I. INTRODUCTION
The principle of non-discrimination has a long-standing history in international trade relations and it has become a central pillar of modern international economic law. The non-discrimination principle provides that contracting parties to an international economic treaty shall not treat domestic market actors more favourably than foreign market actors (national treatment, NT) or differentiate between foreign market actors from different origins (most-favoured-nation treatment, MFN). Non-discrimination obligations are found in all fields of international economic law, notably trade in goods and services, investment protection and the protection of intellectual property rights. They apply to all types of governmental trade obstacles, such as border measures (eg tariffs and quantitative restrictions) and internal regulations (eg taxes and product standards). In addition, it is well established that non-discrimination obligations not only apply to measures which differentiate directly—or de jure—on the basis of origin, but they also prohibit indirect—or de facto—discriminatory measures.
In spite of these commonalities between non-discrimination obligations of different economic treaties, it would be wrong to assume that non-discrimination in international economic law has a firmly defined meaning.Footnote 1 Adjudicating bodies apply different interpretations and standards with regard to the elements of ‘less favourable treatment’, ‘likeness’ and ‘regulatory purpose’, which leads to a variety of different standards of non-discrimination in international economic law. The different standards may in some instances be explained by different intentions, objectives and expectations of the contracting parties or by different structures of the specific non-discrimination clause or of the entire treaty. However, the economic rationale underlying non-discrimination claims are very similar in trade and investment. In both cases, a foreign market actor seeks access to a domestic market under equal competitive parameters compared to domestic market actors. Hence, there are often times when there is no apparent reason, other than pure arbitrariness, for applying different standards in the respective fields of international economic law.
The rules of treaty interpretation set forth in article in 31 of the Vienna Convention on the Law of Treaties (‘VCLT’),Footnote 2 and reflected in customary international law, leave considerable discretion to adjudicators interpreting an obligation of public international law. The rules are not apt to ensure a consistent application of the non-discrimination provisions. Even though it appears that World Trade Organisation (‘WTO’) law has—to a certain degree—assumed a leading role for the general interpretation of international economic law, adjudicators applying a specific treaty have no obligation to take into consideration the jurisprudential developments and precedents from other fields of international economic law or even from prior arbitral tribunals applying the exact same provision of the same treaty. Considering that for historical and political reasons, international economic law is split up in innumerable mostly self-contained bilateral, regional and multilateral treaties, it is also not surprising that most scholarly contributions tend to focus on the principle of non-discrimination as it applies in a specific field of international economic law.Footnote 3
Nonetheless, non-discrimination obligations are constantly applied by international arbitral tribunals and, from a pragmatic perspective, the system appears to function. However, the current situation entails a number of shortcomings. Most importantly, contracting parties have no possibility to accurately anticipate the consequences when entering into an international economic treaty containing a non-discrimination clause. By the same token, private individuals—to the extent the treaty empowers them to assert their rights under the agreement—are virtually left in the dark when assessing their rights and risks prior to an investment decision. The scope, substance and standard of a non-discrimination obligation—and nota bene many other obligations—depends on the interpretation of the individual adjudicators who happen to be appointed to rule on the specific dispute. Moreover, even once the arbitral tribunal makes its ruling by applying a specific treaty, the concerned parties have no guarantee that a subsequent arbitral tribunal will follow the same interpretation. Consequently, the parties to a dispute—whether in state-to-state or investor-state arbitration—have little guidance on how to make their claim, how to present their legal arguments and most importantly, what evidence will be relevant. Granted, any legal proceeding—whether domestic or international—involves a considerable degree of legal uncertainty, but the risks and uncertainties are exceedingly high in disputes pertaining to international economic law in general and to the non-discrimination obligation in particular.
Against the background of these fragmented standards of non-discrimination, this article attempts to establish a typology of the theoretically possible standards on the basis of previous practice in the law of the WTO,Footnote 4 the Treaty on the Functioning of the European Union (‘TFEU’),Footnote 5 the North American Free Trade Agreement (‘NAFTA’)Footnote 6 and bilateral investment treaties (‘BITs’).Footnote 7 Part II compares the different standards which have been developed by the WTO adjudicating bodies, the European Court of Justice (‘ECJ’) and arbitral tribunals with regard to the non-discrimination elements, namely the ‘comparator clause’, ‘less favourable treatment’ and ‘regulatory purpose’. This overview does not pretend to provide an exhaustive and detailed analysis of the specific most-favoured-nation and national treatment provisions in the different agreements. Rather, Part II is designed to present a general analysis of the different possible interpretations in order to illustrate to negotiators of international economic treaties the ambiguities arising from the text of today's non-discrimination provisions and to provide an overview of the possible standards of non-discrimination obligations to adjudicators and disputing parties. On this basis, Part III develops a structure containing the main formal and substantive elements which should be taken into consideration for the legal analysis of a non-discrimination claim.
The purpose of this article is to propose a methodology for the drafting of future non-discrimination obligations which is not based on ambiguous legal elements or conditions subject to interpretation, but on clearly defined factors which need to be weighed and balanced in order to determine whether overall a measure amounts to unlawful discriminatory protection of domestic market actors.Footnote 8
II. CURRENT STANDARDS OF NON-DISCRIMINATION
The principle of non-discrimination consists of two main elements, both of which are comparative in nature. First, the comparator clause calls for a comparison between the market actors subject to differential treatment.Footnote 9 The second element requires a comparison between the treatments accorded to the market actors at issue in order to assess whether one is treated less favourably than the other. The comparator clause and the element of ‘less favourable treatment’ constitute two cumulative legal conditions.Footnote 10 Depending on the structure of a specific non-discrimination provision, additional elements may be taken into consideration, such as ‘(non-)protectionist effect’ or ‘(non-)protectionist purpose’. To date, the comparator clause—in particular the GATT ‘like products’ concept—has received by far the most attention in dispute settlement and legal scholarship. More recently, the focus has also shifted to the element of ‘less favourable treatment’ and its ambiguities.Footnote 11 However, scholarly research and case law does not yet sufficiently take into account the relationship and interdependency between the two elements.
Each element may be subject to different interpretations and standards, which considerably affects the reach of the non-discrimination obligation.Footnote 12 The spectrum varies from very permissive forms of non-discrimination obligations which only outlaw the most apparent and blatant discriminatory measures, to very liberal and integrative forms which considerably restrict the contracting parties’ regulatory autonomy to pursue domestic policy objectives. The present part illustrates the different standards by referring to examples from GATT 1947, WTO law (GATT 1994 and GATS), EU law, NAFTA and BITs.
A. Different Standards and Terminology for ‘Comparator Clauses’
The fragmentation of comparator clauses in different international economic treaties is readily apparent from the differences in the terminology employed. Non-discrimination provisions in WTO law generally use the concept of ‘likeness’, such as ‘like products’ in articles I and III GATT or ‘like service and service suppliers’ in articles II and XVII GATS. In addition, one GATT non-discrimination provision applies the concept of ‘directly competitive or substitutable products’ instead of ‘likeness’.Footnote 13 The TFEU refers to ‘similar products’ and ‘other products’ in its national treatment obligation concerning internal taxes (article 110 TFEU). Finally, non-discrimination obligations in NAFTA and certain BITs apply the concept of ‘like circumstances’,Footnote 14 while other BITs use the concept of ‘same circumstances’,Footnote 15 ‘like situations’,Footnote 16 ‘comparable situations’Footnote 17 or ‘similar situations’.Footnote 18
In spite of these different terminologies, all comparator clauses share the identical fundamental problem of identifying the relevant tertium comparationis, ie the quality or element which two ‘situations’ or ‘objects’ must have in common in order to conclude that they are ‘alike’ for the purpose of the comparison.Footnote 19 The practice of international economic law determined different tertia comparationis which in turn leads to different standards of non-discrimination.
