Climate Labelling and the WTO: The 2010 EU Ecolabelling Programme as a Test Case Under WTO Law

Abstract

Environmental labelling is increasingly used as an instrument of climate protection. This is underlined, for example, by the EU climate change programme, in which various labelling schemes are employed. Cases in point are the EU’s oft-discussed voluntary ecolabelling scheme, which takes a life-cycle approach, and its mandatory labelling scheme for cars.

This contribution draws on several chapters in Vranes, Trade and the Environment. Fundamental Issues in International Law, WTO Law and Legal Theory, 2009, in particular pp. 172 et seq., 191 et seq., 223 et seq., 256 et seq., 302 et seq., 319 et seq., 342 et seq., 379 et seq.; updated in February 2010, it takes into account the new 2010 EU ecolabelling programme, which entered into force in February 2010, and which is described in the following (see pp. 209 et seq.).

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction and Legal Background

Labelling as an Instrument of Climate Protection

Environmental labelling is increasingly used as an instrument of climate protection.¹ This is underlined, for example, by the EU climate change programme, in which various labelling schemes are employed. Cases in point are the EU’s oft-discussed voluntary ecolabelling scheme, which takes a life-cycle approach, and its mandatory labelling scheme for cars.² A further example is the recent discussion on ‘CO₂ backpacks’ in the UK and Austria, i.e. labels on the amount of CO₂ emissions generated by the national and international transportation of foodstuffs.³ After the dubious outcome of the multilateral 2009 Copenhagen Climate Conference,⁴ the importance of instruments of this type may increase even further.

Both mandatory and voluntary labelling schemes risk contravening WTO law: whereas mandatory labels restrict market access for non-complying products, labels that are granted under a voluntary scheme are meant to improve the perceived attractiveness of products that are awarded the label;⁵ hence, such labels may negatively affect the competitive conditions of other products, possibly disadvantaging imported products.

Labelling schemes according to which information on a product’s environmental impacts over its life cycle is included in a pertinent label fall into the category of process-based labels. It is well-known to WTO experts that such labelling schemes – in particular those based on ‘non-product-related processes and production methods’⁶ – raise a considerable number of issues under WTO law.

This contribution examines the EU’s voluntary ecolabelling scheme, which was revised in 2000⁷ and 2010.⁸ Owing to its life-cycle approach, which since the scheme’s inception has taken into account, e.g., energy consumption during production and use (besides further environmental impacts), this scheme has been intensely debated in trade and academic circles since its first version was introduced in 1992.⁹ The new 2010 EU scheme similarly considers the whole life cycle of products, including ‘the most significant environmental impacts, in particular the impact on climate change’.¹⁰ Therefore, and in view of the broad range of issues raised by it, this scheme presents a model test case under WTO law also for other climate-related labelling schemes and ecolabelling schemes more generally.

Process-Based Measures, Process-Based Labelling and WTO Law: The Main Questions

Labelling programmes such as the EU scheme that take into account a given product’s environmental impacts raise a classic set of questions in WTO law: these revolve around the issue of whether and to what extent a WTO Member is allowed to introduce measures affecting trade in goods that are concerned with process and production methods (PPMs) which are not related to the goods concerned in the sense of bearing on their physical characteristics (non-product-related PPMs, or NPR PPMs).¹¹

The notion ‘NPR PPM requirements’ is derived from the 1979 GATT Agreement on Technical Barriers to Trade¹² and primarily refers to measures that target the production of goods, i.e. the stage before they are placed on the market.¹³ It is held by many WTO Members and a majority of publicists that such process-based measures are to be treated differently from product-related regulations under WTO law. Thus, it has repeatedly been held, for example, that physically similar products that differ only in their production or processing methods must be regarded as like products and must always receive identical treatment; this would incur the consequence that any measures that differentiate between like products on the basis of NPR PPMs would inevitably violate clauses such as Article III of the GATT.¹⁴ It has also been argued that NPR PPM-based measures invariably have to be found to be de facto discriminatory, even when they are drafted in origin-neutral terms, on the basis that such measures alter competitive conditions.¹⁵ Furthermore, it has been held that such measures need to be justified under the GATT, even if they are non-discriminatory; moreover, it has even been contended that justification may be impossible in respect of such measures.¹⁶

This doctrine, which treats products and processes differently, is often referred to as the ‘product–process doctrine’, even though – in view of the many variants just mentioned – there is no uniform doctrine. These various views have in common that they break with the GATT system as it relates to ‘standard’, i.e. product-related, measures, which, for example, are not questioned if they are neither de jure nor de facto discriminatory, and which can be justified if they discriminate against foreign products. With the entry into force of the TBT Agreement in 1995, the additional question has arisen whether this agreement applies to NPR PPM requirements. Although the product–process divide is relevant for the status, under WTO law, of NPR PPM-based measures in general, a closely related, but partly self-standing discussion has arisen as to the status of NPR PPM-based environmental labelling schemes, which serve to promote products that are perceived as environmentally friendly owing to their production and processing methods.

The afore-described issues are largely unresolved in WTO practice and academic debate.¹⁷ Their importance is evident, however, given that measures addressing production requirements are significant tools of environmental policy-making in line, in particular, with the rectification-at-source principle. Thus, environmental PPM requirements, in general, and PPM-based labelling schemes, in particular, can be used to address local concerns in the regulating state or another state, transboundary pollution as well as transboundary living resources, and global concerns such as climate change and the protection of the ozone layer.¹⁸ Hence, if process-based measures affecting trade were prevented by WTO disciplines to a greater degree than product-related measures, the resulting structural imbalance might be perceived as problematic from an environmental point of view.

