Competing Transnational Regimes under WTO Law

Against a common perception of CSR being a business concept without binding legal effect, this article discusses legitimate legal effects of private standards in public international law, using the issue of private labels as “international standards” under WTO law. WTO law shows certain openness for external transnational standards. This article argues that the references to “international standards” in the TBT Agreement can be applied for the selection between competing public or private norms that claim relevance. Thereby, the most legitimate standard for governing the problem at issue should be chosen. This is exemplified with the case of Tuna Dolphin II where the Appellate Body has emphasised the requirement of procedural legitimacy. The article argues that the requirements for legitimate standards depend on the interests at stake and that a private standard can well be more legitimate than a (competing) public standard. As the justifying effect of Article 2.5 TBT mainly interferes with economic interests, a relevant “international standard” may well consist of a representative business standard, e.g. a private label. In contrast, an international standard in the terms of Article 2.4 TBT which interferes with a democratic decision in favour of public interests such as environmental protection must reflect these public interests in a legitimate way. The article concludes that CSR can play an important role in defining legally valid justifying or minimum standards in public international law.


I. Introduction
Economic globalisation and its accompanying international legal framework has led to a decrease of the regulatory power of the national parliamentary system related to the balancing of (transnational) economic interests with social and environmental protection aims. This regulatory power cannot be replaced by comparable (public) international law making -neither in content nor with respect to legitimacy considerations.
At the same time, various forms of Corporate Social Responsibility (CSR) instruments have been developed, in particular labels and codes of conduct, in order to address regulatory gaps. These initiatives include rules or regulatory programmes negotiated between different stakeholder groups, voluntary rules initiated by national, supranational or international institutions or by NGOs, but also the unilaterally adopted 'self'-regulation of transnational corporations related to social and environmental aspects of their activities (' corporate social responsibility'). 1 These private rules have gained enormous practical importance for the regulation of public goods such as health and safety concerns, working conditions or the environment. The question remains, however, whether and under which conditions that practical importance may also lead to legal effects, in particular by producing legal standards for socially and environmentally responsible corporate activities.
Traditionally oriented scholars insist on the exclusive validity of classical State law and the respective legitimacy chains; 2 which seems to exclude any legal relevance of private regulation. Another school of thought based on systems theory represents the other end of the spectrum, recognising private (self-) regulation in the different 1 On the various forms of private regulation see also Martin Herberg, Globalisierung und private Selbstregulierung (Campus Verlag 2007); Olaf Dilling and Martin Herberg and Gerd Winter (eds), Responsible Business: Self-Governance and the Law in Transnational Economic Transactions (Hart Publishing 2007 Against a common perception of CSR being a business concept without binding legal effect, this article discusses legitimate legal effects of private standards in public international law, using the issue of private labels as "international standards" under WTO law. WTO law shows certain openness for external transnational standards. This article argues that the references to "international standards" in the TBT Agreement can be applied for the selection between competing public or private norms that claim relevance. Thereby, the most legitimate standard for governing the problem at issue should be chosen. This is exemplified with the case of Tuna Dolphin II where the Appellate Body has emphasised the requirement of procedural legitimacy. The article argues that the requirements for legitimate standards depend on the interests at stake and that a private standard can well be more legitimate than a (competing) public standard. As the justifying effect of Article 2.5 TBT mainly interferes with economic interests, a relevant "international standard" may well consist of a representative business standard, e.g. a private label. In contrast, an international standard in the terms of Article 2.4 TBT which interferes with a democratic decision in favour of public interests such as environmental protection must reflect these public interests in a legitimate way. The article concludes that CSR can play an important role in defining legally valid justifying or minimum standards in public international law.

A. Preliminary Remarks
The approach taken here, towards the legitimacy of law, is based on the ideas of consensus of those who are affected by the relevant norm and of common welfare. The consensual basis of law is reflected in republican ideas of democracy as well as in deliberative theory of law. Whereas private law is directly based on the idea of private autonomy and self-regulation through contracts and other forms of legal acts (ideally) based on the direct consent of the parties involved, State law must, from a classical public law perspective, be constructed as being based on (national) consensus. Here, the legitimacy chain, with the possibility of equal input through elections, replaces direct consensus and safeguards the dedication of the output to common welfare. 9 The dedication to common welfare -which includes all citizens and which requires rules theoretically acceptable to all citizens or members of the polity 10 -is important as an additional legitimacy requirement because the majority rule has to be legitimised with a view to governed minorities, who must not be disadvantaged. 11 Legitimacy of the majority rule also presupposes some homogeneity of the constituency.
The situation is more complicated in a transnational multi-level system with a plurality of -partially highly technical -public, private and semi-private regimes in place. Here, new modes of interplay between these regimes and especially with regard to the different private regulatory contributions -be they negotiated by stakeholders or unilaterally adopted by the business side -need to be found, whilst always keeping in mind their respective legitimacy with view to the relevant constituency.
One important aspect of consensus relates to the level at which a decision is made. This is addressed by the principle of subsidiarity, which builds upon more basic principles such as individual freedom, self-determination, participation and democracy. In a political multi-level-system, subsidiarity demands the allocation of decision competences at the most immediate level that is able to handle the issue in question effectively and that includes all interests concerned. 12 The higher the level of decision making, the less influence individuals traditionally have. People are often mistaken about the significance of (subjectively) distant decision levels. 13 In its original meaning the principle of subsidiarity applied in an all-embracing way so as to include the relationship between societal self-regulation and State law. Yet, the principle of subsidiarity is in a certain tension with the principle of effectiveness of law 14 and with the requirements of free trade and economic globalisation which again raise questions about the relevant constituency. 15

B. The Legitimacy Chain and its Democratic Deficiencies
The classical approach towards the legitimacy of political governance focuses on the legitimacy chain from the voting act over parliament to executive. This hierarchical approach is based upon the strict separation of the State from society and on the assumption of homogeneous peoples without regard to the plurality of societal interests. 16 9  In functionally differentiated societies with growing technological complexity, however, the concretisation of very general laws by the administration or through private norms can hardly be based, any longer, only on the legitimacy chain. 17 Here, substantial decisions with regard to the relation between economic freedoms and different protection aims, for example, health and environment, are de facto already taken far from parliament at the national level. 18 This is aggravated at the international level where legitimacy chains are even longer. First of all, the nationally based democratic mandate does not correspond to the territorial scope of international agreements. Moreover, the regulatory power has shifted from the legislature to the executive with the consequence that decisions can hardly be attributed to societal preferences. Negotiations are carried out by governmental elites that are disconnected from the national political process and may be dominated by 'geopolitical power plays' with the consequence of reproducing power imbalances between States. 19 The shift of power to the executive damages national democratic procedures even more and the more international agreements affect internal national politics. Besides, although there is a need for international co-operation, international agreements are difficult to reach and once an agreement is concluded it is even more difficult to modify it in accordance with changed societal preferences. 20 As a consequence, citizens no longer feel like authors of the (international) rules and national democracies lose credibility. 21