1. Objective Standard
Early GATT 1947 jurisprudence pertaining to articles I (MFN) and III (NT) interpreted the concept of ‘like products’ on the basis of purely formal and objective criteria, mostly ignoring or even denying the relevance of competition.Footnote 20 For instance, GATT Panels ruled that ‘likeness’ does not exist between three types of sardines (pilchard, herring and sprat) or between dimension lumber produced from different tree species (SPF and hemlock-fir lumber) for purposes of article III:2 GATT. Another GATT Panel explicitly refused to consider the competitive relationship between ammonium sulphate fertilizer and nitrate fertilizer and found the two products to be ‘unlike’. Finally, the GATT Panel in EEC—Animal Feed Proteins held that different products used for the purpose of adding protein to animal feeds are ‘unlike’ under articles I and III:4 GATT. All these reports relied heavily on different tariff classifications as well as physical differences between the products as criteria of the ‘likeness’ analysis.Footnote 21 In 1970, a Working Party Report developed a test which later came to be known as the Border Tax Adjustments framework, identifying also ‘end-uses’ and ‘consumer tastes and habits’ as relevant criteria.Footnote 22 While these criteria may have implicitly introduced economic elements into the analysis, most GATT Panels refrained from explicitly recognizing that ‘likeness’ incorporates the economic theory of competitive relationships.
Similar to GATT 1947 practice, early ECJ jurisprudence pertaining to the national treatment obligation for internal taxes of article 110(1) TFEU (ex article 90 TEC, ex-ex article 95 TEC) also relied on purely formal criteria—such as fiscal, statistical or custom classification—for the assessment of the ‘similar products’ concept.Footnote 23
While all of the above examples from GATT 1947 and EU jurisprudence illustrate situations where formal criteria were used to find ‘unlikeness’ between largely competing products, formal criteria may be used to find ‘likeness’ between non-competing products. In the investor-state dispute Occidental,Footnote 24 for instance, the US oil exporter Occidental claimed that the denial of VAT reimbursements to domestic and foreign-invested oil exporters, while granting the reimbursement to domestic and foreign-invested exporters of other goods, constituted a breach of the non-discrimination obligation. The claim was based mainly on the argument that ‘in like situations’ refers not to companies in the same business sector, but to all companies engaged in exports even across economic sectors. The tribunal distinguished the BIT ‘like situations’ terminology from GATT ‘like products’ and essentially defined the ‘act of exporting’ as the relevant tertium comparationis.Footnote 25 In consequence, the foreign invested oil exporter Occidental was considered ‘in like situations’ with domestic exporters of flowers, mining and seafood products as well as lumber and bananas.Footnote 26
In sum, under the objective standard the tertium comparationis may consist of factors such as physical characteristics, tariff classification, end-uses or even the act of exportation. Depending on which criteria are applied, the scope of non-discrimination obligations may be construed very narrowly (eg in case of physical characteristics as criterion) or extremely broadly (eg act of exporting as criterion).
2. Economic Standard
Under the economic standard, the tertium comparationis is defined by economic parameters indicating the extent to which the market actors are in a competitive relationship. This standard was first applied by GATT 1947 panels for the ‘directly competitive or substitutable products’ element in article III:2 GATT and by the ECJ for the concept of ‘other products’ in article 110(2) TFEU.Footnote 27
Under WTO jurisprudence, the distinction between ‘like products’ and ‘directly competitive or substitutable products’ in article III GATT is gradually disappearing.Footnote 28 The WTO adjudicating bodies are more and more prepared to extend the economic standard to the concept of ‘like products’. The Appellate Body Report on EC—Asbestos nicely illustrates this change in approach with regard to the assessment of discriminatory regulations (article III:4 GATT).Footnote 29 However, even though the WTO adjudicating bodies recognized the relevance of competitive relationships for the analysis of ‘likeness’ under this so called ‘market place approach’, they continue, in principle, to apply the formal criteria from the Border Tax Adjustments framework.Footnote 30
The ECJ jurisprudence pertaining to ‘similar products’ under article 110(1) TFEU, which was developed in parallel to the WTO jurisprudence on article III:2 GATT, moved from a formally objective interpretation to a test also taking into account economic considerations.Footnote 31 Consequently, the difference between the concepts of ‘similar products’ in article 110(1) TFEU and ‘other products’ in article 110(2) TFEU is gradually disappearing.
The same economic interpretation of ‘likeness’ will have to prevail for the WTO national treatment obligation pertaining to trade in services, considering that article XVII:3 GATS explicitly states that it is designed to protect competitive opportunities.Footnote 32 In comparison, the WTO has not yet had an opportunity explicitly to confirm an economic standard of ‘like products’ for purposes of MFN (article I GATT),Footnote 33 but numerous scholars rightfully demand such an approach at least with regard to internal taxes and regulations.Footnote 34 Similarly, to date no jurisprudential guidance exists with regard to the standard of ‘likeness’ in article II GATS on MFN-treatment.Footnote 35
Finally, most arbitral tribunals applying the NAFTA rules on investment protection largely endorse an economic interpretation of the ‘like circumstances’ concept, even though the jurisprudence is not entirely consistent. The main criterion generally is whether investors or investments are in the ‘same sector’, including both economic and business sectors. While some tribunals omit explicitly to identify the competitive relationship as the decisive factor to delimit a ‘sector’, others base their analysis more specifically on competition.Footnote 36
3. Subjective Standard
The subjective standard of ‘likeness’ has been developed by different adjudicating bodies of international economic law in order to strike a balance between international obligations designed to liberalize trade and investment on the one hand, and domestic non-economic policy objectives such as environmental and consumer protection on the other hand. The doctrinal reasoning of the subjective standard is to argue that the tertium comparationis is defined by the regulatory purpose of the measure under scrutiny; for instance, if the measure is designed to protect the environment, then the products are compared on the basis of their environmental impact.
GATT 1947 jurisprudence developed a subjective standard with the so called ‘aim and effects’ test as part of the ‘like products’ analysis.Footnote 37 Following this approach, a GATT Panel ruled that low and high alcohol content beers are not ‘alike’ for the purpose of article III:4 GATT because the measures restricting points of sale, distribution and labelling were aimed to encourage the consumption of low alcohol beer. Conversely, wines made from different grapes were found to be ‘like products’ mainly because the respondent was unable to provide any valid public policy purpose in support of its differential tax treatment.Footnote 38 While subsequently WTO panels and the Appellate Body strongly rejected the ‘aim and effects’ test for purposes of both GATT and GATS,Footnote 39 the very recent Panel report US—Clove Cigarettes has again taken into account the regulatory context under the ‘likeness’ analysis for purposes of the non-discrimination provision in article 2.1 TBT Agreement.Footnote 40
Similar to the ‘aim and effects’ test, most arbitral tribunals ruling on NAFTA non-discrimination provisions in the area of trade in services and investment protection interpret the concept of ‘like circumstances’ as containing a subjective element. Following this rationale, the question is not whether the foreign and domestic suppliers or investors are in ‘like circumstances’, but whether the differential treatment occurs in ‘like circumstances’. In other words, the policy objective pursued by the measure under scrutiny may be taken into consideration to define the circumstances in which the comparison of the foreign and domestic comparators takes place.Footnote 41
The same result is likely to prevail in certain BIT non-discrimination clauses. For instance, the former Norwegian draft model-BIT nicely illustrates the subjective standard by means of a footnote to the ‘like circumstances’ concept:
1. Each Party shall accord to investors of the other Party and to their investments, treatment no less favourable than the treatment it accords in like circumstances[fn] to its own investors and their investments, in relation to the establishment, acquisition, expansion, management, conduct, operation and disposal of investments.
