I concur with my colleagues on several significant Issues of the case:

• The questions of jurisdiction, admissibility, control and waivers;

• the rejection of the expropriation (NAFTA Art. 1110) claim;

• the rejection of the "denial of administrative justice" claim;

• the rejection of the NAFTA governments' position that pursuant to Article 1102 Claimant needs to prove that the government had a direct Intention to harm the foreign investor because It is foreign is required for Art. 1102 and needs to be proven by claimant;

• the general view that the principle of legitimate expectation forms part, i.e. a subcategory, of the duty to afford fair and equitable treatment under Art. 1105 of the NAFTA. We also seem to concur on the general conditions for this claim - an expectation of the investor to be caused by and attributed to the government, backed-up by Investment relying on such expectation, requiring the legitimacy of the expectation In terms of the competency of the officials responsible for It and the procedure for Issuing It and the reasonableness of the Investor In relying on the expectation. I do, however, not concur with the application of this standard to the specific factual situation In light of the purpose, specific content and precedents of the legitimate expectations standard as It should be applied under Investment protection treaties based on recent relevant jurisprudence.

I have found the rejection of the national treatment (nondiscrimination) under Art. 1102 more difficult. Guardia, Thunderbird's competitor with comparable operations, continued to operate through success with procedural appeals ("amparos") before Mexican courts and through the lesser success of SEGOB in enforcing against him as compared to its - most effective - enforcement against Thunderbird. He was the domestic investor that was the "most comparable" and "best treated" by the Integral Mexican (administrative and judicial) system. That creates a presumption of discrimination which Mexico has to rebut by showing and proving "legitimate reasons" for such distinct, but more favourable treatment. But I have in the end accepted the view of my colleagues denying a breach of Art. 1102. However, I have been able to identify "discriminatory elements" in the greater energy, focus and effectiveness of the enforcement activities by the SEGOB against Thunderbird - which had arranged (or at least tried to arrange) a clearance of its activities as compared to the main and most successful Mexican competitor, Mr Guardia (who had always taken a non-cooperative approach. But I have not come in the particular circumstances to the conclusion that the much more favourable position enjoyed by Guardia in terms of de-facto practice and effectiveness of enforcement created a corresponding right for Thunderbird under Art. 1102 to continue its gam(bl)lng operations or to receive an equivalent amount of damages. I have been able to solve this dilemma by taking Into account such "discriminatory elements" in the enforcement of the Mexican gambling law In the context of the balancing that is, in my view, required between legitimate expectations of a foreign Investor and an equally legitimate public interest in preserving a large "regulatory space" in particular In the field of gambling regulation under Art. 1105 of the NAFTA.

I will discuss in this separate opinion both the normative scope and contours of the legitimate expectation concept as it should be construed under Art. 1105 of the NAFTA and their significance in the particular factual context. The facts that emerged in this arbitration and on its record provide, as always, not a complete picture of the events. But what has emerged can only be assessed, Including through presumptions and other rules of evidence, on the basis of the particular features of the legitimate expectations concept¹. Since the arbitration has, as often or always, not elucidated all relevant facts, one needs to rely on standard practice of rules of evidence, burden of proof and presumptions to determine when the claimant and when the respondent has to bear the respective burden. In addition, as recently explained again by the Methanex v US award, what is unknown but relevant has to be dealt with by Inference, i.e. by taking the "dots" that are available, drawing explanatory lines between them² and then determine what explanation can be Inferred by relying on burden of proof allocation, prima facie evidence and arbitral determination of the evidence. They need to be assessed not only from the lofty spheres of commercial arbitration law, but also with a real-life understanding of the "coal-face" of foreign Investment practices.

My disagreement is based on a different weight which needs to be accorded to this principle in the particular context of an Investment promotion and protection treaty which protects interests different from those Involved in an ordinary commercial relationship Involving two equal private parties. Commercial arbitration is a suitable mechanism for resolving the disputes of equal parties on equal footing and without need for the purpose of taking Into account the position of the weaker party; nor is there any policy purpose underlying commercial arbitration - such as to protect and promote Investment, enhance transparency and the "rule of law", create employment or enhance trade opportunities. In commercial arbitration, rules Including the caveat emptor and due diligence principle are deeply Ingrained in the culture, approaches and principles applied consciously or subconsciously by the tribunals. By contrast, International Investment law is aimed at promoting foreign Investment by providing effective protection to foreign Investors exposed to the political and regulatory risk of a foreign country in a situation of relative weakness³. The main principles underlying the NAFTA (preamble, Art. 102) as developed in the most recent and authoritative jurisprudence by arbitral tribunals require that in case of doubt, the risk of ambiguity of a governmental assurance is allocated rather to the government than to a foreign investor and that the government is held to high standards of transparency and responsibility for the clarity and consistency in its interaction with foreign Investors. If official communications cause, visibly and clearly, confusion or misunderstanding with the foreign investor, then the government is responsible for pro-actively clarifying its position. The government can not rely on its own ambiguous communications, which the foreign Investor could and did justifiably rely on, In order to later retract and reverse them- In particular In change of government situations.

Investors need to rely on the stability, clarity and predictability of the government's regulatory and administrative messages as they appear to the Investor when conveyed - and without escape from such commitments by ambiguity and obfuscation Inserted into the commitment Identified subsequently and with hindsight. This applies not less, but more with respect to smaller, entrepreneurial Investors who tend to be Inexperienced but provide the entrepreneurial impetus for Increased trade in services and Investment which NAFTA aims to encourage. Taking Into account the nature of the Investor is not formulation of a different standard, but of adjusting the application of the standard to the particular facts of a specific situation.

It is with these principles in mind that I have come to assess the -never absolutely clear and straightforward - factual background of the case and the presumptions and burden of legal justification and proof differently from my colleagues. They rather see the glass of the investor half empty, I rather see it half full. They imply a very high level of due diligence, of knowledge of local conditions and of government risk to be taken by the Investor. I rather see the government as responsible for providing a clear message and of sticking to the message once given and as reasonably understood by the Investor. They view the Investor as having a duty to be close to perfect in its dealings with the government, I consider the government to have a duty to be transparent and consistent, and as responsible for the message conveyed: i.e. how such conduct was reasonably understood by the Investor. They interpret the "Oficio" on its face value; I suggest it should be construed in light of its context, history and the objectively Identifiable common Intentions of both parties and as it was - reasonably - understood by the recipient, i.e. Thunderbird. They attach no importance to the combination of an official comfort letter followed by acceptance by the chief regulator of the Investor's operation to the end of the term of the government's term; I view the governments accepting conduct subsequent to the comfort letter as reinforcing and clarifying the Investor's understanding of the key message conveyed by the comfort letter. My colleagues see no discrimination whatsoever In the fact that the chief Mexican competitor goes on providing the same type of services the claimant offered while claimant loses all appeals and gets shut down; I see here by way of Inference, presumption and burden of proof on the government discriminatory elements of enforcement which reinforce the government's obligation to respect the messages the comfort letter and the subsequent accepting conduct have reasonably conveyed to the Investor.

I would have come to a quite modest obligation of the government of Mexico to pay a part of those Investment expenditures assumed by Thunderbird to the extent such costs can be reasonably and directly related to reliance of Thunderbird on positive assurances and accepting conduct by the competent Mexican authorities. I therefore find the claimant's argument and evidence on detrimental reliance more persuasive, but I then find the respondent's argument on compensation more convincing; In effect, my compensation assessment would have been at less than 0.5 % of the compensation claimed.

I also do not concur with the tribunal's decision to deviate from established practice not to allocate attorney costs to the losing Investor claimant. In my view, such a deviation would have required an In-depth and extensively reasoned justification.

It is likely to lead to highly divergent outcomes If the key NAFTA disciplines at Issue here - 1102, 1105 and 1110 - are not applied with a common understanding of the pertinent principles of legal Interpretation and application to a specific factual situation. Let me therefore highlight the key principles and methodological approaches I consider most relevant to define for the Interpretation and application to the factual situation,the relevant "context" and "purpose" of these principles:

First, the applicable law is - Art. 1131 - "this Agreement" (I.e. the whole NAFTA, not just Chapter XI) and "applicable rules of International law", guided by the authoritative article 31 of the Vienna Convention on the Law of Treaties("Vlenna Convention") where the elements of "good faith", "ordinary meaning", "context" and "object and purpose" are the main principles of treaty Interpretation. These principles - plus the Vienna Convention's reference to "subsequent conduct" - should also guide the Interpretation of the crucial "Oficio" of August 2000.

The Preamble of NAFTA emphasises as authoritative interpretation guidelines, namely the need for "clear rules", "predictable commercial framework for business planning and investment", "promotion of trade in., services" and "creation of new employment opportunities".

Art. 102 NAFTA: Highlights the purpose to "facilitate cross-border movement of..services", transparency, promotion of fair competition, increasing substantially investment opportunities⁴;

Art. 1115: To assure "equal treatment among Investors of the parties"

While the forms and procedures of International commercial arbitration are relied upon⁵, one needs, for the application of such rules, to bear In mind that their purpose is to govern the procedure, but not to inject substantive principles, rules and legal concepts used in International commercial arbitration Into the qualitatively different Investment disputes between a foreign Investor and a host state. International commercial arbitration assumes roughly equal parties engaging In sophisticated transnational commercial transactions. Investment arbitration is fundamentally different from International commercial arbitration. It governs the situation of a foreign Investor exposed to the sovereignty, the regulatory, administrative and other governmental powers of a state. The Investor is frequently if not mostly in a position of structural weakness, exacerbated often by Inexperience (in particular In the case of smaller, entrepreneurial investors). Investment arbitration therefore does not set up a system of resolving disputes between presumed equals as in commercial arbitration, but a system of protection of foreign investors that are by exposure to political risk, lack of familiarity with and integration Into, an alien political, social, cultural, commercial, institutional and legal system, at a disadvantage. Legal principles for and methodological approaches to examining the factual situation, habits, natural instincts and styles from commercial arbitration are therefore no suitable guideposts for Investment arbitration. The relevant legal texts and the factual situation at issue have therefore to be seen In the light of the close link between investment promotion - to get foreign businessmen to come with their capital and efforts into a new, alien and Inherently difficult and high-risk situation - and investment protection, I.e. the protection against governmental risk offered by Investment treaties to increase the attractiveness of the host state economy⁶.

Secondly, while public International law still provides the main principles (in particular Art. 31 of the Vienna Convention, which moreover is an expression of an International consensus on interpretative principles), one needs to bear In mind that Investment treaties such as the NAFTA, deals with a significantly different context from the one envisaged by traditional public International law: At its heart lies the right of a private actor to engage in an arbitral litigation against a (foreign) government over governmental conduct affecting the Investor. That is fundamentally different from traditional international public law, which is based on solving disputes between sovereign states and where private parties have no standing. Analogies from such Inter-state International law have therefore to be treated with caution; more appropriate for Investor-state arbitration are analogies with judicial review relating to governmental conduct - be it international judicial review (as carried out by the WTO dispute panels and Appellate Body, by the European- or Inter-American Human Rights Courts or the European Court of Justice) or national administrative courts judging the disputes of Individual citizens' over alleged abuse by public bodies of their governmental powers⁷. In all those situations, at issue is the abuse of governmental power towards a private party that did and could legitimately trust in governmental assurances it received; in commercial arbitration on the other hand it is rather a good-faith Interpretation of contractual provisions that is at stake. Abuse of governmental powers is not an Issue in commercial arbitration, but it is at the core of the good-governance standards embodied in Investment protection treaties. The issue is to keep a government from abusing its role as sovereign and regulator after having made commitments of a more formal character (contracts and licenses) or of a less formal character (I.e. the assurances by explicit communication or by meaningful conduct that form the basis of the legitimate expectations principle under Art. 1105 of the NAFTA).

The disappointment of legitimate expectations must be sufficiently serious and material. Otherwise, any minor misconduct by a public official could go to the jurisdiction of a treaty tribunal. Their function is not to act as a general-recourse administrative law tribunal. The introduction of direct investor-state arbitration ("arbitration without privity; "transnational arbitration") since the late 1980, resulted in a "discontinuity" which is not as yet fully appreciated and requires attention In cases such as this one. In former times, Investment treaties provided for an Intergovernmental arbitration process only; governments therefore had to "sponsor" private claims. Such governmental sponsorship provided an Important "filter" for screening claims and for avoiding that Investment treaties were used for a multitude of claims that did not justify the machinery of an international treaty to come into play: The risk of opened "floodgates" and the spectre of treaty-based procedures for a single instance of misconduct⁸ of an Individual official. Modern treaties with direct investor-state arbitration rights no longer have such In-built "filters". The construction of key legal terms must therefore provide sufficient filtering so that the treaty is only available to material, substantive and serious breaches and not for the every-day grievances arising from an individual's interaction with the machinery of government. Cases of administrative misconduct which are not serious enough, in terms of materiality of a breach, amount of damage or lack of instant remedy,⁹ do not t justify triggering the operation of the heavy and costly treaty machinery under Chapter XI.

Finally, I wish to highlight the need to pay attention and respect to the consolidating jurisprudence coalescing out of pertinent decisions of other authoritative arbitral tribunals, in particularly the more recent decisions applying the NAFTA and International Investment treaties which have a similar methodology, procedure and substantive content to NAFTA Chapter XI. While there is no formal rule of precedent in international law, such awards and their reasoning form part of an emerging International Investment law jurisprudence¹⁰. This is again a significant difference from commercial arbitration where there is little authoritative and persuasive precedent, largely because the awards are exclusively formulated for the private parties and because they are generally not publicly available. Investment treaty tribunals should therefore place themselves in the centre of emerging International Investment law rather than at or beyond the margin:

"To place one decision In a long tradition of similar decisions give the entire tradition of consistency,an "Integrity" that is a central feature of law as such"¹¹.

While Individual arbitral awards by themselves do not as yet constitute a binding precedent¹², a consistent line of reasoning developing a principle and a particular Interpretation of specific treaty obligations should be respected; if an authoritative jurisprudence evolves, it will acquire the character of customary International law and must be respected. A deviation from well and firmly established jurisprudence requires an extensively reasoned justification. This approach will help to avoid the wide divergences that characterise some Investment arbitral awards - not subject to a common and unifying appeals' authority. Otherwise, there is the risk of discrediting the health of the system of International Investment arbitration which has been set up as one of the major new tools in Improving good governance in the global economy¹³. But it also is also mandated by the reference to applicable rules of International Law (Art. 1131 NAFTA) and thereby Art. 38 of the Statute of the International Court of Justice: An increasingly continuous, uncontested and consistent modern arbitral jurisprudence is part of the authoritative source of International law embodied in "judicial decisions" (Art. 38 (1) (d)) and will develop, with an even greater legally binding effect, into "International custom (Art. 38 (1) (b)), in particular as an arbitral jurisprudence defines in a contemporary treaty and factual context the "general principles of law" (Art. 38 (1)(d).

This case presents neither solely a legal or solely a factual issue, but a situation where the factual situation has to be closely scrutinised, but always with the view of the legal rules and principles that are applicable. The application of a legal rule to a factual situation always Involves a certain feed-back between the way the legal rule is defined and the way the factual situation is viewed. The adjudicator's personal and cultural pre-understanding will Inevitably play a role as much as competent professionals will try to minimise and make transparent such a pre-disposition.

At issue here is a foreign Investment in gambling or, as industry advocates would now call it, "gaming". This industry has a bad press in many cultures and is not accepted in several religions, Including as I understand Canon law, fundamentalist Protestant attitudes and Islamic law. The negative view towards gambling businesses may be the reason underlying the change in attitude that took place between the outgoing Mexican PRI and the Incoming PAN government in 2001. It has also been relied upon by respondent in order to colour the case¹⁴. The former PRI government considered more extensive legalisation of gambling to create employment and re-attract Mexican demand for such services in Las Vegas; the latter seems to have been, to some extent, more closely attached to an attitude negative towards gambling. As far as this dispute is concerned, I have advocated a fully neutral and professional approach, without Inherent bias for or against this particular Industry. There has been no evidence in this dispute about any of the negative effects of gambling often alleged to accompany such entertainment services - crime, prostitution, moneylaundering or similar undesirable by-products justifying an extra-rigorous approach. That liberalisation was in the relevant period under consideration In Mexico is also indicated by a study commissioned by the Mexican Congress on the implications of liberalisation¹⁵ Under the WTO/GATS Schedules of Specific Commitments, certain countries have made specific commitments to liberalize gambling services, namely by offering non-discriminatory treatment to foreign gambling services establishing themselves on the domestic market; recent WTO panel and Appeals Body cases have treated gambling as any normal entertainment services industry as has the European Court of Justice¹⁶. This indicates that gambling services, in particular If not typically accompanied by criminal by-products, have to be treated as a fully legitimate Investment.

There is therefore no general or compelling reason to approach gambling Investment with a negative attitude compared to other types of Investment. All evidence in this particular case points towards operations of computer-programmed slot machines with a "skill" stop button; they have a certain attraction and entertainment value for many people. But from the record of this case and from personal observation of such facilities one now finds throughout the world, it seems not possible to either win or lose large amounts of money in a normal period of time spent. Such video-arcades with gambling machines can not be compared to the high-stakes traditional casinos. None of the criticism of "high gambling" applies here: There is no suggestion of children left destitute because of fathers' gambling, of family assets dispersed, but a rather mundane activity of putting coins repeatedly into multi-coloured slot machines. As a result, Thunderbird's operation should be viewed and treated as a normal operation of entertainment services, without any In-built bias against them and in favour of governmental closure.