C. Economic Globalisation and the WTO
These problems are particularly visible in WTO law and its effects on national politics. The requirements of free trade and economic globalisation have considerably reduced national regulation margins. Uniform standards are needed to make economies and products compatible. Global economic competition leads to competition of national legal orders in the area of production costs. This is true for vivid concerns of national economies such as fiscal revenues and social security systems but also for labour standards, 22 safety at work and environmental protection. 23 This competition creates a need for global standards in order to avoid a 'race to the bottom'. 24 The regulation of transnationally traded products has de facto and de jure become an international or transnational issue. 25 The legal framework of the WTO aims to juridify economic globalisation and the interrelation of the different national economies in order to avoid arbitrary measures or regulation with negative consequences for other economies. 26 The legal emphasis on free trade, however, has also reduced national regulation margins further with effect on social protection regulation such as food safety, agriculture, health and safety or the environment, which could qualify as 'non-tariff barriers to trade'. 27 In addition to this, the WTO has a Böckenförde, Recht, Staat, Freiheit (Suhrkamp 1991) 27 See also Christian Joerges, 'Freier Handel mit riskanten Produkten? Die Erosion nationalstaatlichen und die Emergenz Transnation-fairly effective enforcement mechanism in place, compared to other public international law agreements. Therefore, the legitimacy of WTO law is of major concern. 28 Indeed, serious doubts have been raised as to whether the long legitimacy chains in public international law might be too long to legitimise WTO law and, in particular, its impact on social or environmental protection at the national level. Again, the arguments put forward are that: the WTO is dominated by the executive; bureaucratic-governmental elites enjoy (relative) autonomy; negotiations are secretive; there is no open public discourse which constitutes an essential element for democratic legitimacy; and there is only weak parliamentary control by ex-post ratification on a swim or sink basis, so that in fact no means exist to adjust or change WTO law as negotiated by the executive. Besides, as economic globalisation and economic concentration have led to an enormous shift of power from politics to economy and to relatively close relations between governmental and economic elites at the international level, 29 these detached negotiation systems lack basis, motivation and power to take countermeasures against the power of market forces. 30 As a result, social protection goals are structurally underrepresented. Finally, the WTO has a very independent judicial review mechanism, the decisions of which cannot be politically embedded or corrected. 31 In fact, as can be seen from reports by the Panel and the Appellate Body, precedents do have considerable impact on the further interpretation of WTO law.

D. Consideration of other Levels of Regulation as Compensation Strategy within Multilevel Regulatory Frameworks
In order to compensate for these legitimacy deficiencies, authors have suggested interpreting WTO law in a restrictive manner so as to concretise the principle of non-discrimination in such a way as to not aim at market integration and deregulation (or even at regulatory competition). According to these authors, the law of the Nation States should be given as much regard as possible taking into consideration that the Nation State is the forum that reflects those interests which would otherwise have no standing, such as social security or environmental protection. 32 The reality of the dispute settlement, however, has been quite different. 33 Another approach is to interpret or concretise WTO law in such a way as to give the WTO democratic backing by taking account of factors other than national consensus driven fair and pluralistic regimes or standards, be they private or public. 34 One mechanism to achieve this would be to concretise open-textured terms such as the term 'international standards' with recourse to public or private norms developed in other negotiation systems.
This approach has been criticised as being unsuitable as far as the incorporation of the standards of the Codex Alimentarius Commission into the SPS Agreement is concerned. 35 In contrast, there is potential in the contestation of international standards in the openly worded TBT Agreement and in the scrutiny of their claim for legal relevance. 36 This article takes a second look at this latter approach and at the legitimate potential of regimes based on other strands of legitimacy such as self-regulation of (certain sectors of) society or stakeholder representation in WTO law, thereby taking the above-mentioned collisions-law perspective.  36 See also Scott (n 12) 307 ff. This approach does not mean to put the WTO in a position of supremacy towards other regimes, but rather aims at imposing a legitimacy check according to their legal effect within WTO law, see ibid 349 ff.

Participation and Deliberation
As a strategy to compensate for the weak(ened) legitimacy chain caused by functional differentiation and globalisation, participatory elements have been introduced into administrative decision-making as well as into European and international decision-making, thus aiming to relate the respective decisions back to society and the plurality of interests within society. These participatory rights have gained particular importance in environmental law. 37 Certain possibilities of participation also exist in international economic governance. One example in the framework of the WTO is the possibility to submit so-called amicus curiae briefs. 38 It has long been acknowledged that participatory arrangements contribute to the quality and the legitimacy of decisions since participation increases information and substantive correctness, transparency and control, integration, balancing of interests and acceptance.
Participatory arrangements have also been characterised as democratic, in line with a deliberative approach towards democracy, since they respect citizens as actors in deliberation and acknowledge the plurality of societal interests and the necessity to connect decisions not only formally but also substantively back to society. 39 According to this approach, participatory procedures would not replace the classical democratic legitimacy chain, which secures inclusion and general and equal voting, but constitute a necessary complement. Participation, therefore, does not need to be a perfect copy of society, as long as participatory rights are not exclusive privileges but generally accessible and open to all interests concerned. 40 Another important aspect of deliberative democracy is the legitimating power of 'rational discourse' based on arguments and understanding. 41 This presupposes that participative procedures guarantee that interests and arguments are heard in accordance with their normative weight rather than with the factual power of their representatives -ideally in an institutionalised form. 42 According to this approach, administrative participatory decision making would, notwithstanding certain tensions, be a complement to parliamentary democracy, rather than contradicting it. 43 With regard to transnational rule-making, however, this approach bears the challenge of the absence of a central representative law-making body as a point of reference. Therefore, participative deliberative procedures would have to 'shoulder the democratic burden alone' with the difficulty of defining and representing the respective constituency or public. 44

Stakeholder Representation
With respect to stakeholder representation transnational participatory arrangements bear another legitimacy challenge. Different to national administrative decisions where in fact the people concerned are able to participate, participation at the international level, mainly involves so-called 'stakeholders'. The relevant question in this respect is whether these stakeholders, in particular Non-Governmental Organisations (NGOs), do, in fact, represent the relevant interests in an accountable manner. Critics have argued that NGOs lack democratic legitimacy themselves as they are not elected representatives. 45 Taking a deliberative approach towards democracy again, one cannot overlook that there is an added democratic value in the participation of stakeholders, especially at the international level, as they: transfer information and generate transparency and accountability in the international system; contribute to the formation of a public opinion and to a transnational public sphere; generate a counter-public to official and economic statements; and, last but not least, put in their own expertise and therefore enhance the quality of decisions. 46 It is also acknowledged that NGOs can in principle be of particular importance to the representation of interests, values and preferences of minorities, of vulnerable persons, of those who are not organised otherwise, of those who are not represented by governments and of public interests which somehow got lost in the political process. 47 Whether or not an NGO does in fact represent the asserted interests depends on the individual case. 48 For environmental protection as such, for example, there has been no better representation than by environmental organisations until now. Moreover, NGOs live from their reputation, which puts them under public control or the control of the relevant particular interests. Besides, at the international level, relations between governmental and economic elites are far closer than relations to the peoples or to (unorganised) civil society. Therefore, business exerts influence anyway. 49 Thus, institutionalised participation of stakeholders representing other interests, although perhaps imperfectly, still increases the legitimacy of decisions in relation to those which are taken under uncontrolled influence of the economy. 50 Generally speaking, the quality of deliberation increases with more voices being heard.