[fn] The Parties agree/are of the understanding that a measure applied by a government in pursuance of legitimate policy objectives of public interest such as the protection of public health, safety and the environment, although having a different effect on an investment or investor of another Party, is not inconsistent with national treatment and most favoured nation treatment when justified by showing that it bears a reasonable relationship to rational policies not motivated by preference of domestic over foreign owned investment.Footnote 42
The main challenges of this approach are to determine the legitimate policy objectives and the appropriate standard for the reasonable relationship between the measure under scrutiny and the pursued objective (see below, II.C.3).
4. Combination of Standards
The objective, economic and subjective standards of ‘likeness’ may be applied individually or in combination. WTO adjudicating bodies combine the economic and objective standard for purposes of certain GATT non-discrimination provisions. For instance, ‘like products’ in terms of article III:1 GATT, first sentence, is interpreted as requiring both physical similarity and a competitive relationship. In comparison, the concept of ‘like circumstances’ as contained in NAFTA rules on non-discrimination is usually interpreted as providing both an economic and a subjective standard allowing differentiation between competing investors or service suppliers in order to pursue legitimate domestic policy objectives.
5. Absence of a Comparator Clause
Finally, the comparator clauses are not only highly fragmented in terms of their terminology and application; some international economic treaties even contain non-discrimination obligations that entirely lack a comparator clause.Footnote 43 The absence of a comparator clause may be subject to two different interpretations. The first and preferred approach consists of the argument that a comparative element is inherent to the logic and structure of the non-discrimination principle in international economic law.Footnote 44 Consequently, the claimant would still have to establish the existence of a competitive relationship between the allegedly discriminated foreign market participant and a domestic market participant who is receiving more favourable treatment. Conversely, under the second theory the absence of a comparator clause entails that competitive relationships or any other form of ‘likeness’ between the domestic and foreign comparators are irrelevant. Consequently, equal treatment would have to be accorded to foreign and domestic market actors across economic sectors.Footnote 45
B. Different Standards of ‘Less Favourable Treatment’
The term or standard of ‘less favourable treatment’ is usually not defined in non-discrimination provisions of international economic treaties. As a notable exception, article XVII:3 GATS states that ‘different treatment shall be considered to be less favourable if it modifies the conditions of competition’. GATS incorporates the interpretations from prior GATT 1947 panel reports which developed the principle of conditions of competition under the analogous provision of article III GATT.Footnote 46 Considering that non-discrimination obligations aim to ensure equal conditions of competition for foreign and domestic market actors, it is logically consistent that differential treatment is only relevant to international economic law to the extent it modifies the conditions of competition to the detriment of certain foreign market actors. In principle, this rationale not only applies to international trade, but also to investment protection law.Footnote 47 However, the standards or thresholds with regard to the element of ‘less favourable treatment’ differ considerably.
1. Disproportionate Disadvantage Test
The disproportionate disadvantage test requires assessing the negative (and potentially neutral or positive) economic effect of a measure on the group—as defined by the comparator clause—of domestic and foreign market actors.Footnote 48 The non-discrimination obligation is only breached if the group of foreign market actors is disproportionately disadvantaged as compared to the domestic one.
Even though WTO jurisprudence is not entirely consistent on this issue, it appears that the Appellate Body explicitly endorsed the disproportionate disadvantage test in EC—Asbestos. In this case, the Panel ruled that the French sales ban for products containing asbestos fibres violated GATT national treatment, arguing that products containing asbestos fibres are ‘like’ products containing substitute fibres and that the measure resulted in less favourable treatment of asbestos products not produced in France. The Appellate Body reversed the Panel's ruling primarily on grounds of ‘likeness’, but it also reversed the Panel's approach in regard to ‘less favourable treatment’ in an obiter dictum. The Appellate Body held that ‘a complaining Member must […] establish that the measure accords to the group of “like” imported products less favourable treatment than it accords to the group of domestic products’.Footnote 49 This approach is supported by most commentators of WTO law.Footnote 50
In comparison, the ECJ also adopts the disproportionate disadvantage test in order to demonstrate whether a tax is of protective nature for purposes of article 110(2) TFEU prohibiting tax discrimination:
The protective nature of the tax system … is clear. A characteristic of that system is in fact that an essential part of domestic production … come within the most favourable tax category whereas at least two types of product, almost all of which are imported from other Member States, are subject to higher taxation … . The fact that another domestic product … is similarly placed at a disadvantage does not rule out the protective nature of the system … .Footnote 51
Considering the fact that a measure may have negative economic effects on certain competitors and no or even positive effects for other competitors, the disproportionate disadvantage test requires establishing a ratio threshold for ‘disproportionality’. Assume, for instance, a theoretical model situation where hundred domestic products stand vis-à-vis hundred imported ‘like’ products. Presumably no ‘less favourable treatment’ occurs if domestic and foreign products are equally affected (eg ten domestic vs ten foreign or seventy domestic vs seventy foreign). However, there remains a large range between, for instance, a negative effect on ten foreign vs five domestic or ninety-five foreign vs ten domestic products. It remains in the discretion of the adjudicators to define an appropriate ratio for each specific case.
2. Obligation to Grant the Best Treatment Accorded to any Domestic Market Participant
The non-discrimination principle may also be interpreted as an obligation to grant the best treatment accorded to any domestic market actor.Footnote 52 Following this approach the non-discrimination obligation is already breached if one individual foreign market actor receives treatment that is less favourable in comparison to any individual domestic market participant (NT) or to any foreign market participant from different origin (MFN). For instance, less favourable treatment occurs if a measure negatively affects only one out of hundred foreign market actors, even if ninety-nine out of hundred domestic market actors are also negatively affected. Consequently, the non-discrimination principle becomes an obligation to treat all foreign market participants equivalent to the best treatment accorded to any ‘comparable’ domestic or other foreign market participant. Under this approach, non-discrimination has a strong liberalizing effect and far reaching consequences for the regulatory autonomy of the contracting parties, in particular if additionally the comparator clause is interpreted widely.
This very liberal and intrusive interpretation is mostly adopted in the area of investment protection.Footnote 53 In particular, the jurisprudence pertaining to NAFTA chapter 11 shows a clear tendency towards this ‘best treatment’ approach.Footnote 54 For instance, the tribunal in Pope & Talbot ruled ‘that “no less favorable” means equivalent to, not better or worse than, the best treatment accorded to the comparator’.Footnote 55
The main argument made in support of this far reaching standard is that investment treaties are designed to protect the value of a specific investment, whereas international trade law protects a more abstract value of equal conditions of competition, not the actual value of the exported goods and services. This argument has some merit, insofar as the investment in a foreign market is a more substantial and binding commitment to participate in the foreign market than merely exporting goods and services. A new regulation—for instance an environmental standard—may have severe consequences for the foreign investor who, in case he is unable to comply with the standard, may have to disinvest and suffer actual damages. Conversely, a foreign producer unable to meet the new standard may simply cease to export without incurring actual damages, but only loss of potential gains.Footnote 56
Yet, this circumstance is taken care of in that investment treaties accord to the foreign investor an individual right to claim damages, whereas trade agreements are only enforceable by the governments of the contracting parties which may only claim the abolition of the measure, but not damages. It is not entirely apparent, however, why the arguably different objects of protection in investment and trade law (ie value of investment vs conditions of competition) should also explain different substantive standards of non-discrimination obligations. In fact, most investment treaties go beyond the protection of foreign direct investment, in that they also apply to more mobile forms of investments, such as minority equity investment or debt holdings.Footnote 57 These types of investments are not subject to the same sunk costs as foreign direct investment and may even be retracted from the foreign market. Hence, if different objects protected by the respective economic treaties were apt to justify different standards of non-discrimination obligations, more than one standard would already have to apply in the context of investment law.