Overview of the 2010 EU Ecolabelling Scheme

The EU first introduced an ecolabelling scheme in 1992¹⁹ and this has attracted considerable attention in the literature and international fora.²⁰ In 2000, the EU introduced a revised ecolabelling mechanism which built upon the principles, but abrogated the legal basis of the 1992 scheme.²¹ Like the 1992 regime, this mechanism established a voluntary ecolabel award scheme that intended to promote products with a reduced environmental impact during their entire life cycle. Given that the 1992 and 2000 mechanisms did not prove successful,²² the EU decided, in 2009, to amend the ecolabelling scheme once more.²³

Like the former ones, the new 2010 mechanism pursues a life-cycle approach on a voluntary basis,²⁴ aiming at reducing the negative impact of consumption and production of products on the environment, health, climate and natural resources.²⁵ It is administered by the EU in cooperation with independent competent bodies of the member states²⁶ and the European Union Ecolabelling Board (EUEB).²⁷ Like its predecessors, it strives to promote products which appear more environmentally friendly, during their entire life cycle, than other similar products²⁸ (for the purpose of the EU scheme, the term ‘products’ also encompasses services)²⁹. Therefore, the Ecolabel criteria are set by reference to groups of similar products,³⁰ so only products with superior environmental performance within a given group may receive the EU label.³¹ Relevant criteria are, in particular, a product’s ‘most significant environmental impacts’ including the impact on climate change, nature and biodiversity, energy and resource consumption, emissions, as well as its durability and reusability, and social and ethical aspects.³² Labels are to be awarded on the basis of continuously updated scientifically based information, taking into account appropriate internationally recognized standards.³³

Relevant labelling criteria are set and reviewed in a procedure which involves the Commission, the Member States, competent bodies, the EUEB and other stakeholders. The EUEB consists of representatives of competent national bodies and other interested parties. Member states must ensure that the composition of the competent bodies, within ministries or outside, guarantee their independence and neutrality and that their rules of procedure warrant transparency and the involvement of all interested parties at the national level.³⁴ Following consultation of the EUEB, the Commission, Member States, competent bodies and other stakeholders, which have demonstrated relevant expertise, may initiate and lead the development or revision of EU Ecolabel criteria for a given product group.³⁵ Operators wishing to use the EU Ecolabel have to apply to the aforementioned competent bodies. They can do so also for products originating outside the EU.³⁶ Upon award of the label, the competent body is to conclude a contract with the applicant which lays down the terms of use of the label.³⁷ Ecological criteria were set out, already under the 2000 scheme, for the award of the Ecolabel for a series of products, including personal and portable computers, television sets, dishwashers, washing machines and tourist accommodation services.³⁸

Issues in WTO Law

Although the EU scheme also covers environmental labelling of services,³⁹ the present analysis focuses on labelling of goods for reasons of space constraints and the complexity of the issues that present themselves already in the goods sector. For the same reasons, and since these issues have been treated at length elsewhere, this contribution does not address the questions of the unilateral and purported extraterritorial character of NPR PPM-based measures and labelling schemes.⁴⁰ It first tries to classify the main types of labelling schemes, then turns to the question of the applicability of the GATT and the TBT Agreement to a voluntary labelling mechanism such as the EU scheme, and finally examines central issues in substantive WTO law.

Taxonomy of Labelling Schemes

Generally speaking, it is possible to categorize labelling schemes pursuant to three criteria, that is (1) the issue of government involvement (whether the scheme is administered by public authorities or is privately sponsored), (2) its legal effect (whether labelling is mandatory or voluntary) and (3) its scope (whether it applies to product-related characteristics and product-related PPMs, or whether it – also or exclusively – covers NPR PPMs).⁴¹

Labelling schemes administered by public bodies can be subdivided into mandatory and voluntary ones. A labelling system is regarded as mandatory when the award of the label functions as a legally binding market access requirement; otherwise it is classified as voluntary.⁴² A scheme which functions on a voluntary basis may nonetheless affect the competitive relationship between similar products, and it is normally even meant to bring about this effect: this follows from the common understanding that voluntary labelling schemes are designed to ‘inform consumers and thereby promote consumer products which are determined to be environmentally more friendly than other functionally and competitively similar products’.⁴³ Moreover, both mandatory and voluntary schemes – whether privately or state-administered – can be further distinguished into product-related approaches and NPR PPM-based labelling. This yields the following taxonomy:

• (1) Mandatory government-administered labelling schemes based on product-related characteristics, including product-related PPMs.

• (2) Mandatory government-administered labelling schemes – additionally or exclusively – based on NPR PPMs.

• (3) Voluntary government-administered labelling schemes based on product-related characteristics, including product-related PPMs.

• (4) Voluntary government-administered labelling schemes – additionally or exclusively – based on NPR PPMs.

• (5) Privately sponsored labelling schemes based on product-related characteristics, including product-related PPMs.

• (6) Privately sponsored labelling schemes – additionally or exclusively – based on NPR PPMs.

These labelling schemes raise partially divergent questions under WTO law.⁴⁴ The EU ecolabelling mechanism corresponds to type 4 of this taxonomy, given that it is preponderantly administered by public authorities, is voluntary in nature (i.e. market access is not de jure dependent on the fulfilment of the underlying labelling criteria) and includes NPR PPM-based requirements owing to its life-cycle approach.

Applicability of the TBT Agreement, and of the GATT

Applicability of the TBT Agreement

The issue of whether voluntary, government-administered NPR PPM-based labelling schemes come under the TBT Agreement has been designated as one of the most debated questions in the discussions of the WTO Committee on Trade and the Environment.^{45 , 46} Whereas the EU, Switzerland and Canada have expressed the view that such labels are covered by the TBT Agreement (and do not constitute per se violations of the agreement),⁴⁷ some developing WTO Members have been arguing that the negotiating history of the TBT Agreement shows that NPR PPM-based measures are not covered by the TBT Agreement, as its drafters had no intention of ‘legitimizing’ NPR PPM-based measures.⁴⁸ The view that such labelling schemes do not come under the purview of the TBT Agreement is also maintained in recent academic writings,⁴⁹ and is even regarded as the prevailing opinion.⁵⁰