Standardisation and Deliberation
Institutionalised standard setting activities at national, European and international level are one important field where participative procedural requirements have been introduced in order to ensure the (procedural) legitimacy of the standards themselves. 51 A common formula is that standardisation procedures should be open, plural, participative and transparent, based on well-balanced expertise and ensure reversibility and accountability in order to ensure a fair and deliberative consensus of the interests concerned. 52 For this reason, norm setting in such 'ideal' plural and competent committees has been positively characterised as part of a de-central, work-sharing and deliberative democracy, 53 where far more problemoriented solutions between competing interests -economic interests vs. social, environmental and other public interests -for specific problems can be developed in the traditional political process. This is even more so in the era of globalisation where new balances must be found between the interests of transnational business on the one hand and of civil society aiming for social and environmental protection on the other and where the latter are even less represented at the executive-dominated international level than at the national level. 54 Accordingly, those negotiation systems, provided they satisfy the outlined preconditions, may achieve better results in terms of outcome and legitimacy. Similar considerations could apply to private standards or labels. 55 4. Self-Regulation and Normative Generalisation The integration of private (self-)regulation, such as rules or labels negotiated between different stakeholder groups or even the unilaterally adopted self-regulation of transnational corporations (' corporate social responsibility'), by the legal system would be another means to relate law back to society. This article, however, does not argue in favour of blind acceptance of private regulation, such as 'societal constitutionalism'. 56 In contrast, it advocates controlled legal recognition only in cases where private regulation can indeed be regarded as a sort of self-regulation and where it adds legitimacy or regulatory value to the multilevel regulatory framework. 46 Therefore, NGOs are also called as 'transmission belts' in the international system, see eg Steffek (n 45)  The prerequisites of legitimate recognition by law mainly depend on the legal effect of the normative reference in question and on the extent to which a rule of private origin gains relevance beyond its authors or signatories through its recognition by the respective legal regime. Thus, the concrete legitimacy requirements depend on the interests concerned.
In principle, three levels of generalisation of a private rule through normative reference can be distinguished: the regulation of the internal relationship between those who have agreed on the rule; a broader binding effect also on ' outsiders' of the same group or industry as the signatories ('minimum standard'); and an absolute standard setting effect which also concerns third parties or the public interest ('safe harbour' or 'maximum standard'). 57 Of particular interest here is the legal effect of a 'minimum standard' as opposed to a 'safe harbour' for business conduct as it only negatively affects the 'business side' and can therefore be based on consensus requirements within the relevant line of business, as detailed out below.

a. Internal Regulation
Where the application of a private rule merely binds those who have agreed on the rule and therefore safeguards the private autonomy of the actors, for example in contract law, only the self-commitment or the consent of the authors of the rule is required. Only a few restrictions exist, namely that the agreement must be neither illegal nor immoral according to the values of the applying regime or some safeguards for the parties' private autonomy, particularly where there is a power imbalance between them. Examples would be the protection of trust in private (business) regulation in contract law or in advertising law through openly worded provisions. 58

b. 'Minimum Standard'
Where the application of a private rule or standard may lead to the establishment of minimum requirements for a certain group of actors, for example, businesses in the same industry, including ' outsiders' who have not signed up to the private rule, the legitimacy towards these ' outsiders' must be safeguarded.
The internal regulation of (private) associations, for example of professional associations, has been the classical case for this issue in legal debate since usually no direct approval of each internal rule by every member is possible or required and still the rules are binding on all members of the association. In academic writing, consensus within the group is nevertheless regarded as the basis for legitimacy, but it is complemented by the concept of 'group welfare' or 'group interest' which somehow resembles the political concepts of ' common welfare' or 'public interest'. 59 The idea is that group regulation serves the group interest and, therefore, ultimately, the interest of all members of the group. In order to safeguard this objective, the consensus position must reflect the group interest in such a way that it is theoretically acceptable to all members of the group. 60 This requires that the rule be adopted by a representative variety of the group members in order not to unfairly exclude certain interests or parties. 61 For example, a rule that is decided on by large companies to the disadvantage of small and medium-sized enterprises would not fulfil this requirement. Also, companies from industrialised countries would not be able to establish rules that are equally valid for companies from the developing world. Furthermore, long-established companies or members of the group would not be able to establish rules to the disadvantage of newcomers. 62 57 For detailed analysis of the following see Carola Glinski These legitimacy considerations do not only apply to internal group regulation but can be transferred to a controlled adoption of private rules in order to concretise legal obligations for the relevant group laid down in general terms of an applying regime, such as ' common usage', ' commercial practice', or 'generally accepted standard'; thereby generalising the private regulation with regard to ' outsiders' of the same group. 63 Other examples are general clauses of 'unfair commercial practices' in unfair competition law or minimum requirements of ' due diligence' in tort law. The legal consequence of such adoption would be that certain business practices that are declared as unfair by private rules adopted by a broad and representative variety of companies (of a certain industry) would be regarded as unfair by law. In tort law, the breach of private diligence standards adopted by a representative variety of companies of the relevant category of producers would constitute liability as the private standard at least reflects what can be regarded as predictable and avoidable by the profession. 64 To the contrary, compliance with these private standards does not constitute a legally 'safe harbour'. Neither has every other practice been regarded as fair by law, 65 nor does it prevent liability as there could be normatively unacceptable 'jog trots' in practice, which do not reflect societal concerns and needs. 66 As third parties or the public interest are not bound by a 'minimum standard', legitimacy considerations only need to take account of ' outsiders' of the same group that may be negatively affected. In this respect, it can be assumed that a representative variety of corporations will safeguard their own group interest sufficiently so that stronger protection by law is not required. The disregard of some few corporations ('black sheep') in the consensus can be justified by the fact that the law itself calls for objective standards that do not take account of each individual opinion. Furthermore, standardisation increases legal certainty and therefore serves the public interest and also the group interest. 67 c. 'Absolute' or 'Maximum Standard' Except for concrete (contractual) duties within a certain legal relationship a private rule might even establish an absolute or 'maximum standard' through application by or generalisation through another legal regime. In this case the private rules would in effect amount to the exclusive definition of legal obligations, a legal 'safe harbour' for those (companies) who comply with the rules. Hereby, the interests of third parties or the public interest might be affected and would have to be reflected in the (private) consensus respectively. The concretisation of public law requirements for the protection of health and safety and the environment by private or semi-private standards is the classical example. Here again, the legitimacy of the standards derives less from self-regulation than from deliberation, in particular through open, pluralistic, fair and transparent decision-making procedures based on expert knowledge, which could compensate for the weakened democratically anchored legitimacy chain.
In a weaker form, in particular in multilevel regulatory frameworks, the application of a private rule might constitute a standard the exceeding of which triggers justification duties. An example would be Article 2.4 TBT.  , 29/11/1983, 29/11/ , NJW 1984, at 802 -ice hockey: 'However, the technical norms do not always determine the utmost that can be required in the individual case but they need to be completed and they do not release the judge from his duty to evaluate the interest of the potential victim to have his integrity protected' -translation by the author). 67 See also Bachmann (n 9) 19 ff.