Moreover, distinguishing between a ‘best treatment’ standard for investment protection law and a ‘disproportionate impact’ standard for trade law is not appropriate in view of the fact that trade agreements may also apply to forms of investments. For instance, measures affecting trade in services under GATS mode 3 (commercial presence) would presumably have to be analysed under a disproportionate impact test in line with the Appellate Body's interpretation of ‘less favourable treatment’, even though the concept of commercial presence constitutes a form of investment. Conversely, some WTO panels have applied a standard along the lines of the ‘best treatment approach’ for purposes of the GATT national treatment provision in the trade context.Footnote 58
Similar examples exist in other areas of international economic law. For instance, it is to be expected that non-discrimination provisions in NAFTA chapter 12 on trade in services would follow the same standard as those in chapter 11 on investment protection in view of the similar terminology and structure of articles 1102ff and articles 1202ff, even though in substance chapter 12 is more closely related to the trade rules of NAFTA chapter 3 than to the rules on investment protection. Even within NAFTA chapter 11, the interpretation lacks consistency in that some tribunals seemed to favour the disproportionate disadvantage test which is generally applied in the trade context.Footnote 59
The more practical and pragmatic explanation for the different standards of non-discrimination in trade and investment protection may be found in the form of the challenged measure, rather than in the object protected by the respective treaty. In fact, the measures challenged in investor-state cases often consist of (individual and concrete) decisions by an authority applying to one foreign investor, whereas the measures challenged in trade disputes typically concern (general and abstract) regulations applying to all competitors in a relevant market. The ‘best treatment’ approach is arguably more appropriate with respect to ‘decisions’, as the concerned investor would only have to show that it is the only actor in a relevant market suffering from a competitive disadvantage due to the decision in question. In contrast, in order for a regulation to be discriminatory, the complainant must show that the group of foreign products suffers a higher competitive burden from the regulation than the group of ‘like’ domestic products.
These considerations illustrate that the circumstances may very well justify a ‘best treatment’ or similar standard in a trade case, whereas the ‘best treatment’ approach may just as well lead to absurd results in investment protection law. In view of the inconsistency in the interpretation of this core element of non-discrimination and the resulting lack of legal security, it would be preferable for international economic treaties to either specifically spell out the applicable standard in the agreement or to explicitly empower the arbitral tribunal with the competence to determine the applicable standard on a case-by-case basis.
3. Subjective Standard of ‘Less favourable Treatment’
The subjective standard of ‘less favourable treatment’ takes into account the regulatory purpose in order to determine the true basis of the differential treatment. In the case of de jure discrimination, the measure differentiates directly on the basis of origin. However, cases of de facto discrimination differentiate directly on the basis of a permitted criterion. Under the different approaches to ‘less favourable treatment’ discussed above, the link from the permitted to the prohibited criterion is presumed if predominantly foreign products suffer a competitive disadvantage under the measure (ratio threshold). In contrast, the subjective standard requires determining whether the measure truly pursues an objective related to the permitted criterion, or whether the true intent of the measure is to discriminate indirectly on the basis of origin.
This subjective theory of ‘less favourable treatment’ has not yet been explicitly recognized by the WTO adjudicating bodies or by arbitral tribunals. Some commentators understood the Appellate Body's obiter dictum in EC—Asbestos as a return of the ‘aim and effects’ test under the element of ‘less favourable treatment’;Footnote 60 however, in this case the Appellate Body only addressed the issue of ratio, but not of purpose.Footnote 61
More pertinently, the Panel in EC—Approval and Marketing of Biotech Products seemed to consider a subjective standard of ‘less favourable treatment’:
Argentina is not alleging that the treatment of products has differed depending on their origin. In these circumstances, it is not self-evident that the alleged less favourable treatment of imported biotech products is explained by the foreign origin of these products rather than, for instance, a perceived difference between biotech products and non-biotech products in terms of their safety, etc. In our view, Argentina has not adduced argument and evidence sufficient to raise a presumption that the alleged less favourable treatment is explained by the foreign origin of the relevant biotech products.Footnote 62
Most recently, the US argued in the WTO dispute US—Tuna II, that its ‘dolphin safe’ labelling provisions do not draw a distinction based on the origin of the tuna, but based on whether the tuna products contain tuna caught in a manner harmful to dolphins. The US further argues that there is a relationship between the purpose of the measure—which is to reduce the harm to dolphins and to properly inform consumers—and the conditions set forth by the measure by which tuna products can be labelled ‘dolphin safe’.Footnote 63 In essence, the US argument goes to say that the non-discriminatory intent of the measure shows that the differential treatment is not based on origin, but based on the US objective of protecting dolphins. The Panel seems to have followed the US argument at least to a certain degree, finding that the ‘dolphin safe’ labelling provision does not result in less favourable treatment of Mexican tuna under article 2.1 TBT Agreement:
We first note that, the fact that the measures distinguish between fish based on its capture method rather than its origin, this distinction is not inherently tied to the “national” origin of the fish. The fishing method at issue, setting on dolphins, is accessible to any fleet operating in an area where such method can be practised. Therefore, denying the label to tuna caught by “setting on dolphins” does not, in itself, imply that “less favourable treatment” is afforded to Mexican tuna products. Indeed, any fleet operating anywhere in the world must comply with the requirement.Footnote 64
The Panel then went on to conclude that even if tuna of Mexican origin might be more likely not be eligible for the ‘dolphin safe’ label, this would not necessarily mean that the tuna products processed and canned in Mexico would be less likely to qualify for the label. In the Panel's words, this ‘is because Mexican processors could choose to make their products from tuna of other origins meeting the requirements of the label’.Footnote 65 In sum, it appears that the Panel was at the very least inspired by the purpose of the measure when concluding that, even de facto, the labelling provision does not differentiate based on origin.
C. Different Standards and Relevance of ‘Regulatory Purpose’
The main objective of non-discrimination obligations in international economic law is to outlaw measures which are specifically designed to protect the domestic market from foreign competition. However, even measures which pursue a legitimate policy objective—such as measures setting standards related to health, environment, labour or human rights—may have a protectionist effect. In such cases, most international economic treaties provide that the policy objective of the measure is taken into account for the legal analysis. However, there are two different systemic approaches as to whether the purpose of a regulatory measure is analysed as part of the non-discrimination obligation itself or as a justification under the general exceptions clause.Footnote 66
1. Regulatory Purpose as Part of the Non-Discrimination Standard
As discussed above, the regulatory purpose may be considered as part of the comparator clause or as part of the ‘less favourable treatment’ element, to the extent that a subjective standard is applied. This solution is very rarely explicitly adopted by international economic treaties, but adjudicating bodies occasionally choose one of these two approaches by applying and interpreting a specific non-discrimination obligation. A look at the relevant cases pertaining to non-discrimination provisions shows that this solution is mostly adopted when the treaty lacks a general exceptions clause. This is the case, for instance, for many BITs, NAFTA chapter 11 and the TBT Agreement.