Regarding this contention, however, it has to be stressed that the relevant negotiating history of the TBT Agreement can be characterized as being ambiguous at best, and that systematic-teleological interpretation quite clearly leads to the conclusion that the TBT Agreement is applicable to NPR PPM-based measures in general and – by implication – to labels in particular.⁵¹ The contrary stance taken by several writers and notably developing countries sometimes appears to be based on the misunderstanding that the non-applicability of the TBT Agreements would per se prohibit the introduction of NPR PPM-based requirements by other (developed) WTO members. The core of this possible misunderstanding seems to be rooted in the misconception that the TBT Agreement permits measures that otherwise would be prohibited. However, the TBT Agreement does not introduce permissions; rather, it lays down new obligations, i.e. disciplines that apply in addition to those of the GATT, in particular. Hence, its purported non-applicability would not imply that NPR PPM-based requirements would be per se prohibited (such measures may, however, come under the purview of the GATT⁵²). Moreover, if one takes the view that the TBT Agreement imposes disciplines that tend to be stricter than those of the GATT, then (developing) countries that are troubled by the spectre of NPR PPM-based labelling schemes arguably should in fact advocate the TBT Agreement’s applicability.

Additionally, it has to be noted that the TBT Committee decided in 1997 that the ‘obligation to publish notices of draft standards containing voluntary labelling requirements under paragraph L of the [TBT Code of Conduct] is not dependent upon the kind of information provided on the label’.⁵³ Despite a pertinent disclaimer,⁵⁴ this decision can arguably be interpreted as an indication that there is some convergence of views at least that NPR PPM-based labels should not be regarded as being per se excluded from the scope of the TBT Agreement.⁵⁵

Hence, if one takes the view that NPR PPM-based labelling schemes are not exempted from the scope of the TBT Agreement, then the EU ecolabelling mechanism, being voluntary in nature, must comply fully with the TBT Agreement and its Code of Good Practice in particular.

Applicability of the GATT

Since voluntary labelling schemes are meant to affect the competitive conditions among similar products, the question arises whether such mechanisms come under the purview of the GATT, which, pursuant to the General Interpretative Note to Annex 1A, applies beside the TBT Agreement to the extent there is no conflict between both agreements. This leads to the issue of the attribution of (partially) private conduct, which risks impinging on the order set up by GATT disciplines, to WTO Members.⁵⁶ Importantly, the same question would arise if a panel followed the stance, which was refuted in the preceding section, that the TBT Agreement does not apply to the EU scheme: this scheme would then be governed solely by the GATT. Both considerations constitute reasons why one should examine the applicability of the GATT to voluntary governmental labelling schemes.

As noted, if a product is considered eligible for the award of the EU label, the competent national body is to conclude a contract with the applicant, which covers its terms of use.⁵⁷ The ‘competent body’ may arguably be a governmental or a private body.⁵⁸ In either case, it must be independent,⁵⁹ so that the contract can be concluded between a private party, on the one hand, and an independent governmental or private body, on the other.

In such instances, it has repeatedly been questioned whether eventual distortions of competitive conditions can be attributed to the state. As this issue has already been dealt with in detail elsewhere,⁶⁰ it shall be recalled in the present context that several GATT panel reports have analysed this question.⁶¹ Although these decisions were rendered under individual GATT provisions (namely Articles III:4, XI:1 and XXIII:1(b) of the GATT, respectively), the uniform and generalizable underlying theme is that under the GATT the conduct of private persons will be attributed to the state when they are sufficiently influenced through ‘incentives and disincentives … to act in [this] particular manner’.⁶² This approach, which also interlocks with public international law guidelines and theoretical considerations in jurisprudence,⁶³ is teleologically justified by the fact that the GATT is concerned with non-discriminatory market access and competitive conditions in the internal market, which may also be influenced indirectly by the state through behaviour which appears not to emanate from it in form, but does so in substance.

It follows by implication that the EU environmental labelling scheme can be attributed to public authorities under the GATT, in view of the facts that the system as such has been established by EU legal acts,⁶⁴ that it is preponderantly administered by the EU and that the private bodies involved act, in large part, under mandates of the EU Commission and EU Member States.⁶⁵

In sum, and this corrects a frequent misunderstanding,⁶⁶ the EU’s voluntary ecolabelling scheme is neither exempted from nor prohibited per se under the disciplines of the TBT Agreement or the GATT solely owing to its reliance on life-cycle considerations. Hence, the scheme and existing and future implementing measures must be fully in compliance with the disciplines of these agreements.

Likeness

It is particularly disputed whether and how the fact that a regulatory measure is concerned with non-product-related production and processing measures affects the likeness analysis under central WTO trade in goods provisions such as Article III of the GATT and analogous clauses in the TBT Agreement.⁶⁷ As noted, the fact that several authors take the view that physically similar products that differ only in their production or processing methods must be regarded as like products would – according to many writers – incur the consequence that they must always receive identical treatment. In other words, regulatory distinctions based on environmentally (un-)friendly NPR PPMs would be prohibited (subject to eventual justification under a clause such as Article XX of the GATT, unless one does take the view that even justification is impossible for process-based measures⁶⁸). Therefore, the present section first analyses the notion of ‘like products’ in the GATT and the TBT Agreement on a general level. It then moves on to the specific nexus between likeness of products and NPR PPMs. In the last subsection, the results of this analysis are applied to the EU labelling scheme.

Likeness in the GATT and the TBT Agreement

This section first examines the meaning of ‘like’ and ‘like products’ in the GATT. It then turns to the interpretation of the similar wording of Article 2.1 of the TBT Agreement and the analogous provision of Article D of the Code of Good Practice (Annex 3 of the TBT Agreement), which applies to non-mandatory labelling requirements.⁶⁹

Given that the terms ‘like’ and ‘like products’ are not explicitly defined in Article III:4 of the GATT (nor in Article III:2 first sentence of the GATT, Article 2 of the TBT Agreement or Article D of the Code of Good Practice, which will be analysed later), one has to turn to the context of these terms and the object and purpose of Article III, and eventually of the GATT and WTO law more generally.⁷⁰

A close part of this context is Article III:1, according to which internal taxation and internal regulation ‘should not be applied to imported or domestic products so as to afford protection to domestic production’. Article III:1 is not a norm that is applicable to a concrete case in itself, but constitutes an interpretative principle that is relevant for the interpretation of Article III as a whole,⁷¹ as ensues from its wording.⁷² Being an interpretative principle, it is of special importance for construing the term ‘like products’ in Article III:4 (and Article III:2 for that matter).