Requirements for Consensus
But what could be regarded as legitimate consensus of those interests concerned -given that in fact not each and every party concerned really agrees or is able to agree? How to ensure that deliberation and representation works, that interests and arguments are heard in accordance to their normative weight instead of the factual power of their representatives, that a compromise could be regarded as fair and legitimate?
The International organisation for Standardisation (ISO) Guide requires the absence of 'sustained opposition' to 'substantial issues' by any 'important part of the concerned interests' for consensus, 68 which provides useful criteria to guarantee that arguments count according to their normative weight and that no minority with normatively substantive interests or arguments can be overruled.
These considerations do not only apply to multi-stakeholder decisions but also provide useful criteria for decisions of or within a certain group of actors as laid down above.

Ex ante Procedures; Ex post Recognition
In addition to or as compensation for deficient ex ante procedural legitimacy, a norm or a standard can gain ex post legitimacy through acceptance or recognition by the parties concerned. 69 Again, in order to produce legitimacy, acceptance or recognition should be broad and representative. At least, there should be no wellfounded non-adherence by an important part of the concerned interests. However, the mere fact that a rule is recognised does not automatically mean that it deserves recognition. On the one hand, de facto recognition seems to be a surplus to some (theoretical) criteria why a rule deserves recognition. On the other, de facto recognition could rest on necessities or forces which lack legitimising character.
Whereas a fair drafting procedure guarantees -at least to a certain extent -that all parties or representatives of interests concerned are able to give a relevant input in a situation still open, ex post recognition bears the risk of not being based on the free will of all who adhere to the rules but instead on the power of facts already created by some potent actors. Therefore, legitimacy qua ex post recognition depends on the question of whether those who have adhered to the rules had a real alternative.
This problem of the power of facts is less striking with regard to norms concerning political values where greater plurality and reversibility is possible and even necessary (for example, for eco-or fair trade labels) than to purely technical harmonisation. The acceptance and recognition of technical standards, however, is often based on the fact that they have proven themselves in practice, for example, through effective prevention of damage. 70 In case of economic pressure to comply with a private rule the legitimacy of that rule depends: market forces as such, especially consumer expectations which demand adherence with a private standard, are no obstacle to the legitimising effect of the acceptance as the respective reciprocity of expectations is an integral part of the market (logics) itself. 71

G. Interim Conclusion
The different sources of legitimacy gain more or less importance in different situations, at the same time they face more or fewer challenges depending on the level at which decisions are made. While participative decision-making structures at the, executive dominated, international political level aim at the inclusion of (underrepresented) societal interests in the political decision itself, these decisions formally still rest upon nation based legitimacy chains. Perfect representative structures are, therefore, less important. Likewise, national administration could rest on the respective democratic legal statutes. In contrast, participative structures in transnational administrative structures -as opposed to the administration of international agreements -theoretically have to 'shoulder the democratic burden alone'. 72 The same is true for standardi- 68 See ISO/IEC Guide 2: 2004, definition of ' consensus': 'General agreement, characterized by the absence of sustained opposition to substantial issues by any important part of the concerned interests and by a process that involves seeking to take into account the views of all parties concerned and to reconcile any conflicting arguments. '  sation or private regulation in case of broader legal effects. At the same time, while political or administrative decisions lead to immediate legal consequences, standardisation and private (self-)regulation depend on an applying regime for broader legal effects beyond their signatories. Therefore, there is more room for external legitimacy control by the applying regime and the applying institution depending on its values, legal emphasis, legal effects (on those affected) and its need for complementing of compensating for legitimacy deficiencies.
This article does not aim at trying to define the ideal decision making procedure for a certain situation but instead looks at exactly this interaction between regimes thereby looking for possibilities for improving legitimacy.
In the following, the theoretical considerations concerning the legitimacy of public as well as private transnational rules shall be applied to the notion of 'international standards' in the TBT Agreement and exemplified with the WTO case of Tuna Dolphin II. The Tuna Dolphin II dispute 73 concerns two competing certification programmes on ' dolphin-safe' tuna products. The contested United States (US) provisions, 74 which are heavily based on a widely used transnational private label -originally developed by the US NGO Earth Island Institute (EII) in co-operation with three big tuna producers 75 -comprise stronger fishing restrictions than the international Agreement on the International Dolphin Conservation Programme (AIDCP) 76 by the Inter-American Tropical Tuna Commission (IATTC) with its own ' dolphin-safe' label. 77 . Consequently, the AIDCP label must not be used in the US market. 78