Alternatively, the regulatory purpose could be considered as a distinct and separate legal element within the non-discrimination obligation. However, international economic treaties do not provide a textual basis for such an approach. The only provision allowing for some flexibility in this regard is article III:1 GATT, which states that measures ‘should not be applied to imported or domestic products so as to afford protection to domestic production’.Footnote 67 Even though the Appellate Body emphasized that the test for the wording ‘so as to afford protection’ is about protective application, not about protective intent,Footnote 68 indications of protectionist intent regularly flow into the legal analysis.Footnote 69
From a practical and pragmatic point of view, it seems irrelevant whether the regulatory purpose is considered under the comparator clause, the ‘less favourable treatment’ element or as a distinct and separate element. However, from a doctrinal and systemic angle it would be welcomed if the jurisprudence, or preferably the treaties themselves, would clarify whether and under which title the purpose of an allegedly discriminatory measure may be analysed. Such transparency would enhance legal certainty and facilitate the parties to a dispute in building their legal arguments. For clarity and structural reasons, the regulatory purpose should ideally be considered as its own legal element. However, due to the lack of a textual basis, adjudicating bodies mostly rely on the comparator clause or to a lesser extent on the element of ‘less favourable treatment’ for taking into account the purpose of a regulation.
2. Regulatory Purpose under the General Exceptions Clause
Under the second theory, the regulatory purpose is taken into account only once it is established that the measure under scrutiny is in breach of the non-discrimination obligation. This approach requires incorporating an explicit justification or exception clause in the structure of the respective international economic treaty. Importantly, general exception clauses not only legitimize the violation of a non-discrimination obligation, but also violations of other substantive obligations set forth in the respective treaty, such as the prohibition of quantitative restrictions.
International trade agreements regularly provide for a general exceptions clause. Article XX GATT, for instance, allows WTO Members to deviate from their obligations under GATT in order to pursue public interests such as the protection of public morals, of exhaustible natural resources or of human, animal and plant life or health.Footnote 70 The lists of public interests serving as grounds for justification in article XX GATT, article XIV GATS and most other trade agreements are exhaustive in nature. Consequently, general exceptions clauses only address interests which were recognized as important at the time the treaty was concluded, but due to their static nature, they fall short of addressing new concerns, such as for instance environmental or consumer protection.
One way of addressing new policy concerns in the framework of trade treaties would be to broadly interpret the general exceptions clauses. However, adjudicating bodies ruling on bi- and multilateral trade agreements generally lack the legitimacy to engage in legislative interpretation or to extensively depart from the treaty text of general exceptions clauses.Footnote 71 Only the ECJ transformed the general exceptions clause of article 36 TFEU (ex article 30 TEC) from an exhaustive to an illustrative list of public interests in order to counter-balance its very liberal application of the Cassis de Dijon concept,Footnote 72 generally referred to as principle of origin.
Another way of circumventing the limited scope of general exceptions clauses is to incorporate the regulatory purpose as an element into the standard of non-discrimination by adopting a subjective approach to ‘likeness’ or to ‘less favourable treatment’ as described above.Footnote 73 Adjudicating bodies ruling on trade disputes frequently use this approach, be it explicitly or implicitly, in order to resolve a conflict between free international trade and a domestic policy concern of the State that is bound by a non-discrimination obligation.Footnote 74
In contrast to trade agreements, most investment protection treaties do not contain any general exceptions clause. NAFTA, for instance, provides for a general exceptions clause with respect to trade in goods and trade in services (article 2101 NAFTA), but not with respect to chapter 11 on the protection of investments. This may be the reason why many investment tribunals ruling on a discrimination claim tend to consider the purpose of the measure as part of the non-discrimination obligation by adopting one of the approaches analyzed under Parts II.A.3 and II.B.3 above.
3. Legal Challenges Related to the Analysis of a Measure's Regulatory Purpose
While a proper consideration of a discriminatory measure's policy objective may provide the ‘most just’ result, adjudicating bodies are confronted with a number of very sensitive problems when analyzing the regulatory purpose of a State measure.
First, consideration of the regulatory purpose raises issues with respect to burden of proof, means of proof and standard of review. An unreasonably high bar would be raised by placing the burden on the complainant to prove a protectionist purpose of the respondent's measure. Preferably, it should be up to the respondent to demonstrate that its measure pursues a non-protectionist and legitimate objective. Either way, direct evidence of protectionist intent will rarely exist, considering that numerous governmental actors and interest groups are usually involved in the decision-making process. The disputing parties thus have to rely mostly on circumstantial evidence related to the design, structure, application and effect of a measure. In this context, the question arises of how much deference the adjudicating bodies should give to the respondent's evidence and assertions concerning the purpose of its own measure. This is a very sensitive issue related to the question of the appropriate standard of review.Footnote 75
Second, the adjudicators must determine which policy objectives are considered as sufficiently important and thus legitimate to justify a measure which—indirectly or directly—discriminates between foreign and domestic competitors. Some international economic treaties directly define the legitimate policy objectives. General exceptions clauses, for instance, set forth a list of legitimate objectives agreed upon by the contracting parties during the negotiations. As stated above, such lists are usually regarded as exhaustive; only the ECJ transformed the general exceptions clause of article 36 TFEU (ex article 30 TEC) from an exhaustive to an illustrative list of public interests serving as grounds for justification.Footnote 76 Conversely, adjudicating bodies may have more flexibility in defining new legitimate policy objectives if no such treaty mandated exhaustive lists exist, by simply taking the regulatory purpose into account under the comparator clause or under the element of ‘less favourable treatment’.
Third, a legitimate objective itself is not sufficient to justify the breach of a non-discrimination obligation; there must also be a certain nexus between the measure under scrutiny and the legitimate objective pursued. Most treaties which provide a justification or general exceptions clause explicitly state the required nexus. For instance, articles XX GATT and XIV GATS differentiate between measures which are ‘necessary’ to achieve the pursued policy objective or merely ‘related to’ the policy objective. Again, the adjudicating bodies are more flexible in defining the relevant nexus applying a subjective interpretation of the comparator clause or the element of ‘less favourable treatment’ if no treaty mandated standard exists.
Considering that any decision on any of these issues has far reaching consequences for the contracting parties’ sovereignty and regulatory autonomy, it would be highly preferable that the international treaty either defines the applicable rules with respect to burden of proof, standard of review, legitimacy of objectives and standards of nexus or that the treaty specifically empowers and legitimizes the adjudicating body to develop the necessary rules on a case by case basis.
D. Overlap Between Non-Discrimination and Non-Restriction
Depending on how each element of the non-discrimination obligation is construed, the result may be that non-discrimination overlaps with the more integrative principle of non-restriction (Beschränkungsverbot). The legal concept of non-restriction goes much further in trade liberalization than the principle of non-discrimination. It is fundamental, for instance, to the freedom to provide services under EU law. According to the ECJ, article 56 TFEU (ex article 49, ex-ex article 59 TEC)Footnote 77 ‘requires not only the elimination of all discrimination against a person providing services on the ground of his nationality but also the abolition of any restriction, even if it applies without distinction to national providers of services and to those of other Member States, when it is liable to prohibit or otherwise impede the activities of a provider of services established in another Member State where he lawfully provides similar services.’Footnote 78
Some commentators suggest that the national treatment obligation under WTO law should be interpreted as prohibiting any measure which is more burdensome than necessary for foreign goods, services or suppliers.Footnote 79 Under this approach, the elements of ‘less favourable treatment’ and ‘likeness’ are in effect replaced by a test of necessity and proportionality.Footnote 80
Another way of assimilating non-discrimination to non-restriction would be to combine the ‘best of the best treatment’ approach for ‘less favourable treatment’ with a broad economic interpretation of the ‘comparator clause’. As it would be almost always possible to determine at least one distant competitor receiving more favourable treatment, the principle of non-discrimination would in essence be transformed into an obligation of non-restriction.