The same applies to the corruption hint insinuated by respondent in its submission on the "success fee" paid to Aspe & Arroyo, two of Thunderbird's lawyer-lobbyists, for negotiating the "oficio" of August 15, 2000. Such Insinuations are now frequently employed by both claimant Investors and respondent governments. They should be disregarded - explicitly and implicitly, except if properly and explicitly submitted to the tribunal, substantiated with a specific allegation of corruption and subject to proper legal and factual debate for the tribunal. That is simply the Implication of the "fair hearing" principle. In contrast to, for example, the WTO dispute system and other international adjudicatory bodies, there is in current Investment arbitration only one level of fact-finding. If a tribunal should be influenced by Insinuations, there is no appeal instance (at present) in the NAFTA arbitral system which can correct a factual finding or assumption that has a bearing on the ultimate award. It is therefore particularly Important for a tribunal not to get Influenced, directly or indirectly, by "Insinuations" meant to colour and Influence the arbitrators' perception and activate a conscious or subconscious bias, but to make the decision purely on grounds that have been subject to a full and fair hearing by both parties. Cards should be placed, "face up", on the table rather than be waved around, with hints and suggestions. If the Mexican government had wanted to prove bribery It had the opportunity both to raise it and to try to prove it by providing its officials Involved in the transaction for cross-examination; but it chose not to produce them.

At issue in dispute - and the main area where I disagree with my colleagues - is whether the conduct of the Government, i.e. SEGOB, the federal gambling directorate, has Individually or In its aggregate, created a legitimate expectation for Thunderbird that it could carry out legally its business of computer-driven slot machines Involving some measure of skill and human Intervention. In this context, the government's duty to avoid ambiguity towards foreign Investors, to send clear messages and to pro-actively correct any misperception manifestly created, to take into account the Investor's need for predictability of government conduct and key attitudes is engaged, also its obligation to take its prior assurances into account when "closing" the facilities. It is not sufficient that Thunderbird had an "expectation", and that this expectation contributed in a significant way to its readiness to commit risk capital and effort, but the expectation must also have been "legitimate", i.e. it must have been created by government officials in an official way (i.e. attributable to the government of Mexico), they must have been competent (or at least appeared, credibly, to be competent) for the trust-inspiring action. The procedure for Issuing the assurance ("comfort") letter must have been legitimate and it must have been "reasonable" for Thunderbird to rely on that letter¹⁷.

The following are the key distinct factors on which the determination of "legitimate expectation" under Art. 1105 of the NAFTA must rest:

• First: The letter " oficio" or "criterio" of August 2000, to be read in conjunction with the following: (I) the request ("solicitud") by claimant; (¡i) the prior, prolonged, Informal and preparatory discussions with the competent government officials who later commended the Investor's "cooperative" approach In contrast to the Mexican competitors'confrontational approach and who encouraged Thunderbird to pursue this cooperative approach; and (III) the subsequent legal advice by its legal adviser In September 2000 (I do not concur with the majority's reasoning: paras. 158-163);

• Second: The accepting conduct of SEGOB subsequent to the issuance of the "oficio " in August 2000 which did not raise any questions, did not require any further Information, did not inspect or review the operations and which tolerated and did not interfere in Thunderbird's operation until a new (more antigambling minded) government and a new SEGOB director (on the uncontested facts available in the record more anti-Thunderbird-minded) came into office about six months later;

• Third: The conduct of SEGOB under its new director which targeted less the long-established slot machine operations at various locations of Mr Guardia, and certainly had no success (if there was a serious effort) in factually preventing them from operating but which targeted with priority and soonest after taking office Thunderbird's operation, though (or perhaps because) Thunderbird had gone the legal way and obtained an "interpretative assurance" which did give green light to Thunderbird, or at least seemed to them to give such green light. (The majority award - paras 174-179 - has no problem with both the relative enforcement intensity and its ultimate success against claimant but not against the Mexican competitor and does not deal with the question If the issuing of the "oficio" has anything to do with the later enforcement and closure)¹⁸.

It is in the combination of these three inter-related and consecutive measures of SEGOB - solicitud, oficio and subsequent conduct - that I find a breach of legitimate expectations under NAFTA Art. 1105 in contrast to the tribunal which finds the "Oficio" of August 2000 not clear enough and tainted by insufficient disclosure, attaches no significance to either the preparatory discussions of Thunderbird with SEGOB where a cooperative approach was encouraged,nor to the subsequent accepting toleration of Thunderbird's slot machine operations for over six months and which finds no elements of discriminatory treatment in the enforcement intensity, focus and effectiveness of SEGOB as between Thunderbird (rapidly closed after the new government and then the new SEGOB director took office) and Guardia - who continues, it seems, to this day, winning Injunctive relief ("amparo") and maintaining at least several of his long-standing slot machine operations.

To understand my different application of the principle of "legitimate expectation" under Art. 1105 of the NAFTA to the factual situation It is necessary to understand its background and scope which is far from simple or un-ambiguous.

First the doctrinal structure: "Legitimate expectation" is not explicitly mentioned in Art. 1105 nor In other similar Investment treaties. It is, however, considered to be part of the "good faith" principle which is a guiding principle (also a general principle of International law)¹⁹ for applying the "fair and equitable treatment" standard in Art. 1105, a standard that is repeated, more or less identically, in most of the other over 2500 Investment treaties in force at present²⁰. In the current dispute both parties (and the tribunal) assume the existence of such a standard under Art. 1105²¹. They can, correctly, rely on the recognition of "good faith" principle - either as a separate obligation or, arguably mainly, as a major Interpretative principle that is applied ancillary to a principal obligation (such as "fair and equitable treatment")²². "Good faith" is explicitly mentioned in Art. 31 of the Vienna Convention²³. This principle has been applied in intergovernmental relations to reinforce an obligation, to prevent a state to Invoke formal law against a claim when It has caused the other state to rely on the way it would exercise rights, and to deny a legal argument to a state when its previous conduct indicated it would not rely on such argument²⁴. To cite Derek Bowett in an authoritative statement²⁵:

"Representations.. may be made expressly or Impliedly where, upon a reasonable construction of a party's conduct, the conduct presupposed a certain state of act to exist. Assuming that another party to whom the statement is made acts to its detriment in reliance upon that statement or from that statement the party making the statement secures some advantage, the principle of good faith requires that the party adhere to its statement whether It be true or not. It is possible to construe the estoppel as resting upon a responsibility Incurred by the party making the statement for having created an appearance of act, or as a necessary assumption of the risk of another party acting upon the statement"

But such international inter-state rules are difficult to apply in the context of a foreign investor's reliance on host state assurances as to its law (i.e. a specific interpretation) or as to the way its authorities would proceed²⁶. But what can be used from International, inter-state law is the concept that "good faith" and "legitimate expectation" under Art. 1105 of the NAFTA trump the application of domestic law - such as Mexican gambling law as interpreted by the - then - new Mexican government²⁷. The good-faith and legitimate expectations principle control, for the relationship between the parties (e.g. Mexico and Thunderbird), the way the Mexican gambling law has to be interpreted²⁸. Governments can not, against a determination that under the International law-based "fair and equitable treatment" principle a legitimate expectation of a specific interpretation has emerged, Invoke a dominant contrary Interpretation under domestic law²⁹. The implication of this analysis is that the principle of "legitimate expectation" under Art. 1105 of the NAFTA overrides any dominant interpretation of applicable Mexican law on the legality of the operation at issue if SEGOB can be considered to have given -reasonably and legitimately - such an assurance. Mexico is not compelled by the treaty to change its law or the dominant interpretation of the law at a certain point of time; it is simply obliged to provide financial compensation If its officials have created an Investment-backed legitimate expectation with a specific investor that another, or earlier, Interpretation would prevail.

The principle of protection of "legitimate expectation³⁰" or, In common law, estoppel³¹, has also been applied in comparative contract law, mainly to deny formal rights Invoked by a party if such Invocation contradicts previous statements and conduct that made the other party trust in the particular expectation so created³². But contract law -presuming the existence of two equal parties in a commercial contract - is less relevant than comparative public law with respect to the judicial review of governmental conduct. For example, in its well-established jurisprudence, the European Court of Justice held "legitimate expectations" to be a key principle of the relation between state and individuals³³. The principle requires public authorities (including the European Commission) to respect legitimate expectations it has created with individuals, in particular if such expectations have become the basis for investment³⁴. In ECJ jurisprudence, the public authority can not lightly reverse course once it has created such investment-backed legitimate expectations, but has to take its prior conduct into account when planning to reverse its course with a detrimental effect on Individuals/ Investors. The principle has also been recognised in the jurisprudence of the European Court of Human Rights, here in particular to define the existence of legally protected "acquired rights"³⁵. European law does not prevent a public authority from reversing its course, but requires a balancing process where the strength of the Individual's interest is balanced against the need for flexibility in public policy³⁶:

"An expectation is then legitimate and ought to be protected If "taking a new and different course will amount to an abuse of power" - "Once the legitimacy of the expectation is established, the courts will balance the requirements of fairness against any overriding interest relied upon for the change of policy".³⁷

The principle of legitimate expectation is also recognised in several developed systems of administrative law³⁸. The common principles of the principal administrative law systems are in my view an Important point of reference for the interpretation of Investment treaties to the extent Investment treaty jurisprudence is not as yet firmly established. But its exact scope and Implication are not well established³⁹. There are contradictions between the principle that public administrations have to be utterly clear In their dealing with Individuals and to respect any legitimate expectations they have been responsible for and the concept that only confidence in un-ambiguous assurances by public authorities are protected⁴⁰. Legitimate expectations in EU law can be created by Informal statements, Including sufficiently precise oral representations and by government conduct, either by itself or In combination with written assurances⁴¹. Noteworthy is a reference in Schonberg's comprehensive study to the need for a more subjective approach that takes into account the experience and size of Investors - with greater protection for smaller and less experienced Investors⁴². The principle is recognised in the Spanish and Latin American civil law systems⁴³ and presumably forms - as part of the overall principle of "good faith" - part of Mexican civil and administrative law⁴⁴.

Legitimate expectation has also been recognised as an Important principle guiding the interpretation of other obligations in International economic law. A certain measure of recognition of this principle can be inferred from several WTO panel decisions:⁴⁵ "The protection of legitimate expectations of Members regarding the conditions of competition is a well-established GATT principle, which derives in part from Article XXIII, the basic dispute settlement provisions of GATT (and the WTO)."⁴⁶ In the main, the principle is here used to protect negotiated concessions from being undermined by conduct of member states contrary to the purpose and spirit of such concessions.⁴⁷

While this brief survey of general International, International economic (Including EU), comparative contract and comparative administrative law does not specify exactly the contours of this principle, it suggests that under developed systems of administrative law, a citizen - even more so an Investor - should be protected against unexpected and detrimental changes of policy if the Investor has carried out significant Investment with a reasonable, public-authority initiated assurance in the stability of such policy. Assurance on a particular Interpretation of often open-ended statute against an unexpected detrimental change of such Interpretation is in this context particularly relevant⁴⁸. Such protection is, however, not un-conditional and ever-lasting. It leads to a balancing process between the needs for flexible public policy and the legitimate reliance on In particular Investment-backed expectations. The consulted authorities are Indicative of contemporary state practice and the minimum standards of comparative national and international law. The "fair and equitable standard" can not be derived from subjective personal or cultural sentiments; it must be anchored in objective rules and principles reflecting, In an authoritative and universal or at least widespread way, the contemporary attitude of modern national and International economic law. The wide acceptance of the "legitimate expectations" principle therefore supports the concept that it is indeed part of "fair and equitable treatment" as owed by governments to foreign Investors under modern Investment treaties and under Art. 1105 of the NAFTA. It is before this International and comparative law background that one needs to make sense out of several recent Investment treaty awards which have applied the legitimate expectations principle, both under Art. 1105 of the NAFTA and the equivalent provisions of applicable bilateral Investment treaties. These awards - Metalclad v Mexico, Tecmed v. Mexico, Occidental v. Ecuador, Waste Management v Mexico II⁴⁹ and MTD v. Chile⁵⁰ - may not have explained the doctrinal background of the principle, its scope and contours specifically, but these authoritative precedents have contributed towards establishing the "legitimate expectation" as a sub-category of "fair and equitable treatment" in the for this dispute here most pertinent Investment treaties (Including NAFTA Chapter XI's Art. 1105)⁵¹.

While these cases for the first time appear to consider "legitimate expectation" as a definite subcategory of the "fair and equitable treatment" obligation, there are precursors. In most of them, legitimate expectation is used as a principle to specify the scope and content of primary legal obligations. In Revere Copper and Brass v OPIC, the tribunal Identified the "assurances given In good faith to such aliens as an Inducement to their making the Investments" as contributing to the creation of a protected right⁵². In ME Cement v Egypt, legitimate expectation was used to delineate future profitability of a license, for the purpose of calculating compensation⁵³. In Nagel v. Czech Republic (p. 33), the concept of a legitimate expectation was used to distinguish the existence of a protected acquired right from a mere hope or legally Irrelevant personal expectation⁵⁴. The tribunal's reasoning suggests that the less formal "personal communications", the less likely is the emergence of a legitimate expectation; this means that the greater the formality of an assurance, the greater its ability to trigger a legitimate expectation. That criterium is pertinent to the highly formalised "Oficio" issues in this case by SEGOB. In ADF v US, the tribunal discussed the claimant's expectation allegedly created by existing case law; but it denied the existence of a "legitimate expectation" because the expectation was not created by "any misleading representations made by authorised officials of the US federal government but rather, It appears probable, by legal advice received., from private counsel"⁵⁵. That case suggests, e contrario, that it is representations from authorised officials that provide the foundation for legitimate expectations if these representations become reasonably the basis for the Investor's commitment of capital⁵⁶. In Mobil Oil v Iran, legitimate expectation was used to calculate compensation for the break-down of negotiations for a termination agreement⁵⁷; it functions in this context as the protection of an interest that has not yet grown Into a full-fledged legally binding contract, but is still worthy of protection - similarly to the Shufeldt-case⁵⁸. The key feature of these cases is that a proto-contractual interest is protected, albeit, in terms of compensation, at a significantly different and lower level than would be available to a full-fledged, contractually validated legal interest. In SPP v Egypt, the tribunal held that "certain acts of Egyptian officials"... were "cloaked with the mantle of government authority and communicated as such to foreign investors who relied on them In making their Investment. Whether legal... or not these acts., created expectations protected by established principles of International law"⁵⁹.

As mentioned, the essential difference for the application of the "legitimate expectation" concept under Art. 1105 of the NAFTA and comparable Investment treaties from commercial contract law as applied In International commercial arbitration is that the two parties In Investment disputes are not In an equal position⁶³. If the parties are In an equal position, a much higher degree of due diligence is justified, as for example In Inter-state relations under conventional International law, In WTO law or In transnational commercial relations, as usually adjudicated In International commercial arbitration. Strong parties In an equal position can be expected to deploy more expertise and due diligence to minimise ambiguity In their dealings with each other. Nor can the same requirements as In national judicial review of administrative actions be applied as the foreign Investor is In a much more vulnerable, exposed position than a national citizen confronting his administration before national courts. Investment treaties throughout - witness the Preamble and Art. 102 of the NAFTA - are meant to compensate for this weaknesses of foreign Investors by a regime of Intensified protection. Such special protection for foreign Investors is required In order to encourage Investment, to compensate for the foreign Investors' structural handicap when entering a foreign society and to help governments enhance the quality of their governance systems. As Elihu Root said, In 1910, about the foreign Investor:

"He will naturally be at a disadvantage In litigation against citizens of the country. He is less familiar than they with the laws, the ways of doing business, the habits of thought and action, the method of procedure, the local customs and prejudices."⁶⁴

And recently Jan Paulsson:

"Whatever the rosy rhetoric about the equality of treatment of nationals and foreigners, the very fact of being foreign creates an inequality. The foreigner's obvious handicap - his lack of citizenship - isusually compounded by vulnerabilities with respect to many types of Influence: political, social, cultural."⁶⁵

All International Investment treaties aim at attracting foreign capital in a situation where the domestic Investor is in a much better position: It is as a rule more expert in dealing formally and Informally with the government apparatus; there are often hidden forms of collusion between administrators and local businessmen. What happens in such government-business relationships is usually not visible - it is a "black bo"x Into which foreign Investors - and arbitral tribunals - have great difficulties in penetrating⁶⁶. So if "fair competition" aimed at (Art. 101 (l)(b) NAFTA) and an "increase of substantial investment opportunities" (Art. 101 (l)(c) NAFTA) is to be achieved, there must be an extra attention to "clarity" and "predictability" for "business planning and investment" (NAFTA preamble). The protection of legitimate expectations standard thus says that such competitive opportunities as are protected under Art. 101(1) (b) NAFTA shall not be offset by measures which are in effect detrimental to the "business planning and Investment" of the Investor. NAFTA Chapter XI is to attract foreign Investors to the host state in spite of their greater exposure to the political risk, Including the risk of camouflaged domestic competitor-government alliances; the special protections of Chapter XI NAFTA serve as the principal Instrument for such an Investment promotion policy. As the World Bank Development Report for 2005, the authoritative policy Instrument on foreign Investment and economic development, formulates the relevant standard for government promises and administrative conduct⁶⁷:

"Can firms rely on them, with confidence, when making their Investment decisions?"