Although the exact import of this clause is subject to intense debate in academic writing, there is consensus in general that Article III:1 makes it clear that the function of Article III as a whole is the avoidance of protectionism, a reading which is in line with the overall telos of WTO law.⁷³ Therefore, contextual as well as teleological arguments point to the importance of avoiding protectionism in favour of domestic products as the relevant background for interpreting the term ‘like products’. Moreover, a regulatory intervention, in order to be protectionist in nature, requires that there be a competitive relationship between the domestic products protected and the disfavoured foreign products, since otherwise the protectionist effect would not normally be felt. Hence, Article III should be understood as being primarily concerned with products that are in such a competitive relationship.⁷⁴ Therefore, the term ‘like products’ should be interpreted as a term requiring an examination of the legally required intensity of the competitive relationship between domestic and foreign products.

This is further corroborated by the fact that Article III:2 second sentence, as clarified by the Note ad Article III, refers to ‘directly competitive or substitutable’ products: it follows from the two-sentence structure of Article III:2 and the wording of Article III:2 second sentence⁷⁵ that this sentence is meant to function as a subsidiary clause which shields ‘directly competitive or substitutable’ (DCS) products from protectionist interventions. Hence, DCS products can be regarded as a broader category which comprises like products as a subgroup, in which the competitive relationship is even more evident.^{76 , 77}

In sum, the context established by Article III:1 and Article III:2 second sentence as well as the telos of Articles III:2 and III:4 and the overall object and purpose of WTO law indicate that ‘likeness’ in Article III should be interpreted against the background of protectionism: ‘like products’ should primarily be understood to mean products that are in a competitive relation that is even closer than that of DCS products.

The decisive question is therefore that of when competition does exist between two products. It is obvious that competition inherently depends on consumer perception: even products that differ in their physical appearance and in respect of other criteria may be competitive if they are regarded as equivalent – that is as being interchangeable to a sufficient degree – by consumers; by the same token, products which appear quite similar with regard to criteria such as physical characteristics may theoretically be treated as dissimilar and non-competitive by consumers.

This focus on competition not only has the consequence of making the perspective of consumers central to the determination of likeness. Since consumer perception will normally be influenced above all by product-related criteria, it is a further corollary of this view that the relevance of the perspective that a regulator may have on the similarity of products finds no obvious confirmation in Articles III:2 and III:4. In other words, regulator-related interests should not normally be regarded as relevant in the determination of likeness.⁷⁸

Similarly, it is not convincing that international environmental agreements should automatically be relevant in the determination of likeness,⁷⁹ since government interests, even if they are expressed in international agreements, cannot be regarded to be relevant per se in the likeness context as it is structured by the GATT.⁸⁰ A third consequence (to be examined in the next subsection) of the submission that the pertinent perspective in the determination of likeness is the perspective of consumers is that it appears possible that even processing methods which do not physically affect the product are regarded, by consumers, as rendering otherwise like products unlike.⁸¹

It is submitted that this interpretation of the term ‘like products’ clearly converges with pertinent GATT/WTO dispute settlement practice, which – in relying on the 1970 report of the Working Party on Border Tax Adjustments (BTA) – in particular refers to the following criteria in the determination of likeness: ‘the product’s end-uses in a given market; consumers’s tastes and habits, which change from country to country; the product’s properties, nature and quality’.⁸² It can be argued that these criteria in general reflect the aforementioned focus on competition and consumer perspective.⁸³ The central importance of competition and the inherently intertwined perspective of consumers have meanwhile also been explicitly highlighted in Appellate Body jurisprudence.⁸⁴

Despite the emphasis on competition and consumer perspective, there remains a plurality of criteria in any given case, some of which may militate in favour of likeness, whereas others may indicate dissimilarity. Therefore, it seems appropriate that the Appellate Body has cautioned that the notion of likeness (possibly) varies, its scope depending on the applicable GATT non-discrimination provision, its context and the concrete case.⁸⁵

This consideration is relevant also with respect to Article 2.1 of the TBT Agreement and Article D of the Code of Good Practice, where the term ‘like products’ is pivotal as well and where it remains undefined too. Just as in the GATT, one therefore has to examine the context and telos of this provision. It then quickly becomes clear that Article 2.1, Article D and the TBT Agreement more generally, like the GATT, are concerned with abolishing ‘unnecessary obstacles to international trade’⁸⁶ and, hence, with ensuring international competition. This, and the fact that the TBT Agreement can be regarded as a concretization of the GATT, implies that the arguments which have just been presented with regard to the GATT apply within the TBT context as well. Therefore, in Article 2.1 of the TBT Agreement and the analogous provision of Article D of the Code of Good Practice as well, ‘like products’ should be understood to mean products that are in a close competitive relationship, a determination that has to be made primarily from the perspective of consumers.

Likeness and NPR PPM-Based Measures

It has been contended that the ‘most logical conceptual basis’ for a product–process doctrine, which distinguishes between regulation of products and NPR PPM-based measures, is the concept of ‘likeness’ in the ‘like product’ test of GATT Article III.⁸⁷ According to this view, by redefining the likeness concept, a panel could comply with its ‘belief’⁸⁸ that ‘the relevant community recognizes a normative obligation to limit a certain activity in a certain way’.⁸⁹

This approach converges with views pursuant to which divergent PPMs cannot affect the likeness of otherwise similar products.⁹⁰ Such an approach to the concept of likeness is hardly defensible. The crux of the issue arguably lies in the fact that the terms ‘product-related’ and ‘non-product-related’ seem to imply a (quasi-)scientific approach: if traces of a given process or production method are not physically ascertainable in the final product, then the PPM in question is regarded as non-product-related.⁹¹ This issue must, however, be distinguished from that of the likeness judgment, which is not exclusively concerned with the physical traceability of a given process or production method in the final product, but – as explained above – with the competitive relation that prevails between the products in question.