Finally, the overlap between non-discrimination and non-restriction becomes even more apparent where broad objective criteria such as ‘the act of exporting’ are applied under the ‘comparator clause’, which results in a comparison of different treatments concerning non-competing products. For instance, the true question in Occidental was not whether the differential tax treatment between exporters of oil and exporters of flowers is discriminatory, but whether the tax treatment of foreign invested oil exporters is more burdensome than necessary or whether it violates legitimate expectations. It is highly questionable, however, whether the extensive interpretation of a non-discrimination obligation along the lines of Occidental is in conformity with the contracting parties’ intended level of economic integration.
III. A FLEXIBLE FACTOR-BASED STANDARD OF NON-DISCRIMINATION
Considering that the interpretation of non-discrimination obligations in international economic law lacks coherence—which in turn creates legal uncertainty for the contracting parties, individuals and parties to a dispute—this part attempts to develop a factor-based framework for non-discrimination obligations. Pauwelyn in particular has argued that the non-discrimination analysis should treat ‘likeness’ as a mere threshold question and focus more specifically on ‘less favourable treatment’ as the substantive test, taking into account a mix of elements to determine whether differential treatment is based on origin.Footnote 81 This part takes up this theory, suggesting that the entire analysis of non-discrimination obligations could be viewed as a threshold question. In other words, ‘less favourable treatment’, ‘likeness’ and other ambiguous elements such as ‘so as to afford protection’ or ‘regulatory purpose’ should no longer be implemented as strict legal conditions to be proven by the complainant or the respondent pursuant to the applicable standard of review. Instead, all the relevant elements could be viewed as factors to be weighed and balanced in order to come to an overall conclusion on whether or not a measure amounts to unlawful discrimination of foreign market actors.
Importantly, this article does not address the question whether specific existing non-discrimination provisions of WTO law, the TFEU, NAFTA, BITs or other international economic treaties provide a textual basis for such an approach. The more modest aim is to make adjudicating bodies aware of the significance and mutual relationship of the different legal elements and to propose an alternative approach to the concept of non-discrimination for the negotiations of future national treatment and MFN provisions.
A. Formal Basis of Differential Treatment
The first analytical step should be to determine whether the measure differentiates directly on the basis of origin (de jure) or on the basis of other criteria (de facto). All forms of de jure differentiations affecting the competitive opportunities to the detriment of foreign market actors constitute a strong factor pointing towards unlawful discrimination. The discriminatory effect of such measures does not need any further analysis. However, the respondent must still have the opportunity to justify the measure by proving its legitimate policy objective and a strong nexus between the measure and the objective under scrutiny. In addition, the respondent may show that the measure does not accord a competitive advantage to domestic market actors due to the complete absence of any even remotely competing domestic goods or services.
In case the measure differentiates on a basis other than origin (de facto discrimination), the analysis needs to focus strongly on the effect of the measure in a specific market situation (below, Part III.D).
B. Form of the Measure
A second formal element that needs to be taken in consideration is whether the measure is adopted in the form of an individual and concrete decision adopted by an authority applying exclusively to a foreign market actor, or whether the measure consists of a general and abstract regulation applying to all products in a relevant market. The form of the measure has no value in itself for purposes of the weighing and balancing test. In other words, a decision is not per se more or less discriminatory than a regulatory measure. However, the form of the measure must be viewed in context with its effect.
For instance, a foreign investor subject to a decision should be able to demonstrate the measure's discriminatory effect by showing that only one domestic competitor is not subject to the same competitive constraints (‘best treatment’ standard), provided that the foreign and domestic investors are in like regulatory circumstances (‘subjective standard’ of non-discrimination). Conversely, it is unlikely that one single foreign market actor suffering a competitive disadvantage from a regulatory measure could claim that it is subject to discriminatory treatment if all its foreign and domestic competitors are not negatively affected by the same measure.
C. Extent of Competitive Relationship between Comparators
The current terminology used in non-discrimination provisions referring to ‘like’ or ‘similar’ products, situations or circumstances does not do justice to the underlying question of whether or not there is a competitive relationship between the foreign and domestic market actors. Often the wording of the comparator clause invites the adjudicating bodies to take into account the purpose of the regulation. While this subjective approach may be appropriate from a pragmatic perspective due to the lack of a general exceptions clause in the respective agreement, it does not live up to the required standards for legal security, consistency and transparency. Conversely, current comparator clauses may allow the adjudicating bodies to find illegal discrimination between market entities which are not in a competitive relationship at all. This approach is also not satisfying as it opens the door to an unlimited number of irrelevant tertia comparationis, thereby stretching the principle of non-discrimination beyond its original purpose and scope. By comparing, for instance, an export tax on oil with the absence of such a tax on flowers, the true rationale is not whether foreign oil investors are treated less favourably than domestic growers of flowers, but whether the export tax on oil constitutes an unnecessary obstacle to trade or investment. Hence, regulations with no effect between domestic and foreign competitors should not be dealt with under a non-discrimination obligation, but under the more integrative instruments such as non-restriction in trade or legitimate expectations in investment protection.
To date, the economic concept is best reflected in Ad article III paragraph 2 of the GATT Annex I, which incorporates a standard of ‘directly competitive or substitutable’ products. Another interesting solution had been proposed during the GATS negotiations. An early draft of the GATS national treatment provision contained a specific reference to the marketplace, prohibiting differential treatment of foreign and domestic services and suppliers ‘in the same market’.Footnote 82 Future non-discrimination provisions in international economic agreements should follow these examples, prompting the adjudicating bodies to assess the approximate extent of the competitive relationship by focusing on the economic theory of demand substitutability.
D. Competitive Effect of the Measure
Once the approximate extent of the competitive relationship between the foreign and domestic market entity has been established, the question becomes whether there is less favourable treatment of the foreign entities. For this purpose, the adjudicating bodies need to analyse the competitive effect of the measure, both qualitatively and quantitatively, as well as the ratio of affected domestic and foreign market actors.
1. Qualitative Analysis
The qualitative analysis of a measure requires identifying the parameters of competition which are affected by a measure. Parameters of competition are market factors such as price, quantity, quality, marketing, general conditions, terms of delivery, customer service, etc. For instance, measures banning foreign products from the market, thereby neutralizing all parameters of competition, or measures affecting key parameters such as price or quantity of certain foreign products, are of the highest qualitative burden. In comparison, product regulations, technical regulations, environmental regulations or administrative burdens may result in lower qualitative burdens, depending on which parameters of competition they affect. In such cases, the complainant needs to demonstrate how and to what extent the measure puts the foreign product or investment at a competitive disadvantage as compared do its domestic competitors.
2. Quantitative Analysis
Even more importantly, a measure's quantitative effect on competition must also be taken into consideration. The quantitative effect is primarily determined based on the additional burden placed on the foreign market entities by the measure. Ideally, the additional burden should be assessed in costs, which is relatively easy in the case of taxes. For instance, depending on the cost of the actual product, a tax differential of 1 per cent may be considered to have a low competitive effect, whereas a differential of 30 per cent would presumably result in a strong distortion of competition.