Enforcing rules making such promises effective is both In the long-term and comprehensive interest of the host state and of the Investor (at p. 179 ff). The use of the disciplines of Investment treaties - here the legitimate expectation principle under fair and equitable treatment - is a "potentially powerful tool to enhance the credibility of (governments') contract and policy commitments". As the World Bank Development Report puts it:

"Governments and firms can both benefit. Governments benefit from a commitment device that can address concerns from Investors, and thus help them attract more Investment at lower cost, and also reduce the risk of any later dispute becoming politicized. Firms benefit from reduced risks and a more reliable mechanism for protecting their rights if the relationship with the host government deteriorates."

These objectives of the NAFTA - both the general objective of enhancing the attractiveness of the host state for foreign Investors and the Instrumental tool of using greater transparency, clarity and predictability to enable better Investment planning- have therefore to guide the process of both defining the conditions of the "legitimate expectations" principle under Art. 1105 and of applying It to the particular facts of a specific situation⁶⁸. They are essentially different investment - " (p. 179). It views the use of investment treaties and their "disciplines" as a "potentially powerful tool to enhance the credibility of their contractual and policy commitments" from the approach to the "legitimate expectations" concept in commercial and contract law adjudicated through international commercial arbitration. The UNCTAD survey on the "Fair and Equitable Treatment" for investment treaties⁶⁹ therefore highlights that the "concept of transparency overlaps with fair and equitable treatment in at least two significant ways", one being that "the Investor will need to ascertain the pertinent rules concerning the state action; the degree of transparency in the regulatory environment will therefore affect the ability of the Investor to assess whether or not fair and equitable treatment has been made available..".

The most relevant NAFTA (and ICSID) awards have translated these authoritative objectives and Instruments provided by the NAFTA and similar Investment treaties into an emphasis on "transparency" and a concept of "legitimate expectation" that takes up, but further develops the meaning of this concept in conventional International, comparative contract, administrative and European and WTO law jurisprudence. One can observe over the last years a significant growth In the role and scope of the legitimate expectation principle, from an earlier function as a subsidiary Interpretative principle to reinforce a particular interpretative approach chosen, to its current role as a self-standing subcategory and independent basis for a claim under the "fair and equitable standard" as under Art. 1105 of the NAFTA. This is possibly related to the fact that it provides a more supple way of providing a remedy appropriate to the particular situation as compared to the more drastic determination and remedy Inherent in concept of regulatory expropriation. It is probably partly for these reasons that "legitimate expectation" has become for tribunals a preferred way of providing protection to claimants in situations where the tests for a "regulatory taking" appear too difficult, complex and too easily assailable for reliance on a measure of subjective judgment. objectives are the main "purpose" of the Treaty as relevant for interpretation under Art. 31(1) of the Vienna Convention.

In Maffezini v Spain, the tribunal linked the fair and equitable treatment obligation with transparency. The lack of transparency with a financial transaction carried out on government orders and detriment al to the Investor was considered the core of the breach of this obligation⁷⁰.

In CME⁷¹ v. Czech Republic⁷², the tribunal found a breach of the fair and equitable treatment discipline in the reversal of its previous position on the legal situation of CME. CME had, held the tribunal, a legitimate expectation that its legal position recognised by the Czech regulator would be maintained and not be changed, without bona fide purpose, to undermine its business, in particular favouring domestic investors. Such a change of the regulator's position on statutory interpretation created an opportunity for the squeeze-out of the investor by its local partner (and competitor). It is the "evisceration of arrangements in reliance upon which the foreign Investor was induced to Invest" by the new or subsequent government authorities which was at the core of the determination that there was a breach of the Investment treaty. As in the current case, the change of legal interpretation by the regulatory agency on which the foreign Investor relied allowed a domestic competitor, favoured de-facto, to flourish.

The Metalclad v. Mexico tribunal interpreted "transparency" to mean:

"that all relevant legal requirements for the purpose of initiating, completing and successfully operating investments made, or intended to be made, under the Agreement should be capable of being readily known to all affected Investors.. There should be no room for doubt or uncertainty on such matters. Once the authorities of the central government., become aware of any scope of misunderstanding or confusion in this connection, It is their duty to ensure that the correct position is promptly determined and clearly stated so that investors can proceed with all appropriate expedition In the confident belief that they are acting In accordance with all relevant laws".

The tribunal held that the Investor was entitled to rely on the representations of the federal officials (para 89) and the The Respondent " failed to ensure a transparent and predictable framework for Metalclad's business planning and Investment. The totality of these circumstances demonstrate a lack of orderly process., in relation to an Investor., acting In the expectation that it would be treated fairly and justly". The duty in Metalclad (as in the later MTD v Chile case) is not just a passive duty to ambiguous messages, but to pro-actively take clarificatory action when the government agency knows or should know that the Investor has misunderstood the relevant signalling from the government.

In Tecnicas Medioambientales (TecMed) v. Mexico (para 154), the tribunal held:

"Part of these expectations is the foreign Investor's assumption that the state receiving the Investment will act consistently, without any ambiguities, and transparently with the foreign Investor so that the Investor may know In advance (and thus plan its activities..) not only the rules or regulations., but also the policies pursued by such rules... and the administrative practices or guidelines that are relevant. "The foreign investor also expects the host state not to act in a contradictory manner; this means.. that the state will not arbitrarily reverse prior or pre-existing state-made decisions or approvals upon which the investor relied and on the strength of which it took on its commitments and planned and set in motion its.. operation". And it referred to the standard of a " reasonable and impartial man ".

In Occidental (OEPC) v. Ecuador, the tribunal examined the government's responses to queries by the Investor and found that the official response to such queries was a:

"wholly unsatisfactory and thoroughly vague answer"

and it considered that the "legal and business framework" did not meet the "requirement of stability and predictability" (paras 190-191).

In Waste Management v. Mexico II (para 98) the tribunal considered that Art. 1105 was breached by:

"a complete lack of transparency and candour In an administrative process. In applying this standard it is relevant that the treatment is in breach of representations made by the host state which were reasonably relied on by the claimant".⁷³

The recent CMS v Argentina (Merits) award confirms the principles developed in Metalclad and Tecmed v.Mexico. It draws a close link between the fair and equitable treatment standard, the government duty to provide clear and un-ambiguous signals to the Investor and the treaty objectives to promote investment⁷⁴. It concurs with my explanation of established jurisprudence according to which the breach of legitimate expectations created by specific assurances now constitutes a self-standing subcategory of the "fair and equitable treatment" standard under Art. 1105 of the NAFTA⁷⁵.

his is further confirmed in Eureko v Poland (2005); the award quotes (at para 235) with approval the Tecmed v Mexico (para 154) statement that:

"This provision of the Agreement, in light of the good faith principle established by International law, requires the Contracting Parties to provide to International Investments treatment that does not affect the basic expectations that were taken Into account by the foreign investor to make the Investment"

and determines that the "discriminatory conduct by the Polish Government is blunt violation of the expectations of the Parties in concluding the SPA.." (at para. 242)⁷⁶

These statements on the required clarity mirror established jurisprudence of the European Court of Justice. For example, in Opel Austria v Commission (1997, at para 124):

"According to the case-law, moreover, Community legislation must be certain and its application foreseeable by Individuals. The principle of legal certainty requires that every measure of the institutions having legal effects must be clear and precise and must be brought to the notice of the person concerned in such a way that he can ascertain exactly the time at which the measure comes into being and starts to have legal effects. That requirement of legal certainty must be observed all the more strictly in the case of a measure liable to have financial consequences in order that those concerned may know precisely the extent of the obligations which It imposes on them"⁷⁷

The Implications of the obligation to be clear and avoid ambiguity is that the government agency has to bear the risk of its own ambiguity. This allocation of the risk of ambiguity requires that the Investor did and could reasonably have confidence in the assurance, not as an ultra-perfect lawyer equipped with a hindsight vision facility, but as a reasonable businessman In the position of the Investor would do in the particular circumstances. "Hindsight, of course, is notoriously lucid"⁷⁸; but foresight lacks the sharpness of hindsight. Investors' lack clairvoyance and need to make rapid decisions on the basis of the way facts are and can reasonably be perceived at the time they become known - not the way they appear after years of litigation. Lord Mansfield, in 1761 said:

"The daily negotiations and property of merchants ought not to depend upon subtleties and niceties, but upon rules easily learned⁷⁹"

Lord Denning, in HTV v Price Commission, said that "a public authority which had led traders to rely on one interpretation of a statutory provision could only adopt another Interpretation If there was an overriding public interest to do so.⁸⁰"

The European Court of Justice has recently confirmed its jurisprudence on "legitimate expectation" and "legal certainty"⁸¹:

"With regard to the principle of legal certainty, this requires in particular that rules Involving negative consequences for Individuals should be clear and precise and their application predictable for those subject to them (see, to this effect, Case 325/85 Ireland v Commission [1987] ECR. 5041, Case C-143/93 Van Es Douane Agenten [1996] ECR 1-431, paragraph 27; and Case C-63/93 Duff and Others [1996] ECR 1-569, paragraph 20)"

The conclusion that the risk of ambiguity falls square on the shoulders of the assurance-issuing public authority, is reinforced by the traditional International law principle that the construction of a legal instrument in need of interpretation and with elements of ambiguity should be " in dubio contra proferentem ", i.e. that the drafter and the authority issuing a legally relevant statement has to bear the risks of ambiguity⁸².This rule is primarily concerned with interpretation of unilateral legal acts - such as the "interpretative assurance" given by SEGOB to Thunderbird. Related to the "contra proferentem rule" is the legal principle "Nemo audiatur proprlam turpitudinem allegans⁸³": It implies in the context of the transparency and clarity obligation on governments under NAFTA that a public authority which has evaded -as bureaucratic behaviour often does - the clarity of expression required by investors, then It can later not rely on the obfuscation It intentionally or negligently deployed to avoid the legal consequences of the legitimate expectation thus created and protected by Art. 1105 of the NAFTA. A change of interpretation of the law has to be reckoned with, but it becomes suspicious and "must be viewed with the greatest scepticism If their effect is to disadvantage a foreigner"⁸⁴.

The most relevant Unctad reports - authoritative UN surveys that can not be accused of Investor sympathy - tie transparency explicitly to the "fair and equitable treatment" discipline⁸⁵:

"This interpretation suggests that where an Investment treaty does not expressly provide for transparency, but does for fair and equitable treatment, then transparency is Implicitly included in the treaty (UNCTAD, 1999a, p. 34). Secondly, where a foreign Investor wishes to establish whether or not a particular State action is fair and equitable, as a practical matter, the Investor will need to ascertain the pertinent rules concerning the State action; the degree of transparency in the regulatory environment will therefore affect the ability of the Investor to assess whether or not fair and equitable treatment has been made available in any given case."

Hector Mairal, in his authoritative study of "legitimate expectation" (Actos proprios) for Latin American civil law systems, emphasises that ambiguity in an official representation can not free the public administration from the "legitimate expectations" effect if such ambiguity appears Intentional and contrived in order to leave to the administration two options while appearing to provide predictability to the Individual; if the administration is in a relationship with the individual which obliges it to be clear that imposes an extra duty of transparency and clarity upon It⁸⁶. Schonberg, In his study on comparative law on legitimate expectation has in this context pointed out correctly that smaller and less experienced Investors deserve greater protection than large and experienced companies⁸⁷.

Similarly, In two recent arbitral awards, an Interpretation of ambiguous treaty language in light of the treaty's Investment promotion objective was preferred over a restrictive interpretation that would have allocated the risk of ambiguity to the Investor⁸⁸. While these statements have been made in the context of treaty interpretation, they express a principle that is equally relevant to interpretation of official communications between government and investor, arguably even more so as different from a bilateral treaty, an official communication from government to a foreign Investor is not a bilateral agreement, but a unilateral communication solely under the responsibility of the issuing government agency.

These observations describe the legal contours of the principle of "legitimate expectations" under Art. 1105 NAFTA. But it is helpful to take an even closer look at the decision and underlying rationale of the arbitral tribunal in the very recent MTD v Chile case which should be considered the most relevant authoritative precedent:

In MTD, a Malaysian Investor planned an Investment in housing In Chile. The government made at a high political and administrative level positive noises; it assured the Investor of its welcome. It signed a formal Investment contract which, however, did not include any specific approvals for the Investment project at issue, but rather formalised the grant of foreign exchange and tax-stability related Investment guarantees; this Investment contract also clarified that it was not a substitute for specific zoning permits and other applicable authorisation requirements. The Investor was thus made to believe in the positive attitude of the government. Chile issued, at a senior governmental level, a formal endorsement of the project though the project could not be done under the regulatory framework as it stood at the time of such endorsement. The Investor was unaware of this, was not made aware of this and the minister responsible for the sector was not even Invited to the relevant meetings. Without the Investor's knowledge, other government authorities that were opposed to the project, took active steps to counter local support of the project and in the end ensured the project could not go ahead due to the lack of required zoning permits. The difference to Thunderbird is that the contradiction was there not simultaneous, but in the consecutive actions of government. The MTD tribunal did recognise that a "rigorous due diligence" by the Investor - Inexperienced and new to Chile - would probably have Identified the crucial obstacles to his Investment plan (para 117):

"A wise Investor would not have paid full price up-front for land valued on the assumption of the realization of the project" (para 242),

but:

"Chile has an obligation to act coherently and apply its policies consistently independent of how diligent an Investor is" (paras 165-166)

The tribunal considered such contradictory conduct by the competent government authorities to constitute a breach of the "legitimate expectation" principle - as sub-category of the duty to fair and equitable treatment. It found that the "Investment promotion" obligation was not just a prescription for passive behaviour or avoidance of prejudicial conduct, but had a " pro-active " meaning. Relying on Tecmed v. Mexico and Waste Management II, the tribunal found a breach of the legitimate expectations of the Investor In the governments failure to "act consistently", to be " free from ambiguity and totally transparent with the foreign Investor so that it may know beforehand any and all rules and regulations that will govern its Investments, as well as the goals of the relevant practices and directives" (para 114). The tribunal took account (also in terms of mitigating the compensation claim) of the "unwise business decisions or.. lack of diligence of the Investor" though "counsel for Chile in effect argued for the notion that the claimant was foolish to have relied upon representation of the government"⁸⁹. In MTD, the formal approval of the financial arrangements for the project and official signals about its desirability were in contradiction to urban policy and regulation; the fact that this contradiction was not conveyed to MTD before it committed its Investment, constituted the breach of the fair and equitable obligation; this situation is not that different from the positive "Oficio" of August 2000, confirmed by continuing acceptance of the operations by SEGOB, but then, and in contradiction to the earlier actions and positions taken by the former government, followed by the targeted and prioritised enforcement of the Mexican gambling law against Thunderbird by another, new, set of officials and political forces acquiring powers under the new government.

The MTD award supports the view that even If government assurances were ambiguous and an extra-careful Investor could have found this out, the government still owes a duty of consistency and protection of legitimate expectations to the foreign Investor. This is not a passive duty, but a pro-active duty (as in Metalclad v. Mexico) to ensure investors are not misled and are made to realise where the "true" directions of government policy for the issue at stake lie. This approach is in contrast with the "caveat emptor" and "due diligence" approach In commercial arbitration, but also, to a lesser extent though, to some statements in for example comparative administrative law where the risk of ambiguity in the governmental assurance is either assumed by the citizen, or at least balanced against a duty on the public agency to provide un-ambiguous statements. The MTD v Chile award thus reinforces a reading of the legitimate expectation principle that is distinct for Investment disputes. It acknowledges the structural weakness of the Investor - In MTD as in Thunderbird we have entrepreneurial Investors without extensive country experience - when confronted with a foreign country that wishes to attract such Investment. Such a pro-active duty to ensure the foreign Investor does not succumb to a visible misunderstanding is even more acute in cases where there are substantial indicators of "black bo"x collusion between administrative agencies and powerful domestic competitors. This conclusion has a bearing In particular on my point No. 2 - subsequent accepting conduct by SEGOB - for assuming the existence of a "legitimate expectation" by Thunderbird protected by Art. 1105 of the NAFTA.

t is with these interpretative guidelines that I will now examine the factual situation.

The issues are essentially if the "Oficio" ("official response" or "criterio" (translatable as "legal opinion" In this context) of 15 August 2000 and the subsequent conduct by SEGOB can be qualified as creating a "legitimate expectation" with Thunderbird that it could legally operate its slot machine facilities. The majority of the tribunal rejects this interpretation; I respectfully disagree. The relevant meaning of investor and government conduct and communications with each other can not only be determined from within the "four corners" of the legal documents, but must be appreciated with an approach that recognises realistically the practicalities of the foreign Investment process. The legal documentation has to be understood before the context in which the Investor-government interaction takes place. The interpretation of the key document - the official, authoritative, unilateral assurance in the format of the "Oficio" - needs to rely on International and comparative (civil law) methodology applicable to contractual, and where distinct, unilateral, documentation. That means that the text has to be assessed as it represents a "meeting of the minds" of both parties and in particular as it was, reasonably and for both parties manifestly, understood by the Investor to whom the "comfort letter" was addressed, taking Into account the history of their Interaction, the context, the purpose and the subsequent conduct of the parties⁹⁰.