Given that the competitive relationship is inherently influenced by consumer perception,⁹² it follows that PPMs which do not leave physical traces in the final product (and which are not product-‘related’ in any physically ascertainable way) may nonetheless be perceived, by consumers, as being ‘related’ to the product: if such PPMs are prone therefore to affect the competitive relation on the market, then this may constitute an indication that otherwise similar products may be unlike nonetheless.

This eventual indication of unlikeness must be balanced with other relevant indications militating in favour of likeness, however. It has rightly been emphasized in recent writings that a product’s different production history may render it unlike other products,⁹³ even if this will be the exceptional case rather than the rule.⁹⁴

Likeness and the EU Ecolabelling Scheme

The EU labelling scheme exemplifies well the risks of (inadvertent) discriminatory treatment of imported products that are inherent in the setting of labelling criteria. As noted, the ecolabel is awarded to those products within a given product group that fulfil the labelling criteria defined by the EU. Under the Community scheme, product group means ‘a set of products that serve similar purposes and are similar in terms of use, or have similar functional properties, and are similar in terms of consumer perception’.⁹⁵ This definition largely overlaps with the definition of like products in the TBT Agreement and the GATT that was suggested above. Nonetheless, since the determination of likeness is a context-related value judgment, some products which may not be found to be ‘like’ in terms of WTO law may be included in the same product group under the EU labelling scheme. Inversely, products which are not included in a product group that is defined under the EU scheme may have to be considered under WTO law to be ‘like’ the products encompassed in the EU product group; hence, like products risk being excluded from having access to an ecolabel. This raises the risk of discrimination that will be discussed in the next subsection.

Less Favourable Treatment

The EU ecolabelling regime does not introduce de jure discriminatory treatment, as it does not explicitly differentiate between products on the basis of their origin, given that the EU ecolabel can also be awarded, under the same conditions, to products originating outside the EU.⁹⁶ However, the fact that products that are not eligible, in terms of EU law, for the EU label may appear, in terms of WTO law, ‘like’ other products that are awarded the EU label incurs the risk of de facto discrimination under Article III:4 of the GATT and Article D of the Code of Good Practice (i.e. the aforementioned counterpart of Article 2.1 of the TBT Agreement), which applies to non-mandatory labelling requirements.

Since the notion of de facto discrimination is disputed in WTO law in general and as regards NPR PPM-based measures in particular, the next subsection addresses these issues. The second subsection then deals with the question of whether the EU ecolabelling scheme constitutes less favourable treatment in the sense of the GATT and the TBT Agreement.

PPM Requirements and the Concept of De Facto Discrimination

The product–process doctrine, pursuant to which regulation of products, on the one hand, and that of NPR PPM-based measures, on the other, are to be treated differently under WTO law, is also intricately intertwined with the so-called ‘diagonal test’ in determining the existence of de facto discrimination and the so-called aims and effects or regulatory purpose approach to likeness, which is a complement of the ‘diagonal test’.⁹⁷

The diagonal test’is a method that strives to determine whether regulatory treatment is de facto discriminatory. It does so by merely comparing a disadvantaged subgroup of foreign like products (subgroup 2 in the diagram) with that of the most favoured subgroup of domestic like products (subgroup 3 in the diagram), even if the latter subgroup consists of very few products.⁹⁸ Thereby, this test disregards whether there also exists a subgroup of foreign like products (subgroup 4 in the diagram) that receives treatment similar to that accorded to the most favoured domestic subgroup (subgroup 3 in the diagram). Therefore, the diagonal test also disregards whether the proportions of the favoured and disfavoured subgroups are equal for domestic and imported products.

This is problematic, given that according to the traditional approach to discrimination a measure is de facto discriminatory if it produces a disproportionate disparate impact on foreign products,⁹⁹ a view which is confirmed by the object and purpose of Article III, i.e. the prevention of protectionism. According to this traditional view, one has to compare the treatment accorded to the two entire groups of like domestic products (comprising subgroups 1 and 3) and like foreign products (comprising subgroups 2 and 4), a method also applied, e.g., by the European Court of Justice (ECJ).¹⁰⁰ Thus, the ECJ inquires into whether imported products preponderantly fall into the disadvantaged group (subgroup 2) and whether domestic products preponderantly fall into the class of privileged products (subgroup 3).¹⁰¹ More precisely, the ratio between domestic favoured and disfavoured products must be roughly equivalent to the ratio between foreign favoured and disfavoured products.¹⁰² Meanwhile, this approach has arguably also been applied by the Appellate Body in its much discussed EC – Asbestos ruling.¹⁰³

The aforementioned connection between PPM requirements and the diagonal test is shown, e.g., by a finding in the panel report in US – Malt (rendered under the largely analogous provision of Article III:2 of the GATT), which concerned tax credits granted to small domestic breweries. In alluding to the product–process doctrine, this panel first ruled that production- or producer-related characteristics do not affect the nature of the product at issue. In its view, ‘beer produced by large breweries is not unlike beer produced by small breweries’. It then went on to argue that ‘even if Minnesota were to grant the tax credits on a non-discriminatory basis to small breweries inside and outside the United States, imported beer from large breweries would be ‘subject … to internal taxes … in excess of those applied … to like domestic products’ from small breweries and there would still be an inconsistency with Article III:2, first sentence’.¹⁰⁴ This dictum constitutes an application of the ‘diagonal test’ to determining whether regulatory treatment is de facto discriminatory, which – as has just been explained – merely compares a disadvantaged subgroup of foreign like products (in casu beer produced by large foreign producers) with that of the most favoured subgroup of domestic like products (beer produced by small domestic producers). As has just been pointed out, this test overlooks that there may also exist a subgroup of foreign like products (in casu foreign like beer from foreign small producers) that receives treatment equivalent to that accorded to the most favoured domestic subgroup; and it overlooks that the proportions of the favoured and disfavoured subgroups may, in a given case, actually be equal for domestic and imported products. Put differently, the panel’s approach disregards that it may be possible to draw distinctions in treatment between like products even on the basis of production methods that do not incur geographically disparate impacts on domestic and foreign products and do not, therefore, amount to de facto discriminatory treatment in terms of WTO law.