A. Standardisation of Technical Regulations and Labelling Requirements
The TBT Agreement aims at ensuring that 'technical regulations', including packaging, marking and labelling requirements, do not create unnecessary obstacles to international trade, Article 2.2 TBT. The TBT Agreement acknowledges the right of every WTO Member to establish for itself the objectives of its technical regulations. Article 2.2 TBT provides for a non-exclusive list of legitimate objectives, amongst them the prevention of deceptive practices and the protection of human health or safety, animal or plant life or health or the environment. Further legitimate objectives are market transparency, consumer protection and fair competition. 79 According to the Code of Good Practice in Annex 3 to the TBT, comparable requirements apply to the preparation, adoption and application of non-mandatory standards insofar as they also should not be unnecessarily trade restrictive and should be based upon relevant international standards. 80 The application of WTO law to national regulation in general and labels in particular that are concerned with production methods had long been disputed amongst scholars as well as WTO members. Arguments range from absolute inadmissibility of all sorts of 'non-product-related process and production methods' 73 In the late 1980s, the fishing practice of chasing, encircling and netting of dolphins with purse seine nets in the Eastern Tropical Pacific Ocean in order to get the beneath swimming tuna -which by then killed an estimated 7 million dolphins -was made public by the US NGO Earth Island Institute. This led to consumer boycotts, a privately organised ' dolphin-safe' labelling and monitoring scheme and legal interventions mainly in the US. regulation as an unequal treatment and therefore a discrimination of 'like products' in terms of Article III.4 General Agreement on Tariffs and Trade (GATT) to complete non-applicability of WTO law to voluntary labels as a means of market-based self-regulation, at least if there are no concrete economic compliance incentives by State law, 81 and therefore their lawfulness under WTO law. 82 In particular, the applicability of the TBT Agreement to labels concerning 'non-product-related process and production methods' had been dismissed by the majority of scholars with a view to the wording of Annex 1 to the TBT. 83 Point 1 of that Annex defines technical regulations as 'product characteristics or their related processes and production methods', whereas 'related' has been interpreted as traceable within the characteristics of the concrete product. 84 Under this opinion, this should also apply to labelling requirements despite the wording of sentence 2 of point 2 of the Annex, which does not contain the word 'related'.
In Tuna Dolphin II, the Panel -without further discussion -regarded US labelling requirements concerning ' dolphin-safe' fishing methods for tuna as product-related (as they apply to a product) and therefore as an issue of the TBT Agreement. 85 This finding remained uncontested before the Appellate Body. The fact that these processing and production methods aim at the protection of extra-territorial environmental goods was not discussed at all. Therefore, the end of this so-called 'PPM distinction' has already been claimed. 86 Another issue of this case was whether labelling requirements which gave (mandatory) requirements for a voluntary label have to be regarded as mandatory or as voluntary and, therefore, as technical requirements according to point 1 of Annex 1 to the TBT or as standards according to point 2 of that Annex. To date, the great majority of authors had categorised this type of regulation as voluntary, 87 whereas the Panel 88 classified them as mandatory and therefore as technical regulations. 89 The Appellate Body upheld this classification but argued that a (mandatory) requirement for a (voluntary) label does not in itself render that measure a technical regulation; 90 instead it depends on the circumstances of the case. Criteria mentioned in favour of a technical regulation were: whether the labelling requirements are laid down in legislation by State authorities, whether it contains specific enforcement mechanisms and whether the requirements are exclusive or encompassing for a certain type of label. 91 Consequently, according to the recent decision, the requirements of Article 2 TBT would also apply to (comprehensive or exclusive) requirements for voluntary eco-or fair-trade labels, which are concerned with production methods. Therefore, it is even more important to analyse the preconditions for private (transnational) labels to potentially constitute an 'international standard' within the scope of Article 2 TBT. 92

B. The Effects of 'International Standards' within the TBT Agreement
Articles 2.4 and 2.5 TBT aim at achieving international harmonisation of technical regulation through the recognition of international standards.
1. Justifying Standard, Article 2.5 TBT First of all, national measures in accordance to relevant international standards are '(rebuttably) presumed not to create an unnecessary obstacle to international trade', Article 2.5 TBT. In other words, those national measures that are in conformity with the relevant standard are (rebuttably) justified as not being arbitrary, discriminatory or unnecessarily protective and thus in conformity with the TBT Agreement. 93 2. Limiting Standard, Article 2.4 TBT Secondly, Article 2.4 TBT requires members to use international standards ' as a basis for their technical regulations except when such international standards (…) would be an ineffective or inappropriate means for the fulfilment of the legitimate objectives pursued (…)'. 'As a basis for' is not the same as ' conform to' but gives a certain margin of appreciation. 94 It has also always been highlighted that WTO members are free to define their own protection interests and level of protection as long as a legitimate objective is pursued, such as the protection of human health or safety, animal or plant life or health, or the environment, see Article 2.2 TBT. 95 For the definition of legitimate protection aims (above the relevant international standard), however, existing risks must be made plausible, for example through scientific data, see Article 2.2 TBT. In fact, in the case at issue, the US were required to submit scientific evidence that dolphins were adversely affected by being chased, encircled and netted with purse seine nets although they were released afterwards, without any dolphin killed or seriously injured, which is in line with the AIDCP labelling standard. 96 Also, it has been argued, that, in fact, there is an interrelation between what is regarded to be necessary according to Article 2.2 TBT and what is regarded to be the relevant international standard according to Article 2.4 TBT. 97 In fact, in the case at issue the Panel on the one hand accepted that the AIDCP labelling standard was inappropriate to fulfil the legitimate US dolphin protection objectives. Therefore, the US was not required to base their regulation on that standard. 98 On the other hand, the US regulation was declared not to be necessary by the Panel as it only partially achieves its aims with regard to dolphins outside the Eastern Tropical Pacific Ocean (ETP) and it would be equally suitable but less trade restrictive to allow the use of the AIDCP logo in addition to the official label which is in line with the EII label. 99 Although the Appellate Body in this particular case rejected the reasoning of the Panel, international standards are used as benchmarks for proportionality.
In conclusion, the selection of the appropriate international standard does matter as a benchmark for putting national regulation under justification duties.