In light of the differing opinions on the legal value and meaning of the "Oficio", one needs to bear In mind the burden of proof situation: Thunderbird has to prove that the Oficio conveyed to it, from the perspective of a reasonable foreign businessman In the gambling industry and in the specific context of the interaction between Thunderbird and SEGOB, the message that it could operate the software-driven video poker machines it imported. Mexico, on the other hand, has to prove that the Oficio was tainted by Insufficient, but mandatory disclosure by Thunderbird. This is a high threshold because, first, Mexico has to counter the presumption of the validity of official acts of government which respect for government requires; secondly, it has total control over all the documentation and witnesses - its own past and present SEGOB officials who alone can testify about what they knew and did not know. We therefore have to measure the evidence to see If Thunderbird has met this burden of proof, and, if so, Mexico has met its burden of proof.

It is not contested that Thunderbird was very keen to get a "negative clearance", "green light" or an In its sense, positive interpretative assurance from the government that its "skill" machines were not covered by the Mexican gambling law. It rejected the strategy of Mr Guardia who kept his profitable slot machine operations alive by using a sequence of mostly successful, mainly Injunctive appeals, and seems In whatever way have managed to defeat any attempt at effective enforcement. This is characteristic of the way a foreign Investor approaches a business the legality of which is not certain: A domestic investor, often with tacit allies in the administrative and judicial institutions, can afford more easily a blatantly Illegal conduct. A foreign investor will want more legal certainty - and in Thunderbird's case that seems also to have been urged by its venture Investors. It thus took the most "legal" course by asking SEGOB, and trying to persuade it, to assure it that its "skill" machines - in effect combining chance and skill - were not covered by the law.

Thunderbird's proposed interpretation - that machines Involving a substantial amount of skill (In addition to chance) could be considered legal - was not Implausible⁹¹. As all other players in the industry, It used the label of "skill machines" to highlight the Involvement of skill In order to make the point for the interpretation of the Mexican Gambling Act it advocated. Witness Watson reports that former Gobernación Secretary Labastida, had Informally supported the approach to get a comfort letter (Oficio) from the government that confirmed and repeated the "skill" argument already, reportedly, raised in a litigation In "northern Mexico"⁹². SEGOB does seem to have a certain administrative and interpretative discretion, both with respect to interpreting the law and with respect to directing enforcement efforts⁹³. The older and the more obsolete a law, as the 1940's Mexican Gambling Law, the more grows the need and the space for interpretation. Thunderbird had - this is not contested - received encouragement from a very senior Mexican politician - formerly the Minister in charge of Gobernación (Labastida) -to go forward. Liberalisation of the 1943 gambling law was considered in Mexico during the end of the PRI government in order to bring It in line with modern developments outside Mexico and to re-attract gambling income and employment that had moved to the Caribbean, Las Vegas and US Indian reservations⁹⁴. As Mexico's expert testified on the potential to liberalise the gambling regulation in Mexico:

"There was a great movement right before President Fox was elected" ( i.e. in 2000, the year the "Oficio" was issued).

There were no particular public order concerns with the type of coin-operated, computer-programmed video-slot machines. The then Mexican PRI government had the choice of either changing the law In a formal, time-consuming and politically costly process or to try out a more low-profile liberalisation by Introducing and then testing a reinterpretation of the law to relax its margins. Thunderbird's description of its machines as "not Involving chance" was factually - with the hindsight of this tribunal's expertise - incorrect, but it was a qualification for Interpretative purposes that was also used by other operators (Including Guardia), possibly suggested by Mexican experts. The issue of skill versus chance was well known to SEGOB from its confrontational interaction with Guardia and in several other litigations since 1998⁹⁵ as a suggestion to stretch the prohibition of the 1943 Gambling Law. As the report of the discussion with former Gobernacion Secretary Labastida indicated, the concept of liberalisation by "stealth" through re-interpretation of the "skill concept" is likely to have been common currency among senior officials and politicians with some knowledge of gambling regulation In Mexico. It was also the standard criterium operators and regulators used when desiring to relax more prohibitive gambling regulation⁹⁶.

It is in this light that we have to see the August 3, 2000 request ("solicitud") by Thunderbird for an authoritative opinion ("criterlo"). Thunderbird's "solicitud" makes it clear that such a "comfort letter" by SEGOB is desired to provide legal certainty for the Investment envisaged - and it is not contested that Thunderbird had such an intention and that SEGOB officials understood this perfectly well⁹⁷. In interpreting a unilateral declaration under International law, the relevant ICJ jurisprudence has emphasised the "significance of the intention behind the unilateral declaration made by a state"⁹⁸. It is therefore not a free-standing abstracted from its context text as it appears to a tribunal years after the event, but the Intention as it was conveyed and, moreover, as it was - reasonably - understood by the specific Investor In that specific situation that counts. Literal interpretation purely on an Isolated text is a traditional common law method (itself not applied strictly any longer and least in situations of ambiguous declarations); but it is not appropriate to our situation where, next to the NAFTA, Mexican law, and thereby also interpretation method, is applicable. The relevant ICJ jurisprudence deals mainly with unilateral declarations "erga omnes". Here we do not have a declaration erga omnes, but a governmental representation made in the context of a specific relationship. In that specific relationship, the reading of the interpretative assurance letter needs to be guided by what both parties Involved understood the purpose and factual background of the letter"⁹⁹. The "face" of the letter is largely gibberish If not read before the context, the parties' common Intention and the meaning that was intended to be conveyed and that was reasonably so understood by the addressee of the "Oficio". Even If there was a divergence - i.e. if SEGOB had a more modest Intention with the assurance letter, then the - reasonable - perception of the investor as the relevant specific addressee of the letter has to prevail. The reason is that it is the Investor that is to be encouraged by the assurance letter, the Investor that comes with capital and exposes its capital to government risk. It is therefore the Investor's confidence that is to be reinforced by the Oficio. To quote Reisman & Arsanjani:

".. the Inclination of an International tribunal to infer that a unilateral act has given rise to a binding obligation will probably be reinforced if the state making the declaration expects to receive clear benefits on the basis of the declaration"¹⁰⁰.

Different from the majority (see para 157), I see no lack of required disclosure: Thunderbird disclosed clearly that at issue were video slot machines and Identified them In a way that for a knowledgeable Regulator It was clear that these were video-gaming devices. I accept Mexico's suggestion that these were most likely refurbished gaming devices used by Thunderbird in the US. But this does not detract from the fact that for somebody with experience in the gambling Industry It was clear that these were video-gaming devices. The reference to BESTCO should have alerted the most sleepy gambling regulator that these were video gambling machines produced by one of the largest US producers of such devices. One can not assume, again, that the Mexican gambling regulator who according to its own statements had fought for years with Mexican businessman Guardia over video-gaming machines labelled as "skill" machines would not have been aware both of BESTCO as a major supplier of such machines. A short look at BESTCO's website and a google search confirm this¹⁰¹. The same applies to the reference to "SCI-Support Consultants" as an Identifiable manufacturer of video slot gambling machines, class III, deployed on US Indian reservations. The reference to BESTCO and SCI is therefore not misleading; It is a clearly Identifiable reference to video slot machines. It is consistent with the result of the cross-examination, namely that SCI (K. McDonald) probably refurbished Thunderbird and other operators'video slot machines previously used in US Indian reservations¹⁰².

Its letter otherwise needs to be seen not as a detailed factual description of the functioning of the machines (which It was not asked to provide), but as development of the legal argument as it had emerged in earlier litigation and already indicated in the discussions with ex-Gobernacion Minister Labastida. It made the legal argument that the machines were either only skill-based (para 3, which was overshooting reality), but it then referred in order to suggest as reason for legalisation, that "skills and ability is involved (para 6). This qualification for legal purposes is correct and it advances from the earlier reference that the machines were "only" skill-based. The issue was here to propose to SEGOB a legal qualification to help the liberalisation by "stealth" through a cautious interpretative strategy -that the machines were "skill" machines because they were used "in entertainment where skills and ability is Involved".¹⁰³ Thunderbird's "Solicitud" described the character of these machines in a light so as to make it easy to subsume them under the label "skill" machines¹⁰⁴ - a term that was used, for reasons of suggesting compatibility with the law, throughout the industry In Mexico¹⁰⁵. its statement - that "chance and wagering is not Involved" was Involved, is technically not correct. However, It should be seen not as a scientific analysis but as rather a legal-interpretative term suggesting (or repeating a suggestion Informally made by SEGOB as can be Inferred from Waton's testimony on the discussion with Labastida) how the law could be interpreted to allow such machines.

Virtually all games, indeed all human activity, Involve some element of skill and chance (Including say chess or football);¹⁰⁶ only some games - presumably the more mechanically and machine-based chance-oriented games - have in practice been prohibited in Mexico. As Mexico's principal expert Prof Rose put it:

"The second, element (sc. in gambling law), chance has caused the most problems in the courts. Part of the problem is that if every human activity is mixed skill and chance, the question is simply where you draw the line"¹⁰⁷. And: "In England, any skill at all takes a game out of the prohibited lottery category. California outlaws slot machines if any chance enters into the payoff, but then states that devices that are predominantly skill are legal".

So the legal-interpretative view that is put forward does not amount in my view to a lack of disclosure, but rather reflects the particular interpretative strategy, a strategy that Professor Rose describes in detail as the interpretation normally put forward to justify liberalisation¹⁰⁸. In his extensive study on gambling law - and I have to take this as authoritative as he has been put forward by Mexico as the principal authority on gambling law - he describes courts that recognised video poker as a game of skill and other courts which did not do so. But putting forward a legal view based on several respected US courts in Illinois and Pennsylvania - that video-poker is a game of skill or a game predominantly of skill¹⁰⁹ - can not constitute a deception. Thunderbird was not asked or expected to provide a dispassionate academic study on comparative regulatory approaches on video poker machines to the Mexican Regulator, but did suggest, and was expected to do so, its view on how the machines could and should be legally qualified. It naturally advocated an Interpretation that was in its favour rather than develop the reasoning for an opposing view. Nothing else is expected of professional advocacy, Including In interaction with an Industry Regulator.

The tribunal thus views as factual statements - and in this respect incorrect and lacking In required disclosure - what I consider is not a factual and technical statement, but a legal qualification of the machines made with the very intention to suggest an interpretation that would extend the boundaries of the 1943 law. I should add that I do not consider the "Solicitud" as the most technically perfected document. Thunderbird did not highlight the fact that "chance" was inevitably Involved in playing such machines (but as Prof Rose testified and we all know, chance is Involved in any activity), but it did reduce its original claim that "no chance was Involved " to, later In the Solicitud, that "skill" was Involved (i.e. that it was not exclusively a skills game). That some level of "skill" is Involved has not been disputed in the case; the tribunal has come so far as to suggest that a "considerable degree of chance" was Involved, without, however, being willing and feeling competent to quantify specifically the "degree of chance" (See para 136).

Thunderbird did not say these were refurbished Thunderbird videopoker machines; it did not say that Guardia was using the same type of machines. It did not Invite SEGOB to inspect the machines nor did it provide manuals. But SEGOB did request further Information when It wanted to - such as later In 2000 when a request for a similar "Oficio" was launched by Mr Gomez. That Thunderbird did not provide the Information Mexico now thinks they should have provided is, in my view not material. They were under no duty to do so. If SEGOB had felt in summer 2000 there was a need for more, it should have requested Thunderbird to provide whatever It considered relevant. If Mexico now raises them In arbitration, but did not raise them during the Informal and formal process of Solicitud and Oficio, this suggests that it changed its mind on Information requirements under the impact of a new government and the arbitration. Relevant non-disclosure -deception - would only then have been material if SEGOB had requested such Information and Thunderbird had in response provided false Information.

The signs were there - it must have been clear to anybody Involved -that Thunderbird was testing the waters with a cooperative approach to government for video-slot machines issuing prizes (or US Dollars, as a prize that eschewed offering Mexican pesos - legal tender - as prizes). From the prolonged period of Informal consultations - the claimant's factual assertion is not contested¹¹⁰ - the presumption arises that SEGOB officials knew what was at issue - and most probably suggested or at least approved of the very low-profile and discreet description. I can not agree that the tribunal "cannot rely on presumptions or Inferences, let alone speculation concerning that background" and interpret the 3 August 2000 Solicitud on its "face value" (para 150). The significance and the meaning of the Solicitud and the Oficio can only be understood when the itself undisputedly convoluted and ambiguous text is read before the background of the parties' interactions, their level of knowledge, their role and relationship (regulator vis-à-vis clearance seeking Investor) and interests. That is standard interpretation of contracts and related instruments, and in particular In civil-law countries such as Mexico¹¹¹. Not only does Oficio have to be interpreted on the basis of the parties common Intentions and the context of their Interaction, but also with the principle of good-faith which emphasises transparency, clarity and discourages the abuse of Intentional ambiguity to allow a government to first make the recipient and Investor believe one message and then turn around and claim It really had sent the opposite message. In addition, as we have - as mostly in litigation - a not completely verified factual situation, It is normal and necessary to use Inference and presumptions to derive from the evidence that is available what was most likely to have happened.

The "solicitud" did not come out of the blue; the normal way to go about such matters is to Informally sound out, negotiate and prepare in such an evidently very sensitive matter both the "solicitud" and the "oficio". This has to be the common-sense assessment of the situation. That would make eminent sense in terms of the "stealth liberalisation by interpretation at the law's margins" strategy that can be easily Identified. If the unlikely course of action had been that SEGOB was surprised by a request coming out of nowhere and then reacted a little bit confusedly, as must be Mexico's and the tribunal's understanding, then It was up to Mexico, using its control over SEGOB officials, to prove a course of event that would be strikingly different from the way interaction between an Investor and a regulating agency normally proceed.

Whatever the defects of the letter (and with hindsight and professional perfectionism a technically perfect "solicitud" separating a technical description from suggested legal qualification could have been written), I do not concur that by not providing manuals, complete technical specifications and not forcing SEGOB to inspect and test the machines physically,¹¹² Thunderbird failed with its disclosure duties in a way that any response would be Invalidated. SEGOB was not a group of widows and orphans to whom shoddy goods are deceptively sold at the door and which requires the special protection of the law: It was the chief gambling regulator In Mexico; it had battled with Mr Guardia about precisely this type of machines since at least 1998¹¹³; its legal battles with Guardia had been at the centre of SEGOB activities. According to the chief witness on this issue put forward by Mexico, the issue of the "skill versus chance machines" had been at the forefront of its litigation activity - Including five Supreme Court decisions. It is therefore not conceivable that when SEGOB received the Solicitud it did not think of the issue of using the "skill Involvement" for relaxing the Mexican gambling law prohibition. The machines itself - and that is essential -were identified in a way that allowed SEGOB to know they were videoslot machines used in the US for class 3 gambling.

If SEGOB had had the slightest doubts about the nature of the operations, it had the duty to investigate. The preparation of an administrative decision is not the responsibility of the applicant, who does what the government requires of him, but of the Regulator. It is not - as implicit in the majority's award - the obligation of private applicants to tell the national chief regulator how to run its business, but the public authority has to advise applicants what Information It requires. This is even more so as SEGOB had enough time; the time between the receipt by SEGOB of the Solicitud and the delivery of the Oficio is quite short; the claimant's narrative of several weeks (if not months) of Informal discussions between SEGOB and its lawyer-lobbyists Aspe & Arroyo has not been contested. It is also the way such business is conducted practically and in reality. One does not write out of the blue a request to a government agency, but the rules of the art of interaction with the regulator normally Involve an Informal period ("sounding-out") with the formal inputs and outputs (Solicitud and Oficio) only as the ultimate official documentation of an Informal process of consultation. Again, with full sensitivity of the controversial skill-chance issue created by years of litigation, with clear Indicators of a wish to liberalise gambling policy by interpretation rather than full-fledged legislative change, one has to expect the Regulator knew exactly what the issues were. It must have considered a physical inspection superfluous - much as later Mexico felt a physical inspection of the machines was not necessary for its principal expert, Prof Rose to develop his views later presented to the tribunal. Respondent can not now argue that its federal Gambling Regulator needed more Information which should have been provided by Thunderbird without being asked to do so when its NAFTA defense unit considered such Information for a foreign gambling law expert not necessary.

The consequence is that Mexico has not met the incumbent burden of proof that there was deception of SEGOB by insufficient disclosure. It should have brought the SEGOB officials involved to the tribunal. Since it did not do so, the Inference must be allowed that it considered that production of these key witnesses to the events would not have supported its argument of deception - nor its argument about the meaning conveyed with the Oficio.