Furthermore, in this context, recent academic writings¹⁰⁵ appear to have overlooked that panel practice meanwhile has rightly found that process-based measures do not constitute a special case under the standard of differential treatment. Thus, in 2000, the Canada – Automotive panel rightly decided that a PPM-related import duty exemption ‘cannot be held to be inconsistent with Article I:1 simply on the grounds that it is granted on conditions that are not related to the imported products themselves. Rather, we must determine whether these conditions amount to discrimination between like products of different origins’.¹⁰⁶ The panel explained condemnations of process-based measures in earlier GATT practice,¹⁰⁷ some of which had in fact employed the formula ‘not related to the product’,¹⁰⁸ by pointing out that these cases had been concerned with discriminatory measures.¹⁰⁹

Hence, the panel’s reasoning a contrario confirms what has just been argued, namely that non-discriminatory NPR PPM requirements should be regarded as being consistent with non-discrimination disciplines of the GATT such as Article III:4 when they do not incur disparate impacts between foreign and domestic like products.¹¹⁰ Although this decision has been rendered under Article I of the GATT, the panel’s reasoning is clearly transposable to Article III as well. This view is also in conformity with academic writings according to which process-based measures should be regarded as being GATT-consistent if they do not constitute countrywide (i.e. directly discriminatory) measures.¹¹¹

Less Favourable Treatment and the EU Ecolabelling Scheme

As explained above, under the EU ecolabelling scheme there is a risk that products that are not eligible, in terms of EU law, for the EU label, may appear, in terms of WTO law, ‘like’ other products that are awarded the EU label. Such geographically disparate impact – if it occurs – risks being compounded twofold. On the one hand, imported products which are new on a given national market often particularly depend on the use of marketing means such as labels for successful market penetration. On the other hand, the EU scheme envisages the promotion of labelled products and the labelling mechanism itself;¹¹² this, too, tends to reinforce the competitive disadvantages of products that are not covered by the scheme.

Hence, theoretically there is a risk of disparate impacts on like domestic and foreign products that amounts to de facto discrimination if foreign products turn out to be disproportionately affected by the EU scheme. Whether this theoretical risk materializes in practice depends on a factual analysis which is beyond the scope of this paper, as such an analysis would have to be conducted for any given product group individually. The important point to note, however, is the fact that NPR PPM-based regulations in general and NPR PPM-based labelling schemes in particular do not per se amount to de facto discriminatory treatment.¹¹³

Justification of NPR PPM-Based Labelling Schemes

There is also disagreement as to whether and under which conditions NPR PPM-based measures, in general, and process-based labelling schemes, in particular, can be justified under the GATT and the TBT Agreement. These issues are analysed in the next two subsections.

Justification and NPR PPM-Based Measures in General

A further variation of the product–process doctrine has been developed under Article XX, the general exceptions clause of the GATT. Especially, the two unadopted Tuna panel reports led to a widespread belief¹¹⁴ that unilaterally imposed PPM requirements addressing extrajurisdictional¹¹⁵ concerns are per se incapable of justification under Article XX. If this view were correct, it would be relevant for NPR PPM-based measures which, like the EU ecolabelling scheme, also address transboundary and/or extrajurisdictional concerns and global commons such as the world climate.

In these reports, it was essentially held that such measures do not come under the ambit of Article XX, because otherwise a WTO member could ‘unilaterally determine … policies from which other contracting parties could not deviate without jeopardizing their rights under the General Agreement’,¹¹⁶ and because the NPR PPM requirements at issue were introduced ‘so as to force other countries to change their policies with respect to persons and things within their own jurisdiction’.¹¹⁷ Not least owing to the wide support for the Tuna I and Tuna II rulings by GATT contracting parties,¹¹⁸ it was often held that process-based measures cannot be reconciled with Article XX, even though a subsequent, albeit unadopted, report again indicated that a process-based measure may, in principle, be justified under Article XX.¹¹⁹

This particular prong of the product–process doctrine is difficult to reconcile with international environmental law (in particular Principle 12 of the Rio Declaration, which was arguably adopted under the impression of the first Tuna ruling merely 9 months after its adoption,¹²⁰ and para 2.20 of Agenda 21), which has an undeniable bearing on the interpretation of Article XX.¹²¹ These principles do not distinguish between product-related and process-based regulations. Moreover, unilateral trade measures concerned with transboundary and global concerns are not regarded as unjustifiable pursuant to these principles. Rather, such measures are ‘merely’ subjected to specific qualifications, in particular that they should be subordinated ‘as far as possible’ to cooperative efforts.