The quantitative effect of the measure may then be placed in relation to the competitive relationship. If the foreign and domestic entities are in a very close competitive relationship, then a very small quantitative effect of the measure—such as a small differential tax or a small administrative burden—may be sufficient to constitute a breach of the non-discrimination obligation. Conversely, if the market players only compete very remotely, then the quantitative effect of the measure needs to be of a higher intensity so as to amount to illegal discrimination. This rationale of placing the quantitative effect of the measure in relation to the competitive relationship is currently reflected in the national treatment provisions of articles III:2 GATT and 110 TFEU with regard to taxes. For instance, if the products are ‘like’ in terms of article III:2, first sentence, GATT (ie competing and physically similar), every even very small difference in taxation to the detriment of imported products meets the ‘in excess of’ requirement and thus violates national treatment; no de minimis exception is granted.Footnote 83 In contrast, if the products in question are not ‘like’, but nevertheless in direct competition or substitutable (ie competing regardless of physical differences), the requirements on the difference in taxation are more strict. In these cases, a supplemental de minimis tax on imported products is not sufficient to find a breach of GATT article III:2, second sentence, and, unlike in the case of the first sentence, the taxation must be construed so as to afford protection.Footnote 84
3. Ratio Analysis
Another very important aspect that needs to be taken into account for the weighing and balancing test is the competitive effect of the measure on domestic as compared to foreign market entities. A new regulatory measure may have a positive or a negative competitive effect on a concerned market entity, or it may have no competitive effect at all. The ratio analysis requires assessing how the positive, negative or neutral effect is distributed among the foreign and domestic market entities. Instead of imposing a pre-defined standard, such as the ‘disproportionate disadvantage test’ or the ‘best treatment’ approach, the adjudicating bodies should have the flexibility to weigh and balance the ratio in light of the competitive relationship, the relevant market and the form of the measure. For instance, a measure is likely to amount to unlawful discrimination if it negatively affects predominantly foreign market entities in a narrowly defined market (ie with strong competitive relationship and high demand elasticity). Conversely, on the other end of the spectrum would be a measure which negatively affects only few or one foreign market entity in a broadly defined market (ie low competitive relationships with low demand elasticity).
In sum, the higher the competitive relationship between the market entities which are affected by the measure, the fewer foreign entities need to be negatively affected for the measure to amount to unlawful discrimination (‘best treatment’ approach). Conversely, if the entities affected by the measure are only remotely in competition, then discrimination would only occur if the measure negatively affects predominantly foreign market entities (‘disproportionate impact’ approach).
4. Supply Substitutability and Temporal Considerations
A final aspect that may be taken into consideration when assessing the effect of a measure relates to the issue of supply substitutability and temporal markets. Supply substitutability focuses on the question of what costs and within which time frame a supplier could switch its production from product A to product B. This analysis is particularly pertinent to assess the market position of a company for purposes of antitrust law. In the context of non-discrimination, however, the focus lies not on the market position, but on the ability—in terms of cost and time—of the foreign market entity to escape the negative effect of the measure by, for instance, making its product compliant with the regulation under scrutiny.Footnote 85 More specifically, if a foreign producer is banned from importing its product because it fails to meet a new environmental standard, then the discriminatory effect could be mitigated by the fact that the foreign producer could relatively easily (ie at low costs and within a short period of time) switch its production from the non-compliant to a compliant product. Consequently, high supply substitutability could be considered as a factor supporting the conclusion that a measure does not amount to unlawful discrimination. Ideally, the costs for the foreign market entity to switch from a non-compliant to a compliant product or service should be viewed in relation to the importance and value of the policy objective pursued by the measure.
E. Regulatory Purpose of the Measure
A final element which may be taken into consideration is the regulatory purpose of the measure under scrutiny. Importantly, the regulatory purpose should have its own value and should not be incorporated into the analysis of the previous elements.
1. Protectionist Purpose
It is up to the complainant to produce evidence of a protectionist purpose. In the rare cases where direct evidence shows that a measure was adopted with the aim to pursue a protectionist purpose, such evidence should be considered as a very strong—albeit not by itself decisive—indication that the measure amounts to unlawful discrimination. Direct evidence could be obtained from official documents or press coverage related to the legislative history of the measure.
In most cases the complainant is more likely to produce circumstantial evidence pointing towards protectionist intent of the responding government. However, the most important type of circumstantial evidence is the effect of the measure which is already taken into account as its own element. In addition, circumstantial evidence could relate to the design, structure and application of a measure.Footnote 86 Elements to be considered are, for instance, the purpose and objective of the government and legislature ‘to the extent that they are given objective expression in the statute itself’,Footnote 87 unexplained changes in drafts during the legislative historyFootnote 88 or the application of differential standards.Footnote 89 Finally, the absence of a credible non-protectionist purpose could also be considered as circumstantial evidence of a protectionist purpose.
2. Non-Protectionist Purpose
While the complainant seeks to show protectionism, the respondent may submit evidence demonstrating that the measure was adopted with the objective to protect a public interest, such as health, morals, public order, environment, natural resources etc. Such an alleged non-protectionist purpose must not only be balanced with the previous elements, but it must also be analysed under the principle of necessity or, more broadly, the principle of proportionality.
a) Importance of the Public Interest
The contracting parties have it in their hands to negotiate either an exhaustive or an enumerative list of public interests that may be protected in violation of a non-discrimination obligation. Considering that the need for the protection of unforeseen interests may arise at short notice, an open list of public interests appears to be preferable. In many cases the contracting parties would have great difficulty to find a consensus for amending the exceptions of an existing international economic agreement. However, in order to prevent overreaching justifications, the adjudicating bodies need to have the authority to evaluate the objective value or importance of the allegedly protected public interest. This may be a difficult task as it entails second guessing the national values of sovereign contracting parties with different institutional, political and religious traditions.Footnote 90 Consequently, the value of the pursued public interest cannot by itself be a decisive element in the analysis, but more a consideration that reinforces the tendencies of the previous elements. For instance, the protection of a minor public interest may reinforce the conclusion that a measure is not discriminatory in cases where the measure differentiates de facto between remotely competing market entities and has a low qualitative and quantitative effect on few foreign entities. Conversely, a universally very important public interest, such as human life and health, may be used to justify measures which differentiate on a formal basis or between strongly competing market entities to the detriment of predominantly foreign entities.
b) Nexus between the Public Interest and the Respondent's Territory
Following the general territoriality principle of public international law, a sovereign State is generally prohibited from adopting extraterritorial measures which infringe the sovereignty of another State. Consequently, the question whether and to what extent a nexus exists between the policy objectives pursued by the regulatory State and its territory needs to be taken into account under the aspect of regulatory purpose. This nexus is usually existent in cases where a measure concerns a foreign investor or service supplier who is present on the territory of the regulatory State. In the case of trade in goods, no concerns of territoriality arise if the measure sets certain standards in order to pursue a domestic non-economic public interest, such as the protection of domestic health, environment or consumers. In contrast, an importing State may be inhibited from restricting the import of products for reasons of insufficient process or production methods used in the State of production, unless there is a nexus between the ‘extraterritorial’ value and the domestic territory. For instance, in the case US—Shrimp the Appellate Body recognized a sufficiently strong nexus between the United States’ import ban on shrimp caught without a turtle excluder device and the United States territory due to the fact that sea turtles sought to be protected may migrate to the waters subject to United States jurisdiction.Footnote 91 The ban was designed to protect exhaustible natural resources under article XX(g) GATT.