C. 'International Standard'
The notion of an 'international standard' is concretised in the TBT Agreement. According to Annex 1, point 2 TBT a 'standard' is a '(d)ocument approved by a recognised body that provides, for common and repeated use, rules, guidelines and characteristics for products and related processes and production methods, with 92 Another issue in this regard was the question whether de facto practice of private market actors, for example, their de facto adherence to voluntary labelling requirements, renders these voluntary requirements de facto mandatory. See WT/DS381/R, at 7.166 ff. For detailed analysis, see Alessandra Arcuri, 'Back to the Future: US-Tuna II and the New Environment-Trade Debate' [2012] 2 European Journal of Risk Regulation 177, III. 93 For Scott (n 12) 325 f, this presumption is rather weak. Pauwelyn (n 84) 213 ff, goes even further and discusses whether the TBT Agreement in effect amounts to 'real' minimum standards which may not be undercut. 94 'As a basis for' means that an international standard is a 'principal constituent' or a 'fundamental principle' of a national regulation but not complete identity; see EC - which compliance is not mandatory. It may also include or deal exclusively with (…) labelling requirements as they apply to a product, process or production method'. The explanatory note by the Appellate Body to Annex 1, point 2 makes it clear that a standard has not necessarily to be based on consensus. A standard constitutes an international standard, when it is adopted by an international '(b)ody or system whose membership is open to the relevant bodies of at least all Members', Annex 1, point 4. 100 A 'standardizing body' is defined as a 'body that has recognised activities in standardization'. 101 Therefore, an 'international standard' has to be approved 'by an 'international standardising body', that is, a body that has recognized activities in standardization and whose membership is open to the relevant bodies of at least all Members.' 102 Annex 3 of the TBT -which demands that standards are not unnecessarily trade restrictive and are based upon relevant international standards -explains that its requirements are open for acceptance for all standardising bodies, be they governmental or non-governmental, local, national, regional or international. A non-governmental body is defined as one 'which has legal power to enforce a technical regulation', Annex 1 point 8. 103 The TBT Committee's Decision on Principles for the Development of International Standards ('Committee Decision') 104 adds the procedure-oriented principles of transparency, openness, impartiality and consensus, effectiveness, relevance and coherence, and of addressing the concerns of developing countries for the development of international standards. These principles clearly aim at achieving a broad and representative consensus, where ' all interested parties' 105 really have the chance of giving relevant input. 106 The Committee Decision has gained importance as the Appellate Body has regarded it as a 'subsequent agreement' within the meaning of Article 31(3)(a) of the Vienna Convention 107 regarding the interpretation of the TBT Agreement. 108 Therefore, its principles have been ascribed particular relevance concerning the questions of ' openness' and of 'recognized' activities in standardisation. 109 According to the Appellate Body, the concept of 'recognition' implies a factual (acknowledgement of the existence) and a normative (acknowledgement of the validity and legality) element. For the factual element, in order to be acknowledged to exist, information about the standardising activities of a body has to be disseminated according to the transparency requirements of the Committee Decision. The normative element implies adherence with the other principles of the Committee Decision. 110 Therefore, in order to be regarded as recognised, an 'international standard' now must pass the respective procedural legitimacy test. A wide de facto participation (of WTO members or standardising bodies) in the standardisation activities could be regarded as certain evidence of 'recognition', although there is no prerequisite that the standard is widely used. 111 Also, the development of a single standard could be enough to be regarded as recognised. 112 Moreover, an international standardising body must be ' open… on a non-discriminatory basis… at every stage of standards development '. 113 This again leads to the question of which kind of norm-setting arrangement is required in order to produce legally valid standards? Is it enough to have a system which is open to the relevant bodies of all WTO members, which would include respective international agreements? Is it necessary to have mixed systems which are open to all relevant national bodies but which also adhere to the inclusion of all interested parties and stakeholders and to procedural fairness requirements, which would aim at (well-organised) institutionalised standardisation? 114 Or would it also be possible to have purely private regulatory systems which are open to all interested parties and stakeholders from all WTO Members and adhere to the relevant fairness requirements? I would suggest determining the relevant standard on a case by case basis -thereby trying to find the most legitimate regime for the relevant issue, also with a view to complement for potential existing legitimacy deficiencies. 115

International Agreements as International Standards?
A valid public international law treaty can be adopted on the basis of the consent of the respective States without any further requirements concerning procedure or participation of stakeholders. Pauwelyn has described the different requirements for international agreements and standards as 'thin consent' vs. 'thick consensus. 116 In the case of Tuna Dolphin II, the US have argued that the definition of a legal term (here: ' dolphin-safe') within a (personally and substantially limited) international agreement (here: the AIDCP Agreement) cannot constitute an international standard. 117 This raises the question of international agreements as international standards. How many WTO members would have to form an agreement, to what extent would they have to represent which variety of interests and which procedures are necessary for an international agreement to count as an international standard? 118 The Appellate Body´s emphasis on the procedural legitimacy requirements as enshrined in the Committee Decision can be regarded as a considerable step beyond the hitherto applied narrow formal test according to the wording of the TBT Agreement as to whether a standard derives from an 'international body or system' which had already attracted some criticism. 119 Not only must its membership be ' open on a non-discriminatory basis to the relevant bodies of at least all Members' ' at every stage of standards development' 120 , but standards development must also 'take place transparently and with wide participation' 121 and 'must not privilege any particular interests' 122 Hereby, the Appellate Body also emphasised the participation of all interested parties which also aims at stakeholders 123 These requirements, in particular the finding that the development of a single standard could be enough, do not rule out international agreements but seem to put them under the same scrutiny as other standards. 124