Although the Rio Declaration is not binding, its ‘evidential value’ regarding state intentions¹²² is obvious, given that it has been adopted by 176 states,¹²³ and is considered as expressing worldwide consensus¹²⁴ and as constituting ‘at present the most significant universally endorsed statement of general rights and obligations of states affecting the environment’ which partly restates customary law and partly endorses new and developing principles of law.¹²⁵ The guidelines, which are derivable from the Rio Declaration and Agenda 21, are to a considerable extent mirrored in both US – Shrimp rulings of the Appellate Body, which has pointed out that PPM requirements are not a priori excluded from the scope of Article XX. Rather, in its words, ‘conditioning access to a Member’s domestic market on whether exporting Members comply with, or adopt, a policy or policies unilaterally prescribed by the importing Member may, to some degree, be a common aspect of measures falling within the scope of one or another of the exceptions (a) to (j) of Article XX’.¹²⁶ In the subsequent 2001 Shrimp proceedings under Article 21.5 of the Dispute Settlement Understanding (DSU), the Appellate Body reaffirmed that its ruling on the conditional justifiability of process-based measures constitutes a statement of principle.¹²⁷

Justification of NPR PPM-Based Measures, and the EU Ecolabelling Scheme in Particular

Justification Under the GATT

There are three reasons why it seems appropriate to examine whether eventually discriminatory effects of a voluntary NPR PPM-based labelling regime such as the EU ecolabelling scheme can be justified under the GATT. First, as has been mentioned, the GATT applies beside the TBT Agreement to the extent that no conflict arises¹²⁸; and it was shown above that the GATT in principle applies to voluntary NPR PPM-based labelling regimes such as the EU scheme, insofar as it can be attributed to the EU and its member states. Second, as indicated above, it is contested by some WTO members and several authors whether the TBT Agreement applies to NPR PPM-based labelling scheme at all. Although the present contribution does not share this view, a WTO panel might do so; it would then have to scrutinize a labelling scheme such as the EU regime under the GATT. Therefore, the following considerations on the eventual justifiability of such labelling mechanisms arguably also have practical value, besides their doctrinal import. Third, there is the problem that the TBT Agreement and its Code of Good Practice do not contain a general exception clause modelled after Article XX of the GATT. As will be argued below, one way of overcoming this problem consists in regarding Article XX of the GATT as an overarching exception clause that may become relevant, as a fallback clause, also under the TBT Agreement. This third reason explains why it seems useful to examine the possibility of justifying NPR PPM-based labelling schemes under the GATT before turning to the analogous issue under the TBT Agreement.

Under the GATT, Articles XX(b) and XX(g) are primarily relevant. Under Article XX(b), it has to be shown that a given measure is necessary to protect human, animal or plant life or health. It must also be shown, under the introductory clause (‘chapeau’) of Article XX, that the measure is not applied in a manner which constitutes a means of arbitrary or unjustifiable discrimination, or a disguised restriction on international trade. For reasons of space constraint, and given that the relevance of Article XX(g) and the chapeau standards for NPR PPM-based measures have already been treated elsewhere,¹²⁹ the following considerations concentrate on Article XX(b).

In view of the international efforts to combat climate change and the fact that, e.g., the United Framework Convention on Climate Change has quasi-universal membership,¹³⁰ it seems safe to assume that WTO panels would regard climate protection as a legitimate goal also within the terms of Article XX(b). However, the means adopted – in casu the EU ecolabelling scheme – must also be ‘necessary’ to protect this aim. This necessity test essentially inquires into whether there is an alternative measure, which is less trade restrictive than the regulatory measure actually adopted, and whether the WTO member in question could reasonably be expected to apply this alternative.¹³¹ With the inception of the WTO, the WTO Appellate Body seems to have relaxed the necessity test when the values pursued are vital or particularly important.¹³² Given that the protection of life and health, which is pursued through a means such as an ecolabelling scheme that also aims at the protection of the international climate, undoubtedly is a vital interest, the necessity threshold should in principle be lower in such a case.

However, the necessity test, as applied in WTO jurisprudence, also seems to contain considerations of suitability and effectiveness.¹³³ It is worth noting, therefore, that official EU documents state that the EU scheme (at least in its 1992 and 2000 versions) does not appear effective,¹³⁴ or, put differently, that it is questionable whether it is suitable for reaching its goal. Moreover, it has been held, on a more general level, that ecolabels are typically unsuitable means for environmental policy-making.¹³⁵

In this respect, it must be stressed, however, that one can argue that the assessment of the suitability of a means adopted in pursuance of a legitimate goal should as a general rule employ a very low threshold for legal grounds and for interrelated reasons of decision-making theory,¹³⁶ and that this low degree of scrutiny is reflected also in WTO dispute settlement practice: thus, a measure is regarded, in standing Appellate Body jurisprudence, as suitable unless it ‘cannot in any possible situation have any positive effect on conservation goals’.¹³⁷ Additionally, one must take into account that the effectiveness of labels may increase over time,¹³⁸ particularly when they are combined with further types of measures such as label-related tax incentives and the promotion of the labelling scheme,¹³⁹ as has been quite clearly demonstrated by national experiences with the implementation of the EU’s labelling scheme for cars¹⁴⁰ as well as by the EU-wide experience with the considerably more successful EU energy-efficiency labelling scheme.¹⁴¹

Justification Under the TBT Agreement

When one takes the (contested) view that NPR PPM-based labelling schemes come within the scope of the TBT Agreement,¹⁴² then, as noted above, the EU labelling scheme, being a voluntary regime, would have to be scrutinized under the TBT Agreement’s Code of Good Practice (Annex 3 of the TBT Agreement). Like the TBT Agreement in Articles 2.1 and 2.2, the Code of Good Practice contains two self-standing disciplines that are primarily relevant for an EU-type labelling mechanism: pursuant to Article D, such a mechanism must not be discriminatory; pursuant to Article E, even non-discriminatory measures must not create unnecessary obstacles to international trade.