SEGOB therefore knew full well what these machines were like and what issues they raised; the over two years of litigation occupying SEGOB's core attention focused on one issue: The question If slot machines with stop-functions (video-poker) could be exempted from the Gambling Law because of the publicly and in litigation alleged "skill" character¹¹⁴. I suggest that SEGOB therefore understood the issue at stake quite possibly much better than Thunderbird itself. The uncontested evidence on the interaction between Thunderbird and SEGOB officials suggests that the officials had encouraged Thunderbird to seek a clearance - rather than the confrontational strategy with Guardia which must have cost SEGOB a large amount of resources and loss of face. If SEGOB had had any doubts about the machines, they could have easily asked Thunderbird to provide more Information and inspect the machines - which were available in the offices of Baker McKenzie in Mexico City. The fact that they did not suggests that SEGOB had not the slightest problem In terms of awareness. The confrontation with Guardia and other Mexican operators must have provided to SEGOB all relevant technical understanding and legal sensitivity. They must have known how such machines functioned and how skill and chance played a role, both from a technical and legal/ regulatory perspective. It is not proper to consider a large country such as Mexico with a fully developed legal and administrative system, a 60+ years old gambling law and an experienced regulatory agency as acting, on the highest level of this specialised regulator, as uninformed, naive, inexperienced and not aware of the key issues relevant at the time in their line of business. We have to consider SEGOB as a competent regulator of its industry which knew what it was doing. The respect for government owed by international tribunals requires also respect for its officials and regulatory agencies - and with this respect, naturally, comes responsibility.

Nor did SEGOB have any doubts - or could have any doubts - that the Investor was asking for an assurance in the light of its interest to invest under conditions of greater legal certainty in a "grey area" of the law where the competent government agency's authoritative interpretation would make the decisive difference¹¹⁵. If SEGOB had had any doubts about either what machines were being envisaged, their technical character and the way they functioned, or the interpretative challenges they raised, they could have easily - and should have under the transparency and avoidance of ambiguity rule - requested Thunderbird to amend and back-up its "solicitud". That they did not do this indicates that SEGOB saw the letter - as Thunderbird intended -not as a technical description of the machines, but as a request to confirm the legal qualifications that Thunderbird, after Informal consultations, proposed or was recommended to propose. The same approach was practised by Mexico in the arbitration. Not only did SEGOB never feel it was necessary to inspect and test the machines, but respondent, in its defense, did exactly the same: It let its principal (but foreign) expert, Professor Rose, opine on the machines, their functionality and the legal Implications under Mexican law In great detail, but never felt it necessary to let him see, inspect, review and test the machines (which were in Mexico's hands)¹¹⁶:

Question: "Did counsel for Mexico indicate that they had it in their ability to provide a machine for your review If you could work out the logistics? Answer Rose: "I don't think we ever really got to that stage".

The Mexican approach throughout this case - be it SEGOB at the Oficio-stage, newly directed SEGOB In the prohibition phase or Mexico in the defense stage - has been that the functionality of the machine was self-evident, and no need for In-depth Inspection and examination was necessary¹¹⁷. If, after all the controversy on skill and chance, Mexico still felt it was not necessary to let their principal expert examine the machines physically and directly, then the conclusion to be drawn is that at no time was there any doubt with SEGOB about how the machines functioned and what legal issues they raised. The "lack of disclosure" by Thunderbird argument hence can go nowhere: Re-examining the machines in August 2000 - as during the subsequent NAFTA arbitration from 2002 to 2005 - would have been to "bring coal to Newcastle" or "owls to Athens". SEGOB and Mexico's counsel never thought it was necessary to examine the machines in detail - and the tribunal, I suggest, should not theorise on SEGOB's Ignorance as SEGOB and Mexico's counsel, then and now, act in a way that indicates that they have a perfect understanding of the machines at issue.

To sum up: Since I view the Solicitud as a proposal for a legal qualification of the machines as not being covered by the Mexican gambling law, I can not view the claimant's Solicitud as lacking In required disclosure of the technical nature of the machines. There can be no deception of SEGOB If SEGOB was or must have been aware of the nature of the machines, the legal issues raised, the precedential litigation and if the Solicitud in essence was conceived as and understood as a legal advocacy. The facts were evident and knowledge of them was shared by both parties; what was at issue was the legal qualification. Even Mexico's chief expert describes the moment in time when the Oficio was issued as " a great movement right before President Fox was elected " for liberalisation of the gambling law. And he equally provided the explanation for the subsequent reversal of SEGOB's position under the new PAN government:

"then some sort of scandal or political issue hit, and the government had to back away".

Nothing can be more persuasive for explaining Mexico's attempt to liberalise by stealth, through the "oficio" interpretation and its subsequent reversal (at the cost of the investor) than Professor Rose, Mexico's own chief expert and authority on comparative gambling law.

The formal letter that emerged is an extreme case of bureaucratic obfuscation: While protecting the "back" of the officials that signed and authorised that letter by ambiguous references, sometimes to machines where chance does not "Intervene" (there is hardly any game where chance does not at least have a minor role - so Mexico's principal gambling law expert Rose¹¹⁸), sometimes to machines which "predominantly" ("preponderante") Involve chance¹¹⁹, the main "operative" message of the letter is: Yes - go ahead with the machines if they are as you qualified them, but bear in mind that machines which involve "predominantly" chance are not allowed. A very rigorous analysis, done with hindsight of 4 years of national and International litigation and with the sophisticated expertise of my respected colleagues examining closely the Oficio word for word (paras. 159, 160), can plausibly come to the conclusion that the literal text of the letter did not give unambiguous clearance if chance was Involved in the operation of the "skill" machines¹²⁰. Chance is evidently Involved to a substantial extent, as it is in every respect of human activity, so my colleagues have some justification In suggesting this letter was not the un-ambiguous and clear assurance to Thunderbird that it could go ahead.

On the other hand, if the letter is read from the perspective of the addressee and a "reasonable businessman" of the relevant trade without the benefit of 4 years of litigation, and over twenty lawyers and experts poring over every word in the letter, a different message emerges. The letter does not say: Your machines (which SEGOB knew perfectly well) are not allowed nor did it say: We think your machines are the same as Guardia's machines (which SEGOB knew or should have known) and as you know they can not be operated in Mexico. It did give a positive signal - you can go ahead; its qualification ("as you described the machines") refers back to "legal" interpretation given by Thunderbird in its "solicitud" to the machines. Possibly, it plays intentionally with ambiguity in the "solicitud" which was meant to convey the legal qualification but could also be read as meaning the "factual" or "technical" description. Most importantly, and at first sight out of the blue, comes the reference that machines that are "predominantly" Involving chance are forbidden. The use of the "predominant" criterium Inevitably leads to the conclusion that if an operation that is "predominantly chance" is forbidden, then an operation that is "predominantly skill" is allowed. Predominant means "more than 50%". There is a zero-sum relationship between skill and chance. Something that is more skill, is less chance and vice versa. The "predominant" criterium is - as Professor Rose testified and wrote - the key issue around which legalisation and liberalisation of gambling regulation turns:

"There is the difference between whether It is a game of skill or a game of chance, so if it's predominantly skill, It is not gambling. If It is predominantly chance, then It is gambling".¹²¹

Using the "predominant" criterium is referring to a crucial gambling regulatory standard. A reference to "predominantly chance" as an indicator of prohibition is therefore automatically a reference to "predominantly skill" as an Indicator of legality. I have therefore trouble with the tribunal's rejection of the "e contrario" argument (para 160), in particular as Mexico's chief counsel (same as counsel for Mexico later, In the hearing, accepted quite explicitly the e contrarlo argument as Inevitable logic¹²²:

Question by President: "But does it address also the question predominantly, now the reverse, predominantly skilled?.. It says one thing, but does it also say the other thing.

Answer: You might interpret it as predominantly ability and skill and not betting

Question: So, you would say you can Interpret this?

Answer: Yes, sir

It is virtually impossible to determine if the machines Involve chance under or over 50%; at best, it depends on the level of player skill which, so respondent expert Prof Rose, increases in a "steep learning" curve, i.e. with a rapid increase once a serious effort at learning is made¹²³.

With the introduction of the criterium of "predominantly skill or chance", SEGOB shows the way how the boundaries of the Mexican gambling law's prohibition on games of "chance" can be relaxed. That is fully consistent with the report of witness Watson's conversation with ex-Gobernacion Secretary Labastida supporting the "appropriateness" of using the "skill Issue" from a "northern Mexican litigation" to relax the gambling prohibition. The "Oficio" can therefore be read as suggesting to Thunderbird that it should not qualify its machines as "only skill" (reflecting the label of "skill machines" used for presentational purposes), but as "predominantly skill". While "no chance at all" is a criterium that can not be met (by any game), with "predominantly skill" the door is open to discretionary assessment. A gambling Industry person can only hear when the term "predominantly skill" emerges the message: "Yes - allowed" - as Mexico's chief expert Prof Rose said in describing the Californian approach¹²⁴:

"devices that are predominantly skill are legal"

A dispassionate expert or a tribunal careful weighing up facts ex-post and after Intensive litigation may come to a more nuanced conclusion. That what is relevant for Interpreting the conveyed meaning and message by SEGOB to Thunderbird is not what a dispassionate expert or a meticulous tribunal would or should understand, but what the addressee of the message - the Thunderbird gambling industry investors and promoters -could reasonably understand at the time the message was conveyed.

SEGOB's and Thunderbird's interaction can not be construed on the sole basis of the text of the "Oficio" as would be read in isolation by sophisticated International lawyers, but they need to be read as the "people in the business" - the gambling regulator and gambling professionals - would read them. In proper methodology for construing contractual text and text of unilateral declarations addressed to investors as we have here, it is the "horizon" and perception of a reasonable person In the trade that counts. And here "predominantly skill" means - let us simply trust Mexico's chief expert in this matter: "Yes".

With this criterium, a large leeway of discretionary Interpretation is opened: Do video-slot machines running on software Involve skill at 10%, at 51%? There is no fully objective determination possible; player skill and experience determine the relative proportions of chance and skill. In capturing the main message conveyed by the text in its particular context, we need to acknowledge the desire by Thunderbird to get legal clarification for its Investment. That was perfectly known to SEGOB. We need also to appreciate that SEGOB knew and must have known all about the technical nature of the machines and the legal sensitivity, tested in many litigations and administrative procedures. The August 15, 2000 "criterio" ("oficio") has then to be seen as SEGOB giving a green light (at the end of a long tunnel darkened by ambiguity and obfuscation). The numerous reservations can be explained by the usual self-defensive strategies of bureaucracies¹²⁵. Some of the reservations - i.e. "predominantly" skill-involving versus "Involving no chance at all" - are contradictory. But the ultimate message for a reasonable businessman In that situation was the answer to his question: Can we operate these machines which you know?: Yes, you can, just be careful and note that you - we - have to present this as something that can be qualified as "predominantly" - but not exclusively - skill-involving. That explanation fits perfectly with Mexico's expert Rose's reference to the window of opportunity for relaxation of the rules that existed just in 2000 (before President Fox was elected) and closed rapidly thereafter. What counts for the legal assessment of the letter is not the text per se, but the way it could be and was likely to be understood by Thunderbird to whom the message was conveyed. It was how Thunderbird could, reasonably, have understood the response of SEGOB to its request - the reasonable perception of the addressee of the message.

Thunderbird was no "Fortune 100" multinational company with hundreds of lawyers and country analysts at its disposal. It is a small entrepreneurial company where entrepreneurial activism was not matched with commensurate expertise and caution. But NAFTA would lose its objective of mobilising Investment opportunities if its requirements were only suited to very large, expert, well resourced and suitably super-cautious companies. The vigour and dynamics of entrepreneurial drive would be lost; this is not compatible with the cited objectives both from Art. 102 and the Preamble of the NAFTA. If SEGOB had wanted to keep Thunderbird from operating its - clearly Identified - machines, it should have said so and it could have said so easily, clearly and unequivocally.

This conclusion, I suggest, is the one most consistent with real-life practices and expectations. It takes into account that a private Investor will rarely look at what looks like and is intended to be a positive response with the "rigorous due diligence" and the fine comb of an ultra-cautious litigation lawyer based on hindsight, but will look towards the essential message. It was: "You can go ahead - bear In mind: Such types of games in Mexico need to be presented as "predominantly skill-involving"". While a text-book approach would always require that official opinions be very clear, the messy reality of business life in most places and most times is that bureaucrats tend to use obfuscation for self-protective purposes in sensitive situations even if they want to be supportive. Disputes would not go to arbitration and investment treaties were not necessary if every investor would at any stage in its business manage to execute a legal transaction so that there were no doubts whatsoever over a government's intention. To the contrary, ambiguity is the name of the game in dealing with governments and the task of international investment protection comes into play not in the case of the perfectly executed and documented transaction, but in the imperfect one of real life.

It is here that the legal criteria Identified earlier for "legitimate expectations" need to be applied: The tribunal's majority relies on the ambiguity¹²⁶ and lack of clear, unconditional and un-reserved text of the letter. But if we apply the principle that the risk of ambiguity has to be allocated to the drafting government, that a government agency can not rely on Intentionally inserted obfuscation to extract itself from the key message the Investor relied upon and that the drafter and the public authority in a position of superiority over the foreign Investor has to be clear, unambiguous and consistent - then the positive message that a reasonable businessman could have taken from the "Criterio" of August 10, 2000 must prevail over the manifold reservations and contradictions my esteemed colleagues rely on. Similarly, based on the rules developed in particular In the Metalclad v Mexico and MTD v Chile cases, but also reflected in other precedents on the duty of governments' to provide pro-actively legal certainty to investors, one can conclude that if SEGOB did not want to accept Thunderbird's type of operation, It should have said so, clearly, and if it saw that Thunderbird did not get the message properly, It should have repeated the message and ensured it was clearly conveyed and understood¹²⁷.

The "Oficio" or "Criterio" is not private legal advice - the claimant did not need any more legal advice having contracted several respected lawyers and law firms already. It comes, as respondent concedes¹²⁸, with the presumption of being an official and authoritative act by the competent government agency. It comes with the full authority of government - on SEGOB letterhead, multiple official seals or stamps of the "Secretaria de Gobernacion" - the Mexican Interior Ministry. It is not a furtive note handed out secretly to Thunderbird to avoid the light of day, but it is formally copied to at least two senior Gobernacion officials; it presents itself as an official unilateral statement intended to have legal Implications. It is signed, every page is Initialled and it has reference to an official case Identification code¹²⁹. There is also a formal act of notarisation of the document. The more formal a communication by an administrative agency to an Individual in a specific case, the more likely it is to create a legitimate expectation; the threshold for Informal or general communications is much higher.

Formal acts of government have to be treated with full respect¹³⁰; it would not be respectful to treat a government's formal declaration as if It were the uninformed utterings of an Ignorant minor In need of protection against shady dealings. Thunderbird did not want or need a restatement of the letter of the law - It wanted, as was clear to the government, a statement if its "skill machines", Identified properly, could be operated in Mexico. It wanted an Interpretation - and with the "predominant criterium", It received one. We have to assume that SEGOB did mean what it said and was ready to provide "green light" to the Investor. The presumption is that such a formal legal advice, sought by an Investor, is valid, has an effective meaning, responds properly to the request and in its operative conclusion gives to the investor a clear response. Respondent bears the burden of proving that the "oficio/criterio" was emitted in an Improper procedure by officials acting manifestly outside their powers and that it did not convey the main message which was the reply to the main question of the investor: Can we operate our machines - the BESTCO and SCI machines which we (as the other operators) call "skill machines", in Mexico?

We also have to assume that it was intended to say something substantial on the request for "green light" by Thunderbird - rather than just a re-copying of the text of the law. A view that reduces the conveyed meaning of the letter to something close to zero, lacking a true substantive response to the "solicitud", does not do justice to accepted interpretation methodology for legal Instruments which include a legally significant unilateral statement such as contained in the Oficio. Legal Instruments formally emitted are in doubt to be interpreted for an "effet util". If they serve as a formal and official reply to a request for clarification of the law by a foreign Investor, then they have to be an effective response to the request. If it did so with so many reservations and ambiguity, then the government has to bear the risk for such ambiguity. There is a presumption - both In international and in comparative administrative law - of the legitimacy of official acts¹³¹. That is the risk that the government, as price for the due respect to official acts, has to bear.

That the "Oficio" gave green light was also the opinion of Thunderbird's Legal Adviser Mr Ruiz de Velasco of Baker McKenzie. While he reiterated the reservations of the "Oficio" - which lawyer does not equally try to protect his back when giving legal opinions, the operative conclusions, and this is what counts, he confirmed that Thunderbird could go ahead and operate its video skill machines. He may not have understood nor Mexican gambling law nor the functionalities of the machines; possibly, he did not appreciate the implication that the introduction of the criterium of "predominantly skill-involving" machines in the Oficio opened the interpretative door of the Mexican gambling law. But his opinion must be weighed primarily by its clear conclusion rather than by its lawyerly and self-protective reservations. While other sophisticated lawyers are competent to appreciate the self-protective legalese in legal opinions, in particular with hindsight ex-post, our Impression from the hearing was not that this applied to Thunderbird. Mr Ruiz Velasco got in cross-examination increasingly confused about disclosure as it should have been, as it was done, about the functionality of the machines and their legal implication In Mexico, but that was because he had little if any understanding or Interest in the technical and legal issues of the Mexican gambling law. Had he understood the Implication of the "predominantly skill or chance" criterium Introduced by the Oficio properly, then he would have been able to give a clearer legal opinion and represent this accordingly before the tribunal.