c) Nexus between the Measure and the Objective
Finally, the most important element that needs to be considered in the analysis of the regulatory purpose is the extent of a nexus or causality between the trade restrictive measure and its objective. Such a nexus is crucial in order to avoid the risk that a measure under scrutiny is overly trade restrictive in view of achieving the pursued objective. For instance, a total import ban of cigarettes may not be necessary to achieve certain objectives related to ensure the quality or to reduce consumption of cigarettes.Footnote 92 This aspect is currently embodied in the general exceptions clauses, which require that a measure must be ‘necessary to protect’ or ‘related to the protection of’ a certain public interest (see eg articles XX GATT and XIV GATS). The factor-based approach to non-discrimination proposed here would not prescribe a fixed threshold, such as ‘necessity’ or ‘related to’, but it would allow the adjudicating bodies to weigh and balance nexus-related factors in light of the analysis of the previous factors. The necessity test, for instance, requires a process of weighing and balancing a series of factors, namely (i) the relative importance of the interest protected by the measure, (ii) the contribution of the measure to the protection of the policy objective and the public interest, (iii) the impact of the measure on trade, (iv) the existence of alternative measures in light of (v) the level of protection chosen by the responding Member (eg zero-risk level).Footnote 93 Under the factor-based approach, the arbitral tribunals would not be bound by a specific threshold, such as ‘necessary’ or ‘related to’, but they would add the extent to which the measure contributes to the protection of the policy objective as an additional factor to the overall analysis.
IV. SUMMARY AND CONCLUSIONS: A FLEXIBLE RANGE OF STANDARDS
This article analysed how each of the legal elements ‘less favourable treatment’, ‘likeness’ and ‘regulatory purpose’ constituting the non-discrimination obligation may be subject to different interpretations and how each interpretation affects the scope and integrative effect of the non-discrimination principle. Depending on how each element is interpreted and combined with the respective interpretation of another element, the non-discrimination obligation can be extremely intrusive or very permissive.Footnote 94 The following interpretations are possible:
Comparator Clause
– Under the objective standard the tertium comparationis may consist of factors such as physical characteristics, tariff classification, end-uses, environmental impact or even the act of exportation;
– Under the economic standard, the tertium comparationis is defined by economic parameters indicating the extent to which the market actors are in a competitive relationship;
– Under the subjective standard the tertium comparationis is defined by the regulatory purpose of the measure under scrutiny;
– The objective, economic and subjective standards of ‘likeness’ may be applied individually or in combination.
Less Favourable Treatment
– The disproportionate disadvantage test analyses whether the group (as defined by the comparator clause) of foreign market participants is disproportionately disadvantaged as compared to the domestic one. Different thresholds of disproportionality are theoretically possible;
– The ‘best treatment’ approach leads to an obligation to grant the best treatment accorded to any domestic market participant to all foreign ‘comparable’ (as defined by the comparator clause) market participants;
– The subjective standard of ‘less favourable treatment’ takes into account the regulatory purpose in order to determine the true basis of the differential treatment.
Regulatory Purpose
– The regulatory purpose may be considered as part of the non-discrimination obligation itself (definitional stage), either (i) within the comparator clause (aims [and effects] test), (ii) within the ‘less favourable treatment’ element, or (iii) as its own substantive element;
– Alternatively, the regulatory purpose may be considered as part of a general exceptions clause (justificational stage).
Considering this wide variety of different possible interpretations and the lack of any guidance in most international economic law treaties, it would be preferable for such treaties explicitly to empower the adjudicating bodies to construe the non-discrimination obligation on a case-by-case basis. At the same time, the treaty should spell out the factors which the adjudicating bodies need to take into consideration under an overall weighing and balancing test. This article suggests an approach along the following lines:
First, the treaty should specifically spell out that all forms of de jure discrimination are considered a prima facie violation of the non-discrimination obligation and thus create a presumption of illegality. The respondent may only justify the measure by showing (i) that it has no impact on the competitive relationship or (ii) that it protects an important public interest and that there is a strong nexus between the measure and its objective as well as between the measure and the respondent's territory.
Second, prima facie violation of the non-discrimination obligation may also be established in case the complainant is able to produce direct evidence proving the protectionist intent on behalf of the respondent. Measures specifically designed to protect certain domestic market actors would not be justifiable. However, as a complainant will hardly ever be able to produce such evidence, this intent-based standard of non-discrimination is likely to remain theoretical.
Third, with regard to measures not formally differentiating on the basis of origin, the treaty should explicitly state that such measures are subject to an analysis of their protectionist effect in the market place. The following criteria need to be taken into consideration for the assessment of the measure's protectionist effect:
– The extent of the competitive relationship between disadvantaged foreign market actors and domestic market actors;
– Form of the measure (ie individual and concrete decision or general and abstract regulation);
– The measure's qualitative effect on competition (ie parameters of competition restricted by the measure);
– The measure's quantitative effect on competition (ie extent of the additional burden or the competitive disadvantage imposed by the measure);
– The measure's effect on domestic vs foreign market entities (ie ratio of negative competitive impact between domestic and foreign market participants);
– Supply substitutability (ie costs and time for foreign market actors to escape the negative competitive impact of the measure);
– Regulatory purpose of the measure, taking into account:
(i) the relative importance of the public interest sought to be protected;
(ii) the nexus between the measure and its objective; and
(iii) the nexus between the measure and the regulatory country's territory.
The adjudicating bodies need to conduct a weighing and balancing test on the basis of these factors in order to determine whether a measure amounts overall to unlawful de facto discrimination. For instance, a clear violation of the non-discrimination obligation would occur in case (i) a measure affects only or predominantly imported market entities (ratio analysis) (ii) and constitutes a strong negative impact on an important parameter of competition (quantitative and qualitative elements) (iii) in a narrowly defined relevant market consisting of products, services or investments with high demand substitutability (competitive relationship); moreover (iv) foreign participants are unable to substitute their good or service with another one not subject to the trade restrictive measure (supply substitutability) and, finally, (v) the measure does not pursue a legitimate policy objective (regulatory purpose).
In practice it would be very rarely the case that all the factors clearly show that the measure under scrutiny either violates or complies with a non-discrimination obligation. Hence, under the factor-based approach suggested here it would still be difficult to accurately predict whether a given trade restrictive measure violates a certain non-discrimination obligation. However, the main advantage of this approach is that at least there would be significant legal security with regard to the applicable legal test and, to a certain degree, the legal standard of non-discrimination obligations. Due to the clarity of the legal test, it would be much easier for the parties in a dispute to make their case by bringing forward the pertinent arguments. At the same time, such a factor based-test for non-discrimination is sufficiently flexible to allow adjudicating bodies to consider the framework and purpose of the agreement in which the obligation is found. Most importantly, the arbitral tribunals could set the appropriate standard in light of the fact that trade agreements aim to protect competitive opportunities and equal conditions of competition, whereas investment protection agreements are designed to protect the value of a specific investment.
Finally, it is important to point out that the factor-based solution for non-discrimination provisions entails the risk that the analysis will focus much more on the purpose and necessity of a measure than on the elements of differential treatment and comparability. In fact, adjudicating bodies may be tempted to qualify a measure as discriminatory if the measure does not pursue a legitimate policy objective or if the respondent fails to demonstrate a nexus between an alleged objective and the measure. Hence, arbitral tribunals must pay attention not to overemphasize the possible absence of a legitimate purpose in comparison to the elements of ‘likeness’ and ‘less favourable treatment’, in order to avoid undue extension of the non-discrimination provision into a more integrative obligation of non-restriction. The interpretation and application of a non-discrimination obligation must respect the contracting parties’ intended level of economic integration.