Purely Private Regulation as International Standards?
Clearly, a private organisation such as the ISO can be an international standardisation body in the terms of the TBT Agreement. Such an organisation can be composed not only of governmental or administrative personnel but also of private representatives. It should, however, be open on a non-discriminatory basis to the 'relevant bodies of at least all WTO members' (Section B of the Committee Decision). Would that rule out purely private regulation that is only set or negotiated by (transnational) private systems but not organised via national delegations, for example agreements between industry and environmental or consumer organisations? 125 The requirement of participation of national delegations had long been disputed between the US and the European Union (EU). The EU insisted upon delegations, arguing that the composition of the standardisation organisation was essential, and that only non-discriminatory access for all national bodies guaranteed the necessary international consensus, given that it is the national bodies which have to implement the Christian Joerges and Erst-Ulrich Petersmann (eds), Constitutionalism, Multilevel Trade Governance and International Economic Law (Hart Publishing 2011), 402 ff, who calls this 'private intergovernmentalism' and argues in favour of the ISO as the relevant standardisation organisation. 115 Scott (n 12) 331 ff, argues that the respective procedural and legitimacy requirements could be part of the ' appropriateness'-test. 116 See Pauwelyn and Wessel and Wouters (n 19) 124 The assumption that a standard adopted by States could have less procedural legitimacy requirements than a standard by the 'standards community' and the respective suggestion to apply the principles of the Committee Decision (n 100) flexibly (see eg WT/ DS381/R, 7.654 ff) seem to be obsolete now. 125 See,however,Schepel (n 114) 405, who observes that the requirement of an international standard to be set up by an international body is not explicitly laid down in the TBT Agreement, which, therefore, enables bodies or systems not open to national bodies to set up international standards.
international standards or take them as a basis for their national norms. The US, in contrast, argued that it was not composition and membership that count, but a fair and open procedure, especially with regard to the inclusion of technical expertise, regardless its origin. According to this approach, standardisation has to consider safety and regulatory aspects, technical expertise and (global) market forces. 126 With regard to regulatory philosophy, this has been reconstructed as a conflict between a European '(…) integrated, formalistic and policy-driven' approach and an American 'pluralistic, sometimes fragmented, (…) market-driven' approach. 127 The principles of the Committee Decision have not solved the conflict as now delegations 128 and a fair including procedure are mentioned. The question concerning the best or most legitimate representation of the interests concerned remains. First, aspects of membership or participation and procedure cannot be regarded separately but rely on each other. Second, the conflict cannot reasonably be reduced to 'policydriven' delegations vs. 'market-driven' private actors as there could be a sub-optimal representation of interests in 'policy-driven' systems, whereas a private market-based negotiation system can surely be a forum of political debate and representation of competing interest. Third, even a unilateral business or user made standard can constitute a legitimate minimum standard as shown above. 129 Indeed it may be doubted that only a negotiation system that is primarily or exclusively composed of State delegations meets the requirements of an 'ideal' standard based on the principles mentioned above. This kind of organisation of international standardisation leads to a situation where the representatives of business and of other interests -such as safety or the environment -do not meet directly at the international level but only indirectly through members of national delegations. It has been criticised that those interests that have already been superseded at the level of the national standardisation bodies are not reflected anymore at the international level. 130 ISO, for example, has often been said to be dominated by industrial interests. 131 Here again, the 'legitimacy chain' within standardisation organisations could lead to the exclusion of certain legitimate interests, which ought to have standing in order to 'back' the WTO democratically.
Vice versa, it could be that a private standard that has been directly negotiated at the global level between the relevant interests -for example, industry, environmental protection and consumer interests -meets the above-mentioned principles of fair procedure and therefore offers greater legitimacy, in particular if it includes industry and civil society from developing countries. In such a case, the necessary consensus can be achieved more directly through participation of the different stakeholders at the relevant negotiation stage and the reflection of the different protection interests, local particularities, political and social realities and not least economic conditions in the standard is better ensured, 132 as it is said to happen in the regulatory system of the 'Forest Stewardship Council' or with regard to 'Fair Trade' labels. 133 The recognition of purely private standards in the framework of the TBT Agreement is getting the more pressing the more encompassing these private transnational certification programmes are becoming and the more national regulation refers to these programmes or to their contents. 134 In fact, not only standardisation organisations such as ISO have developed elaborate procedural requirements 135 but also private transnational certification programmes have begun to systematically adjust their structure and procedures to the above mentioned requirements of the Annex to the TBT Agreement in order to become recognised as 'international standards'. The 'Forest Stewardship Council', for example, has registered with the World Standards Services Network (WSSN). 136 Moreover, other private standards apart from the requirements for dolphin-safe tuna, such as the standards of the 'Forest Stewardship Council' or the 'Marine Stewardship Council' have been incorporated into (several) national regulatory systems in various ways. 137 Therefore, the relevance of private transnational negotiation systems which are open to stakeholders from all WTO members has already been discussed. 138 Pauwelyn argues in the same direction that non-traditional sources of global regulation such as private standards could at least play a role as benchmarks when deciding whether or not a concern is legitimate, a standard is appropriate and whether or not a country has acted in a nondiscriminatory manner. 139 Counter-arguments, such as private programmes being too independent from national regulation to be able to constitute a valid justification or limitation to State laws, 140 would lose weight with their growing recognition by State laws. 141 Also, the argument that acceptance of private transnational standards would lead to a multiplicity of standards (in clear contradiction to the goal of harmonisation of technical requirements), 142 is not convincing when it comes to certification programmes and labels because these do not aim at technical harmonisation as a precondition for the compatibility of products but at a competition of different values and protection aims and their persuasiveness. Besides, this problem is not specific to private transnational standards, but could as well arise with regard to different public standards. 143 Of course, the often mentioned problem of the legitimacy of privately set standards and the representation of interests of citizens, consumers and in particular of developing countries must be taken into consideration. 144 Therefore, their legitimacy for the legal effect in question requires scrutiny in each individual case. In this respect, the Appellate Body´s emphasis on the procedural legitimacy requirements as enshrined in the Committee Decision can be regarded as a considerable step forward also in terms of the recognition of purely private standards in case they fulfil these requirements.

IV. Applying Legitimacy Considerations: Tuna Dolphin II
In the following, I will use the example of Tuna Dolphin II to apply these legitimacy considerations and requirements to the interpretation of the notion of 'international standards' in the terms of the TBT Agreement. 145 There, particular emphasis is laid on the differentiation between the legal effects of Article 2.5 TBT and Article 2.4 TBT. The characteristics of ex post recognition and of labelling as (direct) market selfregulation are other relevant issues.

A. The AIDCP Label as International Standard?
In the case at hand, the WTO Panel had regarded the AIDCP standard as the relevant 'international standard' and in effect used it as a benchmark for putting the US regulation under justification duties. The Panel had argued that the AIDCP standard was the relevant 'international standard' in the terms of Article 2.4. TBT because it was ' open to the relevant bodies of at least all Members', because 'there were no limitations to or prohibitions of fishing in the agreement area, (…) (and) any country whose fishing fleet was operating in the ETP could have signed the AIDCP'. 146 In contrast, the Appellate Body rejected this approach. Essentially, the Appellate Body regarded the AIDCP as unduly favouring fishing interests over consumer or environmental conservation interests and indeed as not ' open on a non-discriminatory basis to relevant bodies of at least all WTO-members'. 147 A possible membership in the AIDCP was based on the requirements of either having a coastline bordering the ETP or having vessels fishing for tuna in the ETP or being otherwise invited to join the agreement. It seems quite obvious that this is not open on a non-discriminatory basis for all members of the WTO. The requirement of having or having to set up a fishing fleet in the ETP in order to participate in the agreement is of course a factual obstacle to non-neighbouring countries. The possibility for any State to be invited to join the agreement would qualify for openness only if the invitation was a pure formality. 148 Besides, the criterion of transparency in Section A of the Committee Decision requires transparency throughout the standardisation procedure, 149 not only that the standard was made available to the public and all interested parties after adoption, which the Panel had regarded as sufficient. 150 The view of the Appellate Body corresponds well with the legitimacy requirements set out above. When an international standard puts national protection aims under justification duties (Article 2.4 TBT) it is precisely these protection interests that are concerned. Here, the labelling requirements aim at the protection of dolphins and the protection of consumers from misleading statements and half-truth. Since the AIDCP standard is mainly based upon an agreement of the Nations fishing in the ETP, this arrangement by no means guarantees the representation of consumer and environmental interests, especially of those situated outside the fishing nations. This is confirmed by the fact that environmental NGOs call the label a ' death certificate' for dolphins, 151 and consumer organisations have classified it as not reliable and misleading. 152 Accordingly, that label has not gained considerable market acceptance. 153 The legitimacy of the AIDCP label to limit protection aims is further questioned by the fact that there is another, more protective, label which national measures could be based upon. 154 In contrast, no legitimacy considerations would strive against the use of the AIDCP in the context of Article 2.5 TBT, where national measures in accordance with relevant international standards are '(rebuttably) presumed not to create an unnecessary obstacle to international trade'. The agreement on the AIDCP monitoring and labelling scheme was ratified by, amongst others, Mexico, the US, Costa Rica, Ecuador, El Salvador, Guatemala, Honduras, Nicaragua, Panama, Peru, the Vanuatu and Venezuela. Bolivia, Colombia and the European Union are applying the agreement provisionally. The agreement is therefore based on a consensus between those countries whose vessels fish for tuna in the relevant area; which must regarded as reflecting the economic interests concerned sufficiently. 155 Thus, it could legitimately be used as an international 'minimum standard' that justifies equal or equivalent national legal labelling requirements according to Article 2.5 TBT, if such existed.