The "Oficio" was also within the competence of the government officials who signed it¹³². Interpretative and similar official assurances and representations must be "legitimate", i.e. they must be issued by competent officials and not, at least from the due-diligence horizon of the recipient, be against the law¹³³. SEGOB is the highest federal authority in Mexico for regulating the gambling business. Such authority Involves a competence to determine, for the purpose of the administration of SEGOB, the boundaries of the law. That Inevitably implies interpretation of the terms - even If such Interpretation was not legally binding In the way courts act and subject to judicial action. Since virtually all games involve some elements of chance and skill, It is a normal and legitimate activity for the principal national regulatory authority to determine (and to convey to an Investor) its own view of the precise line constituting that boundary, even more so as the underlying law, of the 1940s, was quite old and had not kept up with modern commercial and technological developments. International regulatory practices - on which Prof Rose testified for Mexico - had developed the "predominantly skill or chance" distinction; accordingly, it was perfectly appropriate for SEGOB to interpret the 1940s' Mexican Gambling Law In the light of such practices, in particular as there was a political idea of liberalising the Gambling law around at the time. Liberalising It at the margin - rather than seeking a wholesale legislative change - is often If not mostly used to introduce policy changes in a way that is faster, more efficient and more politically palatable. Ex-Gobernacion Minister Labastida, Mexico's chief expert Professor Rose and Thunderbird witness Watson all in effect concur that there was, in 2000, a window of opportunity for "stealth liberalisation" using the openness of the "skill" condition and SEGOB and Thunderbird exploited this window. The SEGOB officials therefore issued their "Oficio" well within their real and apparent competence and within the then emerging (but later reversed) official policy¹³⁴.

Thunderbird's view that the "Oficio" was giving formally (even If cautiously worded in bureaucratic language) green light to their operations was also reasonable. First, the machines and their mode of operation were well known to both parties. Second, Thunderbird had made clear to SEGOB that it considered the issuance of a comfort letter as significant to their operation, and also engaged on the path of cooperation with the government rather than the confrontational strategy applied by Mexican competitor Guardia. They might have been more cautious; they might have seen that the "Oficio" left many escape routes to SEGOB and was not an absolutely clear and unambiguous assurance. But they were not unreasonable in drawing comfort from what appeared in the context of their communication the positive attitude of the Oficio towards Thunderbird's machines and the confirmation of this positive message in the operative paragraph of their legal adviser's subsequent legal opinion letter.

To sum up: The expectation was created, by the competent officials in their normal conduct of affairs, with Thunderbird and it was also reasonable by Thunderbird under the circumstances to draw confidence from the Oficio. We do therefore have a "legitimate expectation" protected by Art. 1105 of the NAFTA. Thunderbird evidently understood the "Oficio" to give green light, but my analysis also suggests that it could reasonably and in the context of the regulatorgambling business interaction understand the operative message and the "predominant" criterium to mean that green light was given, and for the machines it had named, envisaged for its operations and ultimately deployed for operations. Mexico's case in the main rests on the "deception of SEGOB" argument, but as I have determined earlier, it did not meet the Incumbent burden of proof for deception.

In spite of my different view attributing effectiveness to the "Oficio", I might have become swayed by the eloquent arguments of my colleagues dissecting the Oficio in a painstaking way that the "Oficio" was just not enough to create a legitimate expectation that Mexico's SEGOB was ready to use its powers to tolerate Thunderbird's operations. But the "comforting" messages coming from SEGOB to Thunderbird did neither start nor stop with the "Oficio" of August 15, 2000. As is recognised in "legitimate expectations" jurisprudence,¹³⁵ conduct, as the "consistent and prolonged treatment of a person in a particular way, can create a reasonable expectation that the treatment will be continued until further notice". Given the difficulty of enforcing Mexican anti-gambling laws throughout the country swiftly, I would not have been willing to qualify the about six months of toleration of Thunderbird's operations alone, without preceding Oficio, by the then outgoing Mexican government as sufficient for creating a legitimate expectation under Art. 1105 of the NAFTA. But even If one considers the "Oficio" as not sufficiently strong and the post-Oficio toleration as not sufficiently prolonged, the combination of the two creates a much stronger case for a protected legitimate expectation¹³⁶. This is also in line with the interpretation guideline of Art. 31 (2) of the Vienna Convention where subsequent conduct of the parties is taken as a significant indicator of their common Intention. In comparative administrative law - In particular In legal systems of the Latin tradition - subsequent conduct by the administration is generally relied upon to interpret earlier, ambiguous, administrative acts and contracts¹³⁷.

If SEGOB had been effectively deceived by dressing up a video-poker operation as an Innocent video arcade game, as the majority suggests, then It had sufficient time to inspect the operations, realise that they were not what was submitted and for which SEGOB had given green light, but something else that was against Mexican law as then interpreted by SEGOB. Given the sensitivity of the issue and the long legal battles of SEGOB with Guardia starting In 1998, it would have been natural for SEGOB to check on the facilities soon after the "Oficio" of August 15, 2000. Lie. Alcantara testified to his ever present will to pursue vigorously and consistently any perpetrators¹³⁸. Nothing would be more normal after a so carefully drafted Oficio than to inspect Thunderbird facilities to see that the "warnings" were observed and the machines were as what they were presented to SEGOB. But there was no action by SEGOB throughout 2000 and beyond - until a new government and thereafter a new Director of SEGOB - Guadelupe Vargas - took office. The first actions against Thunderbird, reflecting the change of interpretation and enforcement attitude, started in February 2001, i.e. only after a new government and a new SEGOB Director had taken office. I do therefore not share the tribunal's view (para 165) that "approximately six months" is "Insufficient to establish that prior to that date SEGOB had authorised (or was Intentionally tolerating) Thunderbird's operations. It was not just the mere passage of time from August 2000 to February 2001 that is relevant, but the fact that toleration and an absence of any action of monitoring, inspection, request for Information or enforcement lasted throughout the whole period remaining for the outgoing PRI government. It only ended when a new, PAN-appointed SEGOB director, took office. As we have to read the "Oficio" in a way that is most likely to reflect the true intention and common understanding of the parties in the context of their Interaction, It is only that period - of the same group of players motivated by the same type of approach and attitude to gambling regulation - that we have to look at. We do not have simply a period of six months' toleration - short some might say for many government agencies to get their acts together, but the full remaining period of the tenure of the government which negotiated and later Issued the "Oficio". I note that in Biloune v. Ghana¹³⁹ the tribunal Identified the about 12-months' long toleration of a visible construction as a key factor for a finding of expropriation, i.e. a sanction that reaches much further than the Art. 1105 NAFTA breach at issue here¹⁴⁰. But 12 months of toleration of a construction process indicates much less than the combination of a formal, though ambiguous, interpretative assurance combined with toleration not only of the prior process of establishing the gambling facilities, but also of their operation subsequent to the Oficio to the very end of the government's tenure. On the Biloune principles, Thunderbird had therefore a much better case for the lesser Art. 1105 NAFTA claim. Different from Biloune where a positive signal from the regulating agency was alleged, but contested, Thunderbird had a very formal assurance letter following its formal request plus a subsequent toleration of the very operations for which the Oficio had been requested.

In Thunderbird, the assurance letter was given In light of a well known interpretative dispute, where the facilities were not only established, but up and running and where the government had a specialised agency charged with monitoring and enforcing the regulation-intensive gambling law and where the government prided itself on rapid and energetic enforcement. The comparison with the Biloune case thus reinforces the view that SEGOB's conduct subsequent to the Oficio letter, throughout the outgoing PRI government, not only expressed toleration, but allows us to read the preceding Oficio in light of the subsequent toleration.

The combination of the "Oficio" with the continuous tolerating acceptance of Thunderbird's operation by SEGOB to the end of the term of the government - which had been responsible for Issuing the Oficio - suggests that SEGOB knew exactly what it gave a green light for and was content with It. The conduct of both parties subsequent to the key "Oficio" - Thunderbird's continued Investment and SEGOB's tolerance - confirms that the "Oficio" was meant to give green light to the installation and operation of exactly the type of software-programmed slot machines Thunderbird operated and that SEGOB was perfectly aware and accepting of this fact - regardless of circuitous and convoluted way it formulated the Oficio.

In Interpreting legal acts, what counts in the end is what the parties intended and what the recipient of a legally relevant communication did and could reasonably understand the main message to be. The fact that it took a new government and a new director - with his own sets of attitudes, affiliations and alliances¹⁴¹ -to reverse the course that the Oficio of August 15, 2000 had most cautiously taken, suggests that the earlier Mexican government had indeed given green light to Thunderbird, had been fully conscious of it and accepted the consequences of Thunderbird now backing its expectation with substantial follow-up Investment. The fact that it took a new government and a new SEGOB director to suddenly reverse the course - and the fact that "the first closure order was issued" against Thunderbird in early 2001 - and not against the confrontational Mexican competitor Guardia - is unlikely to be coincidental: Thunderbird was penalised for having collaborated with the (earlier) government and for having been part of the earlier government's attempt to gradually relax the gambling prohibition.

If this is not enough to explain what SEGOB meant and the investor understood with the Oficio, then the "pro-active" duty of government to avoid contradiction and confusion of the Investor -developed in the MTD v Chile, Tecmed v. Mexico and Metalclad v Mexico cases - would come into play. Given the close interaction between SEGOB and Thunderbird, one has to assume that SEGOB was aware that Thunderbird started to operate with its video-poker and related machines (identified as BESTCO and SCI machines) after the Oficio. If this was not covered by the Oficio - as the majority of the tribunal believes - then SEGOB had a duty to advise the Investor accordingly and to ensure no legitimate expectation would arise. That they did not so, both confirms the meaning SEGOB and Thunderbird assigned to their Oficio, but also that SEGOB would have breached the duty of transparency and fair dealing with the investor by letting him run blindly into an open knife.

The element of breach In the case of legitimate expectations under Art. 1105 of the NAFTA does not consist in the act of creating them, but in the disappointment of such expectations i.e. when a government changes course after the investor made its investment. We need therefore to examine not only how the expectations were created, but also how they were breached. Legitimate expectations -under Art. 1105 of the NAFTA or equivalent Investment protection treaties - is never to be seen as an Iron-clad guarantee - comparable to a long-term concession contract with a stabilisation guarantee - that policies will not change. Throughout the extensive jurisprudence surveyed, we find that if governments reverse their previously communicated and relied upon course, a balancing process takes place between the strength of legitimate expectations (stronger If an Investment for the future has been committed) and the very legitimate goal of retaining "policy space" and governmental flexibility. Equality between Individuals and absence of favouritism - i.e. nondiscrimination - plays a role in the assessment of legitimate expectation¹⁴². That is even more relevant in Investment treaties where the prohibition on discrimination In favour of domestic competitors is formally enshrined, as in Art. 1102 of the NAFTA. Courts have made reference to transitional measures¹⁴³ to smooth a reversal of policy. But this is not what occurred here. With the change of government and SEGOB director, enforcement started with priority and focus on the weakest player: the foreign investor¹⁴⁴. As Licenciado Alcantara confirmed: The first closure order, under the new director, was issued against Thunderbird's Nuevo Laredo facility¹⁴⁵. The new SEGOB director did not go first, as one would have expected, against Guardia who had never sought or obtained a comfort letter from government, but against the foreign Investor who had engaged with the (previous) government and obtained an assurance, as disputed as such assurance later became. Enforcement attempts against Guardia followed, but they were ultimately not effective. It is hard to tell and the evidence is not conclusive if Guardia was simply more skilful with his "amparos" before Mexican courts or If SEGOB was pursuing Guardia with less intensity than Thunderbird, a much easier and politically less protected target. Lie Alcantara's, SEGOB's enforcement lawyer, cross-examination Indicates that the direction of enforcement was not in his discretion but ordered from above by senior authorities ("Unidad de Gobierno")¹⁴⁶ in the "Secretarla de Gobernación". Lie Alcantara - keen as he said he was to enforce vigorously - was excluded from such deliberations and acted simply as a lawyer executing enforcement directions given from above. His cross-examination Indicated quite clearly that when he was given an enforcement job, he went about energetically, but the targets were given to him from above. Nothing has come to light or been produced by Mexico on who these officials were, how they went about their business and if they directed enforcement actions with equal energy against both Thunderbird and Guardia. This is another "black bo"x In Gobernacion overseeing SEGOB. But the results speak against such equality. Since the prima facie results indicate that Thunderbird was singled out without good reason (Guardia's confrontation should in normal circumstances made him the first target), and since access to these people and their conduct controlling enforcement is under Mexico's exclusive control, the prima facie presumption is that they favoured Guardia or at least had a particular reason to go after Thunderbird first rather than after Guardia¹⁴⁷. That leads to another -rebuttable but not rebutted or explained and proved - presumption that there was an Intention to discriminate against Thunderbird and quite plausibly to thereby favour the chief and most potent and visible Mexican competitor. Support comes here again from the method of the Feldman v Mexico tribunal¹⁴⁸ which Inferred from a number of factors -including the willingness of the foreign Investor to raise a NAFTA claim and the better-treatment of a well-connected Mexican Investor - that there was a good case for an Intention to single out Feldman because he was a foreign Investor; with the unwillingness or Inability of the government of Mexico to rebut that plausible conclusion based on available factual "dots" which the tribunal was able to connect with an explanatory "line",¹⁴⁹ the tribunal rightly Inferred from the available "dots" that they were connected by the "line" of discriminatory intention.

While I have come to an agreement with my respected colleagues that such conduct may not have amounted to a full breach of the national treatment duty of Art. 1102, I find more than enough "discriminatory elements" that have to be taken into account when judging the disappointment of legitimate expectation inherent in the rapid priority enforcement of closure against Thunderbird. "Discriminatory elements" may per se not amount to a breach of Art. 1102 of the NAFTA (and I concur that breach of one NAFTA Chapter XI duty does not necessarily Indicate the breach of another one), but in particular In the context of fair and equitable treatment (Art. 1105 of the NAFTA) discriminatory elements have to play a role in the process of determining If problematic conduct has risen to the required threshold of intensity required under Art. 1105. I am comforted here by the similar (or Identical) approach of the prestigious Eureko v. Poland (2005) tribunal; it has also linked "discriminatory conduct" with a finding of a breach of the fair and equitable standard¹⁵⁰.

It is clear from the uncontested evidence and my assessment of the witnesses, in particular Lie Alcantara, that the reversal of government attitudes towards Thunderbird started right after the new PAN government and its new director of SEGOB, Guadeloupe Vargas, took office and that it developed a special vigour In enforcing the law against Thunderbird. That is evidenced by the not contested fact that the first closure order was against Thunderbird. Under normal circumstances, one would expect that the first target of a more vigorous anti-gambling policy should have been Guardia who had pioneered the "skill" machine operation since 1999 and openly defied SEGOB, going as far to brag in public about his success of running such operations "with or without the law"¹⁵¹ - rather than Thunderbird who had chosen the approach of not confronting, but cooperating with the Regulator. Guardia is described in the most illustrative article as "friend of PAN politicians". I therefore find the justification for a presumption of discriminatory enforcement energy and direction In the result. We do not know and hardly are able to know what happens exactly in the "black bo"x of government administration, In particular in sensitive matters and where domestic competitors are linked with government services against foreign competitors. That is why a distinct treatment by result raises the presumption of a discriminatory strategy and Intention. SEGOB was, with respect to Thunderbird, successful before the Mexican courts. But such combination of exceptional enforcement energy and success did not occur against the competing operations of Mr Guardia. What happens within the Mexican courts is not separate from the measures SEGOB took nor does it provide Immunity for SEGOB action: Mexico is before the NAFTA responsible for its courts as it is for the conduct of SEGOB¹⁵².

Without an extensive analysis of the national treatment obligation - Art. 1102 - I read the relevant jurisprudence - Pope-Talbot v. Canada, Myers v Canada, Feldman v Mexico and Occidental v Ecuador - as requiring the claimant to prove "likeness" and different treatment at least de-facto, with the burden of proof that such difference in treatment is either linked to legitimate policy objectives or unrelated to the foreign nationality of the claimant going to defendant. I also read these persuasive precedents as suggesting that the best-placed major domestic competitor¹⁵³ has to be compared with the foreign investor. The fact that there may be other domestic competitors who are also not treated as favourably as the best domestic competitor does not detract from this approach. The reference to "most favourable treatment" in Art. 1102 (3) suggests that it is "the most favourable treatment" accorded to a domestic competitor, and not an "average treatment" or the "worst treatment afforded to a domestic competitor" that is the required benchmark. There is no defense of equally bad treatment for some, politically not favoured, domestic companies.

Treatment means the consolidated conduct by national authorities (Including courts). I accept that discrimination requires a certain materiality and weight; it also requires that it can not be remedied rapidly and practically by an administrative or judicial appeal readily available¹⁵⁴. It also does not Involve a duty of "affirmative action" by the state to equalise all the Informal handicaps which are Inherent in the foreign origin of the Investor nor does it require that bad luck and lack of litigation skill of the Investor In judicial processes be automatically seen as a breach of national treatment. Nor do I see the Art. 1102 obligation to require a government to afford the same toleration to Illegal operations just because it is foreign owned - such as supporting a foreign Mafia group just because a local police chief is in cahoots with a domestic Mafia group.

