Appellate Body for not considering explicitly the possibility that a Member's obligations under Article II of the GATT could be discharged: the case where a product could be classified reasonably and objectively under either of two headings, that is by applying the duty corresponding to either heading, not necessarily the lower rate of duty associated with one of the two.

With respect to printers with photocopying and fax capacity, the Panel noted that the WCO in its response to the Panel's questions concerning the meaning ‘of a kind’ suggested that given that the WCO had not taken a view on this nomenclature, that it was left to individual administrations to interpret on a product-by-product basis:

Despite the WCO raising the possibility of a margin of deference to individual Member's interpretations of terms left open-ended or not fully defined by the WCO, the Panel – here following it would seem the approach of the Appellate Body in EC–Chicken Cuts – suggests that it is for the Panel to determine the correct interpretation of each term in a Member's schedule, not merely whether the Member's interpretation for customs-administration purposes is an objective and reasonable one. One of the challenges in relying upon definitive interpretations by the WCO – as opposed to non-binding advice from its Secretariat – is that the HS Committee, which alone can issue such definitive interpretations, meets only twice a year. There is no obvious basis in the DSU for the equivalent of lis alibi pendens, which would allow suspension of a WTO dispute-settlement proceeding pending a definitive interpretation by the HS Committee. In the absence of lis alibi pendens, the Panel focusing its inquiry not on the correctness but the reasonable and objective character of the Member's classification practice has the salutary effect of not preempting, as it were, a subsequent definitive interpretation by the WCO. Nevertheless, it is notable (and somewhat of an improvement over the Appellate Body's approach in EC–Chicken Cuts) that the Panel did feel it necessary to consider not just the scope of the ‘input output’ classification but also the competing photocopier classification. Of course, as noted, the Panel found that the latter applied only to analogue photocopiers. The result was that the Panel was able to determine not only that the products at issue could be classified under the first concession, but also that they could not be classified under the latter. This raises the issue of what the Panel would have done had it found that the photocopier classification also applied to digital copying technologies. There is a hint in its discussion of set-top boxes of what the Appellate Body would have done in such a situation and that it would have applied a somewhat different logic than the Appellate Body in Chicken Cuts – namely, the Panel might have held that, on a product-by-product basis, it was for the EC to make a holistic judgment about whether, given all of the relevant characteristics and functions of each product, it made more sense to classify it as ‘input−output’ or a photocopier. This implies a degree of discretion to make a good-faith determination, based on all the relevant characteristics of a product, as to whether it falls into one of two possible or plausible classifications (because each captures some salient features of the product or combination thereof), even if the exercise of this discretion may result in choosing the classification to which the higher rate of duty applies.

But the underlying conceptual issue, avoided or ignored by the Appellate Body in EC–Chicken Cuts and not directly engaged by the instant Panel, could be framed as follows: if we acknowledge a genuine, irreducible element of ambiguity, which cannot be resolved by a Member's good-faith application of HS rulings, guidelines etc., is the nature of state responsibility under Article II such that a Member is always required to apply the classification that results in the lower rate of duty? If we return to our discussion of state responsibility above, and the significance we attached to providing certainty and predictability to private economic actors with a view to maximizing the welfare gains from tariff liberalization, such a lex specialis of state responsibility may make sense, at first glance. At the same time, it would risk some private economic actors manipulating the characteristics of products so as to create ambiguity, on the basis of knowing that ambiguity will always result in the lower rate of duty being applied.