B. The EII Dolphin-Safe Label as an International Standard?
Whether or not the EII dolphin-safe label could constitute an international standard in the terms of Articles 2.4 and 2.5 TBT was not even touched upon in the decisions of the Panel or the Appellate Body.
In principle, the same rules set out above can be applied to a private rule, which can legitimately be adopted when it is based on the consent of the interests concerned. Protective regulation can only possibly interfere with contradicting economic interests that might wish a lower standard of protection. Thus, each (private or public) trans-or international regulation that does reflect the relevant economic interests according to the above mentioned principles related to 'group regulation' could be regarded as a legitimate justifying minimum standard in the sense of Article 2.5 TBT.
As the EII label only affects economic, here fishing, interests, it would have to be accepted by the tuna fishing industry, according to the principles of 'group regulation' as laid down above in order to be sufficiently legitimated. In this consent the interests of tuna producers from developing countries fishing in this area are of particular interest.
The private EII monitoring and labelling scheme was originally developed by a US NGO in co-operation only with three big US tuna producers. 156 By now, it is used in more than 50 countries, with more than 90% of the world's tuna companies 157 from more than 50 countries voluntarily participating, including such from developing countries fishing in the ETP, such as Costa Rica, Ecuador, Colombia, Peru and Venezuela. 158 The EII labelling scheme thus seems to be based on broad and representative transnational acceptance and adherence within the relevant industry.
Although there is no ex ante legitimacy through an inclusive procedurally correct drafting process, legitimacy could have gained ex post qua broad and representative recognition within the relevant line of industry.
Here, the US market has exerted considerable pressure in favour of the application of the higher certification requirements along the lines of the EII label, not only as a consequence of American consumer expectations but also because of the US legal framework. This, however, was not the case on other markets where the fishing industry uses the EII label truly voluntarily. Alternative labels exist (and have existed) yet, the EII label has been applied worldwide amongst the vast majority of producers including those from developing countries and a variety of markets. Provided that there are no important details to the contrary, the EII label could therefore be considered a legitimate business rule qua recognition within the group of tuna producers.
As it can be assumed that a broad and representative variety of corporations will safeguard their own economic group interest sufficiently, stronger protection by (WTO) law is not required. The EII label could therefore also be regarded as a legitimate international 'minimum standard' which is justified according to national requirements. The disregarding by a few corporations ('black sheep') in the consensus -here: mainly the Mexican fleet -can be justified by the very purpose of standardisation.
As a first consequence, the US labelling requirements which conform to the private EII requirements would have to be considered as being in accordance with WTO law, according to Article 2.5 TBT. 159 And even if a private labelling requirement was not based on a 90% adherence within a given industry, but only on a broad and representative consensus of those companies aiming at a certain market segment, for example, the fair trade market or the market for eco-products, it could constitute a justifying standard for a national regulation concerning that particular market segment. 160 Besides, labelling requirements only concerning a certain market segment can more legitimately be based upon market self-regulation by the relevant market actors or respective stakeholders than rules which impact upon whole societies or national economies.
Further legitimacy requirements would be needed in case the EII might be taken to the effect of Article 2.4 TBT. Indeed, the EII standard has not only been accepted by the tuna industry but is also supported by different environmental NGOs from different countries, such as Greenpeace (US) and Friends of the Earth (UK), and promoted by consumer organisations from different countries as reliable and not misleading. 161 Thus, it may be concluded, that the EII standard seems to reflect not only fishing interests but also dolphin and consumer protection interests at least to an extent that has been regarded as sufficient by the relevant stakeholders. The consideration of these interests -however limited to the US -is already reflected in the drawing procedure. In the meantime, it has gained ex post recognition by consumers and the relevant stakeholders. The EII standard could therefore far more legitimately be used as a benchmark for justification duties for stronger national regulation than the AIDCP standard.

V. Conclusion
This article is based on the assumption that there is a multiplicity of, competing or complementary, regimes at the global level which rest upon different strands of legitimacy as to why they should ' deserve recognition': parliamentary democracy and the respective legitimacy chains up to the international level, selfregulation, stakeholder representation, participation and deliberation, recognition and effectiveness. These different strands of legitimacy could gain more or less importance in different legal frames as alternative or supplementary sources of legitimacy depending on the relevant constituency and on those interests concerned by the regulation. Interactions and legal relevance of these regimes in a multilevel framework should be guided by aspects of mutual compensation of democratic deficiencies through recognition of other regimes or levels of regulation and by the principle of subsidiarity.
Accordingly, rules that only concern a certain market segment can more legitimately be adopted by the relevant market actors or respective stakeholders than rules which impact upon whole societies or national economies. Likewise, self-regulatory voluntary or minimum standards for private economic behaviour could far more easily, be regarded as legitimate than binding, absolute, or maximum standards with an impact on public interests.
To that effect, Tuna Dolphin II nicely demonstrates that a private CSR standard can be superior in terms of participative legitimacy and recognition than a standard that has been produced by an international organisation. Therefore, a transnational private CSR standard should well be suitable to constitute an international standard in the terms of Article 2.5 TBT and to justify measures in accordance with that standard. This justifying function could even be accomplished by a representative business standard. On the other hand, a public international standard should not be used as a limiting benchmark for national regulation, if there is a legitimate and more far reaching private standard available on which a national measure could be based. In order to gain more legitimacy the WTO should open itself to private standards including those that are not negotiated by national delegates but directly by the respective actors or stakeholders, which of course presupposes that the private standard meets the necessary legitimacy requirements and that the national measure is based thereupon.
The Appellate Body's emphasis on the procedural legitimacy requirements of the Committee Decision fosters this approach which, however, can be more fine-tuned with regard to the interests really concerned by a legal effect. More concretely, as the justifying effect of Article 2.5 TBT mainly interferes with economic interests, a related 'international standard' could also consist of a representative business standard. Alternatively a standard in the terms of Article 2.4 TBT that interferes with a democratic decision in favour of the protection of certain public interests, at least has to reflect these public interests at the transnational level in a legitimate way.
This approach would be a step into the direction of an ' orchestration' 162 of the plurality of regimes at the global level based on different strands of legitimacy with regard to the normative ideals of consensus of those interests mainly concerned by a regulation and of (common) welfare. CSR instruments could then gain not only political and ethical but also true legal relevance. They can serve to establish at least minimum hard law rules concerning the protection of social and environmental values that can, in the current global economic system, be defended against the otherwise prevailing interests of free trade.