But we are not faced here with a criminal conduct, but rather an often legitimate gam(bl)ing service the legality of which depends on legal interpretations of the boundaries of the law; these boundaries are neither a thin line nor a bright line, but rather a fuzzy grey area. One can therefore identify government conduct (reinforced by court conduct) - by SEGOB In its enforcement intensity and focus - that leads to the result that the foreign Investor who committed its Investment after a reasonable comfort letter by the previous government suffers first, while the domestic Investor who always played the card of legal confrontation continues to thrive. The most legitimate way to test the margins of a about 60 years old law bypassed by technology is surely to ask the government for an interpretation that takes into account emerging technologies and comparative regulatory practices. Why Thunderbird seems to have been penalised for this approach while the confrontational approach of the major domestic competitor is - 4 years later - still reaping rewards, has not been explained, neither by the government nor by the tribunal in its majority award. The reasons for this difference in result are hard to ascertain. It has to do with what happens within the "black bo"x of interface between government and domestic business people. But my conclusion is that we have at least a presumption of discriminatory and arbitrary elements in the SEGOB enforcement activity. That presumption has not been rebutted by a satisfactory explanation. Mexico has kept studiously silent on Guardia's relationship with the government and the reason for his relative success with the courts. That such difference only emerges after a new government and a new SEGOB director have taken office reinforces the idea that SEGOB went after Thunderbird because it was seen to have reached a deal with the prior government.

The Mexican government could have cleared up such a presumption by producing the key players on its side: The Gobernacion Director General under whose authority the Oficio was executed; the SEGOB official who prepared and signed it; the SEGOB officials to whom the "oficio" was copied to, Guadeloupe Vargas who was the instrument of the reversal of policy with energetic targeting of Thunderbird. Nor did the respondent produce any member of the "Unidad de Gobierno" which, though left faceless and unidentified, ordered Lie Alcantara to focus on and go first after Thunderbird and which must bear responsibility for the relative ineffectiveness of enforcement against Guardia. Alcantara's testimony on the first target of enforcement - Thunderbird - and on the location of the "command and control center" within Gobernación indicates only one thing: That the unnamed powers in the "Unidad de Gobierno" had earmarked Thunderbird as the first and prioritised target. That the government did not produce any of these key players - both In the PRI and the subsequent PAN period - supports a not rebutted presumption that Thunderbird was singled out in enforcement. This is the same legal operation as was carried out by the Feldman v Mexico tribunal which found evidence, though never fully explained by Mexico, that the foreign Investor was targeted by effective audit-based enforcement procedures, while the politically well connected and economically more powerful competitor was left alone¹⁵⁵.

Accordingly, I find that a breach of the duty to respect Investment-backed legitimate expectation under Art. 1105 of the NAFTA has taken place at the time when enforcement began against Thunderbird in 2001 without a similar enforcement effort displayed (on the evidence available and as determined by the operation of the presumption of discriminatory elements) against Guardia. The presumption that at least some discriminatory elements were present In the enforcement against Thunderbird strengthens the position of Thunderbird in the necessary balancing process between its Investment-backed legitimate expectation and the equally legitimate acknowledgement of the need for governmental flexibility. Since it is in the end the de-facto situation of different treatment that is compared -Guardia continues to operate from 2000 throughout 2004 at least -the presumption is that there is discrimination. Perhaps it is mere difference of relative luck and litigation skill (though that is not very probable),¹⁵⁶ perhaps Guardia was protected by higher government authorities and had a better way to persuade the courts. But that is not essential: With the evidence of a de-facto more favourable treatment of Guardia by the Mexican state (administration plus courts), Mexico has the burden of proof of explaining satisfactorily and justifying the available prima facie evidence of discrimination. I find the explanations not satisfactorily as there was no proof that enforcement was equally directed; Mexico did not present witnesses from its "Unidad de Gobierno". Similarly, there was no satisfactory explanation why SEGOB singled out, after the change of government, Thunderbird rather than Guardia.

The acknowledgement of the legitimacy of government flexibility can not justify that Thunderbird was pursued with most vigour and priority when the dominant domestic competitor managed - by whatever ways - to continue his operations throughout at least 2004 - though Thunderbird did and could rely on the positive signal from the then government in August 2001 followed by six months of toleration. SEGOB should have given Thunderbird a negotiated transition period to recoup its expenditures and relocate its operations and equipment within a reasonable period in 2001. The government of Mexico is not prevented, in case of a change of government, to change its interpretation of the law - from the view that the Gambling Law prohibited only machines which "predominantly Involved chance" to one where there was some substantial Involvement of chance in addition to skill; but such a change of interpretation can not override the legitimate expectation created by the earlier government, in particular If It was reasonable for the Investor to have confidence in such expectation and if it was clear to the old and new government that the Investor had carried out substantial Investment because of the government-created expectation that the earlier, more liberal interpretation of the law would be respected. As Jan Paulsson has said:

"Surprising departures from settled patterns of reasoning or outcomes... must be viewed with the greatest scepticism If their effect is to disadvantage the foreigner"¹⁵⁷.

One issue has played an Important, but not very visible role in the arbitration: The Implications of the - uncontested - payment of a 300 000 $ success fee to two Mexican lawyers - Aspe & Arroyo - for obtaining the August 15, 2000 "Oficio" from SEGOB. In principle, it is not exceptional that a success fee was paid for successful negotiations that produced a document that Thunderbird considered important for its Investment process (Including its relations with its financial backers). Both the Thunderbird CEO and Mexico's expert Prof Rose testified on the considerable economic value that a license, or a sublicense legal Instrument has when It allows operation In a not generally open market. There was an Insinuation - never maturing to a full-fledged assertion backed by substantiated facts and evidence - hinting the possibility of corruption. One can not exclude that this Insinuation had some Influence on the case. The role of the success fee and its implication for the existence or not of a "legitimate" expectation has therefore to be squarely addressed as it would undermine a fair hearing, If the issue were allowed to fester, but would not be made transparent and fully discussed. The tribunal notes (para 150), after an extensive discussion of the success fee arrangements, that "these facts do not have a bearing on the tribunal's analysis below" and that it can "only Interpret the 3 August 20000 Solicitud letter on its face value". But the Insinuation about the success fee arrangements hangs like a heavy dark cloud over the case. It is difficult to see how It can not have an effect on the analysis of in particular the "Oficio" which, as I have suggested earlier, can not be done purely "on its face value" as it is part and parcel and in the end the formal outcome of a prolonged interaction between both parties; an examination of this interaction only allows to place the "Oficio" properly In the context of the Investor seeking a regulator's clearance by a formal letter confirming a more liberal interpretation of the Gambling Law and the regulator's accommodation of this request, albeit in a convoluted and ambiguous format "protecting its back". Since I consider that attention must be paid to the context of the Oficio, that it can not be interpreted purely on its own as a free-floating document without history and purpose, I consider that the "success fee" story may have coloured the tribunal's award. For this reason, the success fee story and its possible implications for the legal effect of the "Oficio" in creating a legitimate expectation has to be faced head on - rather than be developed in detail, then hang ominously over the legitimate expectation claim, but finally be dismissed as formally Irrelevant and as such no longer a suitable object for a proper examination.

First, there is no doubt that the use of illicit practices such as direct or indirect bribery of government officials would be a reason to invalidate any legal effect of the "Oficio", as indeed the legitimacy of Thunderbird's claim as such. There is ample jurisprudence that a legitimate expectation protected by Art. 1105 of the NAFTA can not be created if deception, fraud or other Illicit means were used to obtain the governmental assurance or other rights obtained from the government in this way¹⁵⁸. There can be no International treaty protection for rights obtained by illicit means. In such cases, there may be an expectation, but not a "legitimate" one. It is generally very difficult to prove bribery as there is usually little if any paper trail. However, arbitral tribunals and courts, in particularly of more recent date and under the Influence of the authoritative International conventions (mainly, but not exclusively the OECD anti-bribery convention) have been ready to use presumptions rather than full-fledged and hard to obtain full evidence. If a transaction creates enough suspicion so that - in the practice of the US Foreign Corrupt Practices Act - a "red flag" should show up on the face of the transaction, It is sufficient to require the party in control of such a transaction to prove that it was contrary to "red flag" indicators a proper one¹⁵⁹. Note again the Methanex v. US award:

"The tribunal is not averse to trying to "connect the dots" as a way of testing Methanex's hypothesis", and: "Inference is an appropriate mode of decision In circumstances in which firmer evidence is not available" (Part III, B, para. 57)

But in this dispute, respondent has hinted, insinuated, focused in cross-examination on the role of the two lawyer-lobbyists Aspe and Arroyo, raised and queries the payment modalities of the success fee (a transfer made from Mexico to an account in the US), but it has never explicitly and properly asserted and tried to substantiate that the success fee had been an Instrument of bribery or that at least it indicated - as a "red flag" - a suspicion of bribery of SEGOB officials. Neither Mexico nor Thunderbird have made the key players - Mssrs Orozco Aceves (Director General de Gobierno); Martinez Ortiz; Antunano - the signatory of the "Oficio", Guadeloupe Vargas, the successor director of SEGOB, the members of the "Unidad de Gobierno" which decided on enforcement priorities nor lawyers Aspe & Arroyo, available. I have advocated throughout this procedure for pressure under the - limited - powers of the tribunal to make them appear, but in the end the Insinuation remained what it was - an insinuation without substantiation and without being available for proper and full testing before the tribunal. The tribunal therefore should, in my view, have drawn Inferences from this failure of Mexico to produce these key witnesses and officials - I follow here the as mostly very persuasive view of the late F.A. Mann, one of the past masters of International Investment law¹⁶⁰.

In this situation, as always when we are faced with a "black bo"x In which relevant events occur but which we do not see, tribunals have to work with a system of presumptions and tests of plausibility. It is theoretically not impossible that the success fee and the work of the two lawyer-lobbyists Aspe & Arroyo had to do with illicit influencing of public officials. Similarly there is testimony by Mexico's expert Prof Rose on the great economic value and natural attractiveness that lies in collusion between a domestic gambling operator and the national gambling regulator to foreclose the operation of a foreign competitor:¹⁶¹

"If they have closed one down and don't close down a competitor who is very public, then there is the possibility, very strong possibility of bribes".

One can Infer therefore, the possibility that the energetic closure action against Thunderbird after Mr Guadeloupe Vargas took office (without a commensurate enforcement energy and result against Guardia) might Involve an underlying Guardia/SEGOB alliance¹⁶². But both these theories are conjecture rather than proven fact. Having worked in Investment negotiations in developing and transition countries for over 30 years, I have rarely encountered a deal that was not surrounded by corruption gossip. Relying on gossip - as plausible as it may appear In particular In conspiratorial explanation models - is never a professional way to proceed in such matters.

A legitimate interpretation of the events in 2000 is equally plausible¹⁶³: The outgoing government did consider liberalisation of gambling. Prof Rose, for Mexico, alluded to the window of opportunity that was open for a short while in 2000. The Mexican Congress had commissioned a study which, subsequently, in 2002 Indicated the benefits of bringing the gambling Industry back to Mexico. Senior politician and presidential hopeful Francisco Labastida had - as is uncontested - raised and supported the idea of "stealth liberalisation" through an Interpretative assurance. Thunderbird's uncontested narrative of government contacts indicates that SEGOB officials were appreciative of Thunderbird's willingness to engage rather, as Guardia, confront the government. The sudden emergence of the "predominantly either skill - then yes, or predominantly chance, then no" - criterium In the "Oficio" of August 2000 attests that the criteria used for liberalising gambling regulation In other jurisdictions had come to the attention of SEGOB and found favour with its senior officials. That success fees or "lump sum payments" are paid for lobbyists (often lawyers) for achieving results - rather than just letting them maximise billable hours - is not unusual and not in Industries where government licensing - by way of formal concessions or less formal interpretative comfort letters - is of great value. Dealing with governments, Including, but not only, developing countries is always a difficult matter, particularly for foreign Investors. It is likely to be rare to find a case where local lobbyists - "government relations experts" -do not have to be employed. They come with risks, but their Involvement is in practice Inevitable.

But what must ultimately decide this issue is that Mexico has the burden of proof - even If such burden can be discharged in an easier way by evidence of sufficient "red flag indicators". Mexico is responsible for the very formal conduct of its officials; there is the presumption of the validity, legitimacy and effectiveness of the Oficio of August 2000. Insinuating corruption but not submitting It for proper testing in legal combat is not an Instrument that tribunals should pay any attention to, directly or Indirectly, explicitly or Implicitly. To quote the recent Methanex v US award in the context of examining the prospect of Inferring conduct for which Indicator "dots" might be available, but not the proof of the full story:

"therefore, to establish undue Influence, Methanex would, at least, have to be in a position to allege if not also to demonstrate that a legal violation took place" (part III, Chapter B, para 22)

The Methanex tribunal was not impressed with the claim for Improper behaviour on the part of the regulating state, though the assertions and facts in Methanex were stronger than the hints and Innuendo in Thunderbird. As in Methanex, there was a reasonable explanation for the context and underlying policy of SEGOB under the earlier government to test and marginally expand the boundaries of the gambling law embodied in the Oficio. As a result, the Thunderbird allegations should deserve even less consideration than similar, but factually much more substantiated, allegations in the Methanex case.

Mexico, however, has not put forward any substantiated assertion or evidence; it has refrained from putting forward the main witnesses under its control, that is the SEGOB officials past and present. Thunderbird has equally refrained from putting forward Aspe & Arroyo, over which It presumably has less control than Mexico over its own officials. But the issue of bribery affecting the Oficio is something that Mexico has to prove, while Thunderbird only has to come forward with counter-evidence once Mexico has provided prima facie evidence of at least "red flag signals". Insinuation without the readiness to come forward and have a substantiated allegation properly debated and tested before the tribunal is a poisonous way to conduct litigation. It has become more and more frequent in Investment arbitration as both claimants and defendants raise such hints, without being ready to submit them to a full and fair trial. Tribunals should actively discourage this tactic and ensure it plays no role, directly or indirectly, in their deliberation. For these reasons, I see in the light of the evidence available and the defense made by the respondent no reason to question the validity of the legitimate expectation created by the "Oficio" in combination with SEGOB's subsequent conduct and sudden reversal once new powers took over. If Mexico had wished to question the legitimacy of the expectation created, it should have openly and directly, with substantiated assertions and proper evidence - mainly making its own officials in SEGOB (Including the higher-level SEGOB officials which directed its enforcement efforts above Lic Alcantara and which Issued the Oficio) available for testimony and cross-examination before the tribunal¹⁶⁴.

Since the majority of the tribunal rejected all claims by Thunderbird, I do not need to get into the details of how compensation should have been calculated. But I can provide an outline. I concur largely with Mexico's back-up argument that at most "reliance damages", that is damages which were directly and reasonably caused by reliance of Thunderbird on the "Oficio", later confirmed by SEGOB toleration, are owed. It is widely recognised that a "legitimate expectation" can only then lead to compensation If there was "detrimental reliance", i.e. a link between the expectation and Investment made - a principle which In American takings law has led to the notion of "Investment-backed expectations"¹⁶⁵. That detrimental reliance must also be a "reasonable" one (see Waste Management II v. Mexico, supra). For a normal business person engaged in foreign investment in Mexico, the "Oficio" and the subsequent conduct by SEGOB must have allowed the conclusion that the government was ready to accept the operation of the gaming machines envisaged - something which not only the Oficio, but also other factors (the high-level encouragement of Thunderbird, the discussion about liberalisation of an obsolete gambling law) supported.

The fact that the "Oficio" may have been only one of the various factors in its Investment process is not an objection. Business decisions are usually made on the basis of several significant reasons and a single, causative relationship between one key factor - the Oficio - and the overall subsequent conduct by the Investor is hard to establish. There is enough evidence that the interpretative assurance by SEGOB was an Important factor for Thunderbird for opening the facilities which were already more or less ready and for adding new facilities. There was credible evidence by the CEO of Thunderbird, Jack Mitchell, by P. Watson, the business development consultant and by other credible references to the importance attached to this "comfort letter" by the financial backers. Plus, the payment of the success fee Itself indicated that the comfort letter was for Thunderbird a matter of great significance. If the "Oficio" had not been very Important for Thunderbird's Investment process as the majority award (para 164) suggests, why did then Thunderbird pay instantly the not Insignificant amount of 300 000 $ to those who helped to arrange it?¹⁶⁶

Thunderbird's position is that compensation were owed (estimated at over 100 M US $) as if its operations had been well established, were likely to run at a high rate of profitability unencumbered by future competition or regulatory measures and should be compensated on the basis of projecting an Initial measure of profitability, after disregarding Initial start-up costs, into a long-term future. That, however, is not a legally viable proposition: First, that would equate a "legitimate expectation" with a firm, long-term concession contract. But a legitimate expectation under Art. 1105 of the NAFTA is a much weaker legal position than a long-term concession contract. As all precedents show, governments retain flexibility to reverse a legitimate expectation In a reasonable way with transitional measures. A conmfort letter may create a legally protected legitimate expectation even If It is not crystal-clear; but It is by far not the equal of a proper long-term concession contract. Even If we had a long-term, legally valid concession contract, one would have to take Into account that the Initial high profitability stemming from a successful start-up operation of a newcomer In a hitherto largely closed market is likely to give way as other competitors move In and thus, In the normal process of economic logic, depress the profitability. In cases of legitimate expectation (detrimental reliance), at most the government owes the Investor the "negative Interest", i.e. the expenditure the Investor has undertaken with confidence In the reliability of the government position communicated. But It does not give a claim to the "positive Interest", i.e. to be placed Into a situation as If the government had committed In the form of a valid long-term concession contract.