Turning first to Article D, and assuming that the EU ecolabelling scheme were to give rise to de facto discriminatory effects, the problem arises that the Code of Good Practice and the TBT Agreement more generally do not contain an explicit exception clause (modelled after Article XX of the GATT) that relates to Article D. An analogous problem exists under the TBT Agreement, as the architecture of Articles D and E of the Code of Good Practice has a structural counterpart in the similar architecture of Articles 2.1 and 2.2 of the TBT Agreement.¹⁴³ Some writers have doubted, therefore, whether it is possible at all to justify discriminatory measures under the TBT Agreement.¹⁴⁴ However, the preamble of the TBT Agreement underlines that justification that discriminatory measures must be possible also under this agreement, provided in particular that such measures are necessary and ‘not applied in a manner that would constitute a means of arbitrary or unjustifiable discrimination…or a disguised restriction on international trade’. Moreover, there are several arguments that support the view that justification of measures found to violate Article D of the Code of Good Practice may be possible under principles similar to the GATT. First, it could be argued that all WTO provisions being cumulative in principle, Article XX of the GATT should, as an overarching provision, also be regarded as being applicable in respect of the TBT Agreement.¹⁴⁵ Second, one could argue that the possibility of justification provided in Article E is also applicable to infringements of Article D. Third, one could submit that the notion of discrimination under Article D is different from that of Articles I and III of the GATT: whereas, under the GATT, a measure which is found to be ‘discriminatory’ under Articles I or III can still be justified under Article XX, one could submit that a measure should only be regarded as ‘discriminatory’ for purposes of Article D if it amounts to unjustifiable discrimination.¹⁴⁶

As mentioned before, Article E of the Code of Good Practice (just as Article 2.2 of the TBT Agreement) also constitutes a self-standing discipline that requires that even non-discriminatory measures do not represent unnecessary obstacles to international trade. Hence, even if labelling measures, adopted within the EU ecolabelling scheme for given product groups, were not to incur de facto discriminatory effects, they would have to be scrutinized under Article E as to their necessity. Moreover, the Code of Good Practice requires members to ensure that their standardizing bodies base their measures on appropriate and effective international standards,¹⁴⁷ and sets out transparency requirements similar to that relating to technical regulations under the TBT Agreement.¹⁴⁸ A complete assessment of the EU labelling scheme as to its consistency with these provisions would however require an examination of the large series of EU measures, in which individual labelling criteria for specific product groups have already been defined¹⁴⁹ or will be defined in future, and of the processes in which they are adopted. Nonetheless, it should be noted on a general level that labelling is commonly seen as a suitable and comparatively rather non-restrictive means for pursuing environmental goals¹⁵⁰; that the EU scheme mandates public and private bodies involved in criteria-setting to take account of relevant international standards¹⁵¹; and that it aims to provide openness and transparency in the criteria-shaping process and in conformity assessment procedures.¹⁵²

Mention should also be made of the view that the TBT Agreement’s necessity test may require members to forgo state-administered voluntary labelling in favour of privately sponsored schemes.¹⁵³ However, this contention cannot stand unqualified, given that legitimate concerns may designate state-run schemes as more effective in the sense of the necessity test: thus, verification of compliance with labelling criteria may turn out to be more reliable in concrete cases, which may in turn lead to broader consumer acceptance and increased effectiveness of the label. Additionally, government involvement may be necessary to establish uniform labelling mechanisms that help avoid the consumer disorientation¹⁵⁴ which risks being incurred by an overly wide array of competing privately sponsored labels.

Finally, as regards the justification of EU-type discriminatory and non-discriminatory NPR PPM-based labelling schemes under the TBT Agreement, regard must also be had to Agenda 21, whose legal import on the interpretation of WTO law has already been referred to,¹⁵⁵ and the Implementation Plan of the 2002 World Summit on Sustainable Development (WSSD). Although both emphasize the importance of ecolabelling as an instrument of environmental protection,¹⁵⁶ the WSSD Implementation Plan goes even further and explicitly endorses that countries should adopt NPR PPM-based labelling schemes that do not act as disguised trade barriers.¹⁵⁷

Summary of Conclusions

This contribution has examined the new 2010 ecolabelling programme of the EU as a model test case, under WTO law, for voluntary government-administered labelling schemes that rely on NPR PPM-based criteria. With respect to this highly contested type of labelling scheme, this contribution has arrived at the following main conclusions:

• The EU ecolabelling scheme and similar voluntary NPR PPM-based labelling schemes are not exempted from the scope of the TBT Agreement, and must fully comply with the TBT Agreement and its Code of Good Practice in particular.

• This type of scheme is not per se prohibited under the TBT Agreement solely owing to its reliance on NPR PPM-based criteria.

• The EU ecolabelling scheme can be attributed, in terms of GATT law, to public authorities. It is neither exempted from nor a priori prohibited under the GATT solely owing to its process-based approach.

• The term ‘like products’ should primarily be understood to mean products that are in a close competitive relationship, a determination that has to be made primarily from the perspective of consumers.

• This approach to the determination of likeness is in line with GATT/WTO dispute settlement practice. In the determination of likeness, this practice has traditionally relied on the criteria developed by the Working Party on Border Tax Adjustment: these criteria can be regarded as indicators of a close competitive relation. The Appellate Body has recently confirmed this focus on a competitive relationship and the inherently intertwined perspective of consumers.

• It follows from the fact that the likeness determination is inherently influenced by consumer perception that divergent NPR PPMs can theoretically affect the likeness judgment, rendering otherwise similar products unlike in terms of WTO law.

• Under labelling schemes, there is a clear risk that products which are not included in a product group that is awarded a label may have to be considered, under WTO law, to be ‘like’ the products that are entitled to such a label.

• Although this incurs a risk of (de facto) discrimination under labelling programmes such as the EU scheme, there is no per se violation of relevant non-discrimination disciplines in the GATT or the TBT Agreement.

• Put differently, non-discriminatory NPR PPM-based regulations in general and NPR PPM-based labelling schemes in particular should be regarded as being in conformity with relevant WTO non-discrimination disciplines if they do not incur disparate impacts between foreign and domestic like products.

• In its rulings which reflect relevant international environmental law, the Appellate Body has confirmed that NPR PPM-based regulations are not per se incapable of justification under Article XX of the GATT.

• NPR PPM-based labelling programmes of the EU-type are likewise capable of justification under Article XX of the GATT and under the TBT Agreement and its Code of Good Practice. WTO jurisprudence has recently applied a deferential standard of review when inquiring into the suitability and necessity of regulatory measures whenever vital interests such as human life and health are at stake. This approach is clearly relevant also when assessing climate-related measures such as the EU ecolabelling scheme.