The claimant can only reasonably be assumed to have relied on the "Oficio" from about August 2000 to February 2001 when the first dark clouds started to cover the sky over Thunderbird. By then, It had received a warning, could have easily appreciated the weakness of its legal and political decision In light of the ambiguities of the "Oficio" and the entry Into power of a new government. By February 2001, It can no longer be assumed to have continued to Invest In full confidence In SEGOB's comfort letter. By October 2001, It was clear that there was a serious problem and the wise course of action would have been to stop operations and take the machines out of Mexico. The relevant expenditures Incurred In direct detrimental reliance are therefore quite modest. They can also not Include the 300 000 $ success fee which was not an Investment after and because of the "comfort letter", but rather a payment to the lawyer-lobbyists for getting the comfort letter. Finally, in line with Art. 39 of the ILC Articles on State Responsibility and the MTD v Chile tribunal (paras 240-243), the absence of a rigorous due diligence¹⁶⁷ in terms of ambiguous qualification of the machines in its "Solicitud"¹⁶⁸ and the unquestioned reliance on the "Oficio" in spite of its manifold obfuscations and ambiguities, should lead to a reduction of the compensation due under the concept of mitigation of damage and contributory negligence. While I do not have at this stage to calculate the hypothetical compensation In detail as the tribunal has rejected the claim, I would not have advocated a compensation award exceeding 500 000 $.

To sum up: The award I have advocated would have provided a fair and equitable solution. Neither would it have produced exorbitant damages likely to undermine the acceptance of the Investment arbitration regime nor would it have let Mexico - which played contradictory games with Thunderbird - come out of the arbitration without a good-governance signal: To be more careful with official assurances to Investors and to be more respectful of the expectation created with such assurances even In the context of a change of government, senior staff and policy direction¹⁶⁹. This is not a zero-sum Issue: If Mexico's official declarations and assurances are given legal effect by way of application of the legitimate expectations concept under Art. 1105 of the NAFTA, Mexico can enhance the credibility and effectiveness of its policy tools to encourage foreign Investment. To deny such effect is to reduce the effectiveness of instruments available to governments required to micro-manage an Investment promotion policy in relation with specific Investors. To Thunderbird and other entrepreneurial companies in a similar situation, the award would also have sent a due-diligence and good-governance signal as well: To be more careful with ambiguously drafted government assurances, with local lobbyists promising to control government conduct and to phase Investment more prudently in alignment with the degree of legal assurance received - and not to make exorbitant damage claims with no legal foundation. Both parties should, in my view, have settled the matter much earlier, In the sense of a negotiated "velvet exit" of Thunderbird from Mexico and not in the style of an abrupt expulsion of the Investor out of Mexico.

The tribunal orders the claimant to pay % of the arbitration cost and to pay to Mexico 3/4 of the costs of Mexico's own legal expenditures. It applies therefore the principle of "costs follow the event" to attorney costs. Such a practice is relatively frequent in civil-law litigation, less so in International commercial arbitration, but mostly with considerable limitations and judicial and tribunal competence to reduce such costs¹⁷⁰. It is not at all practice in North American litigation and arbitration; some US courts have prohibited awards of attorney fees in arbitration¹⁷¹. "Fee shifting" is as a rule only allowed in case of misconduct - contempt of court, incompetent or unacceptable litigation conduct, bad faith in arbitration or frivolous claims"¹⁷².

One of the US federal judges most respected for an understanding of economic analysis recently ruled that in international sales the loss claimed under Art. 74 CISG did not include attorney fees¹⁷³. The Uncitral rules - Art. 40 - constitute a compromise between the "European" and the "American rule"¹⁷⁴ by establishing a slight (but not mandatory) preference for the "costs follow the event" rule for the arbitration cost, but leave it in the tribunal's discretion to allocate the costs of legal representation (Art. 40 (1) and (2). I can not follow the tribunal's position (paras 213) that the Uncitral rules prefer that the losing investor pays to the prevailing government legal representation ("attorney") costs: Section (1) of Art. 40 of the Uncitral rules prefers (without obligation) the "loser pays" principle for arbitration costs, but then, in section 2, different and distinct from section (1) leaves it fully open to the tribunal how to allocate legal representation ("attorney") costs¹⁷⁵. The distinction between Art. 40 (1) and Art. 40 (2) can not be simply explained as giving "larger discretion" (a term that is hard to appreciate - what is the difference between "discretion" and "larger discretion"?). It must mean something sensible. The only explanation is - in accordance with Myers v Canada¹⁷⁶ - that the limited preference for the "European Rule" for arbitration costs in Art. 40 (1) is omitted in favour of complete neutrality between the European and American rules with respect to "attorney costs" in Art. 40 (2). If it were otherwise, the distinction between arbitration costs (Art. 40 (1) and legal representation costs (Art. 40 (2) would not have been necessary. It is true that according to Art. 1135 the tribunal is empowered to award cost and that reference is made to applicable arbitration rules - including Art. 40 (2) UNCITRAL rules. But I suggest that a tribunal has to exercise its discretion under Art. 40 (2) of the UNCITRAL rules in conformity with well established standard practice; it is well established in administrative law that discretion is not unfettered and arbitrary, but needs to be exercised in line with established principles and practice. What is "appropriate and reasonable in the circumstances" (para 216) does not confer arbitrary discretion on the tribunal in its cost decision, but only discretion within the boundaries of established jurisprudence. That jurisprudence has developed the principle that legal representation costs can only be awarded in case of spurious claims or bad-faith litigation tactics.

The tribunal's decision to order the losing Investor-claimant to pay most of the costs of legal representation of the winning state respondent is a significant departure from established jurisprudence by all previous NAFTA tribunals and by most, if not virtually all other BIT-based ICSID cases. It is not required as a measure of damages; the distinction of the award of attorney costs for Investor-claimants and not for respondent governments that can sometimes be observe, can be advocated on the basis of this concept: Only - successful -claimants can reasonably argue that attorney costs form part of their damages claim; respondent governments can only argue procedural law principles. It is not required by the Uncitral rules which are relied upon In Investment arbitration to provide a standard set of procedural rules; reference to them was not intended to import the "European rule" to Investment arbitration under the NAFTA nor does their specific language (Art. 40 (2) require application of the "European rule". Since Art. 40 (2) of the Uncitral rules provides for arbitral discretion, such discretion must, I propose, be exercised in harmony with the well developed jurisprudence of, first, the NAFTA tribunals and, second, other ICSID-based BIT awards¹⁷⁷. I am therefore bound to disagree with the tribunal's departure from well established jurisprudence. Should an Investment tribunal operating under a treaty decide to diverge from well established jurisprudence, it should provide in detail and depth the why such deviation should be exceptionally justified. It is also my view that such deviation should be the subject of a proper hearing (orally or In writing) for both parties and based on extensive, in-depth reasoning to establish the compelling need for such an exceptional approach.

There are three recent and relevant surveys of the cost decisions:

M. Buehler, In a survey that is focused on International commercial arbitration - not Investment arbitration - concludes that in "mixed arbitration (meaning Investor-state), too, the loser-pavs rule seems to be the exception rather than the rule ". "In most cases, the tribunals simply ordered each party to bear half of the procedural costs and left the parties' costs where they fell." "Waste Management v. Mexico seems to be the only case where the private party was ordered to bear the procedural costs because it lost its case (nonetheless, legal costs were not allocated"¹⁷⁸. - The survey of N. Rubins¹⁷⁹ focusing on Investment arbitration notes that :

"awards of costs or legal fees against unsuccessful claimants In Investment arbitration cases appear to be exceedingly rare " and that "Investment arbitration tribunals have examined the issue on a case-by-case basis, more often than not dividing the arbitration costs equally between the parties, and, more frequently yet, ordering each party to bear its own legal fees".

The most recent and exhaustive survey by Professor Gotanda finds that:

"One trend that has developed is the tribunals' hesitation In awarding attorneys fees against a private party under the ICSID rules. Where there is case law awarding attorneys' fees against a losing government party, there is a noticeable lack of cases where tribunals order a losing private party to bear the winning party's cost of representation. Even where the private party's claim or defense fails in its entirety tribunals have opted not to award attorneys' fees and split the costs of the proceeding between the two parties"¹⁸⁰.

A most recent survey on "ICSID Arbitration Awards and Cost"¹⁸¹ finds, on a survey of 14 awards:

"The fourteen arbitration decisions reviewed indicate that, with few exceptions, the tribunals generally allocate one half of the arbitration costs to each party. In addition, the tribunals generally hold each party responsible for their own representation costs".

And it concludes that only "reckless" or "bad faith" claims have led to claimant responsibility for respondent's representation costs:¹⁸²

"Unless there is a significant error, Inconvenience of unreasonable action on the part of one party, it is most likely that each party will bear half of the arbitration costs and their own respective representation cost".

Since this tribunal is departing from general practice, unanimously identified by all recent commentators, the issue requires closer analysis, both in terms of precedential cases ("how exceedingly rare is this tribunal's approach"?) and in terms of the particular criteria that might, exceptionally and in specific cases, justify the U-turn as made by this tribunal. Before I do this, an observation is called for on the significance of precedent in international Investment arbitration. The difference of the role of precedent in commercial arbitration from its role in International Investment arbitration may explain why the tribunal made its unusual cost decision, without giving the parties the chance to comment on this departure, and without detailed reasoning to justify its departure. In commercial arbitration, there is no formal and very limited practical "persuasive" precedent. The reason is simply that most awards are still confidential; only a few are published or made public in sanitised form. The award is in the main an explanation to the parties of the reasoning of the tribunal, and there is no requirement or expectation of transparency, Including its consequence of respect for established jurisprudence or the need to explain a significant deviation from well-established principles. Similarly, commercial arbitration as a rule applies specific rules of contracts; Investment arbitration, on the other hand, applies treaty provisions that are general; in their Investment protection core content, the Investment treaties (with the equivalent of the multilateral treaties now well over 3500) express common principles and very similar, often identical language¹⁸³. Every Interpretation that is public is likely to exercise a general effect and will be taken up by counsel and tribunals in subsequent cases.

But that is different in Investment arbitration: It is not two equal parties who agreed specifically to submit a dispute to confidential resolution, but it is the Investor only which raises a matter usually Involving public policy and administrative misconduct (as measured under a treaty's obligations) against the host state. Investment arbitration is in substance a special form of International quasi-judicial review of governmental conduct using as a default the methods of commercial arbitration¹⁸⁴. Following criticism for alleged "secrecy" (i.e. confidentiality in commercial arbitration language), awards are increasingly made public and debated - this is the now uniform practice in NAFTA Chapter XI arbitration and Increasingly also so in ICSID cases, with all indicators pointing towards greater transparency. As a result of this primarily International and public-law character of Investment arbitration, with transparency and public debate, the principles and practices of International law with respect to precedent become more relevant compared to the almost non-existent and at most illustrative character of precedent in commercial arbitration. That difference still needs to be appreciated¹⁸⁵. In International and international economic law - to which Investment arbitration properly belongs - there may not be a formal "stare decisis" rule as in common law countries, but precedent plays an Important role. Tribunals and courts may disagree and are at full liberty to deviate from specific awards, but it is hard to maintain that they can and should not respect well-established jurisprudence. WTO, ICJ and in particular Investment treaty jurisprudence shows the importance to tribunals of not "confronting" established case law by divergent opinion - except if it is possible to clearly distinguish and justify in-depth such divergence. The role of precedent has been recognised de facto in the reasoning style of tribunals, but can also be formally Inferred from Art. 1131 (1) of the NAFTA - which calls for application of the "applicable rules of international law"; these include, according to Art. 38 of the statute of the International Court of Justice /'International custom, as evidence of general practice accepted as law" and " judicial decisions " as "subsidiary means for the determination of rules of law".

In consequence, it appears to me that at the very least that, if a tribunal wishes in a significant question, to adopt a novel philosophy that diverges from well established principles is under an obligation to provide the parties with an opportunity of a full debate - such as calling for a "separate argument on the allocation of fees and expenses after rendering a decision on the merits"¹⁸⁶ - and to provide extensive reasoning which shows that the tribunal is both familiar with established jurisprudence and is prepared to justify its departure from such jurisprudence with in-depth reasoning. While the parties have both claimed all relevant costs, one should assume that they expected and assumed reasonably that the tribunal would follow general NAFTA and ICSID practice with respect to the attorney cost issue. If the tribunal wishes to diverge, it should give the parties the opportunity of a full hearing, at least in writing, to focus on this issue.

Arguably, in the context of NAFTA jurisprudence it is not proper at all for any tribunal - whatever their "depth of reasoning" and regardless of whether full hearing was afforded to the parties on the point of divergence - to diverge in a significant way, as this tribunal does here; for such purposes the NAFTA has set up the intergovernmental NAFTA Free Trade Commission (Art. 2001); in cases of similar significance this Commission has provided an authoritative guideline¹⁸⁷.

A review of publicly accessible prior NAFTA awards indicates that there is no precedent for ordering a losing claimant to pay the legal expenses of government except in the case of spurious claims or bad-faith litigation. Waste Management v. Mexico¹⁸⁸ is often seen as an exception In so far as the claimant was ordered to pay the cost of the arbitration - but it was not ordered to pay the legal expenses of the respondent. In Azinian v. Mexico, in spite of finding fraudulent conduct with respect to the contracting out of a municipal concession contract, the tribunal split the costs of the arbitration, with each party bearing its own legal expenses. It argued among others that the "novelty of the issues" and professional conduct by counsel as reasons for its cost allocation in a case that might be seen as a "spurious" or "frivolous" claim by a company that did not measure up to a reasonable standard of integrity and competence¹⁸⁹.

In order to seek justifications for the tribunal's cost award, we need therefore to seek out the few cases where an award of legal expenses against the losing claimant was - at least to some extent - determined or mentioned: There has been up to now almost no ICSID case¹⁹⁰ where losing claimant had to pay respondent government's cost. In some cases, the legal costs of the claimant had to be paid by the respondent government, taking Into account relative success, efficient conduct and the principle of full compensation of Investor damages suffered¹⁹¹. In the Myers case¹⁹², the government of Canada pointed out explicitly that It was NAFTA chapter XI practice not to award legal representation costs. The tribunal itself held that "

"Some arbitral tribunals are reluctant to order the losing party to pay the winner's representation costs, unless the winner has prevailed over a manifestly spurious or unmeritorious position taken by the loser", (para. 33)

Even in long-term concession contract cases resembling investment disputes, the practice is rather to let each party bear its legal costs, even if the losing respondent has to bear 100% of the tribunal costs¹⁹³. In other recent cases (e.g. Noble v Romania), the "Arbitral Tribunal deems it fair and reasonable that the cost burden be shared equally between the parties, each bearing its own legal and other expenses and 50% of the arbitration costs" (para 236) though "all the claims ultimately failed" (para. 235).

Dr Benhamlda, in a recent publication In TDM¹⁹⁴, reviewed about 26 NAFTA and ICSID decisions rendered against the Investor In favour of the state. In 19 of these cases, the tribunals decided that every party has to bear its own legal representation costs and to share the arbitration expenditures (tribunal, supporting Institution). In all others where the tribunal awarded the winning respondent all or part of legal representation costs an element of either spurious claim or bad-faith litigation tactic was present. In Soufraki v. UAE (not a NAFTA case), the tribunal ordered the claimant to bear 2/3s of the arbitration cost -but each party to assume its litigation expenses. The only possibly relevant Investment awards we have been able to Identify¹⁹⁵ outside the context of the NAFTA and outside the context of arbitration rules mandating the "costs follow events rule" (e.g. Stockholm Chamber of Commerce) are the - unanimous - Generation Ukraine v. Ukraine award and the Uncitral-based (non-NAFTA, non-ICSID) Link Trading v. Moldovia case.

In the Generation Ukraine case¹⁹⁶ the tribunal did award all costs Ukraine had paid into ICSID and added a contribution of 100 000 $ to Ukraine's legal fees (In the Link Trading case the "contribution" was 22 200 $). But one needs to read the award to get a flavour for the reasons. Not only was the claim rejected in its entirety and no possible reason for a justified claim was found to exist, but the tribunal was also extremely dissatisfied with the claimant's conduct before the tribunal. In other words: It considered the claim as spurious and a not justifiable waste of the resources and attention of respondent and tribunal, both in terms of jurisdiction, merits and conduct before the tribunal.

"The Claimant's written presentation of its case has also been convoluted, repetitive, and legally Incoherent. It has obliged the Respondent and the Tribunal to examine a myriad of factual issues which have ultimately been revealed as irrelevant to any conceivable legal theory of jurisdiction, liability or recovery. its characterisation of evidence has been unacceptably slanted, and has required the Respondent and the Tribunal to verify every allegation with suspicion", (para 24/2_; - "The Claimant's position has also been notably inconsistent." (para 24.3) - Moreover, the Claimant's presentation of its damages claim has reposed on the flimsiest foundation, (para. 24.4) - The Claimant's presentation has lacked the intellectual rigour and discipline one would expect of a party seeking to establish a cause of action before an International tribunal. This lack of discipline has needlessly complicated the examination of the claim, (para 24.6) ". "The claimants submissions.. have been seriously flawed due to the absence of a coherent analysis" (para. 20.26)

The "contribution" of 22200 US $ to the respondent's attorney costs in the Link Trading case (under Uncitral rules) could prlma vista be seen as supporting, marginally, the cost decision In this award. There is, however, no review of precedent and only a very general reference to Art. 40 of the Uncitral rules in the award. But the tribunal also expressed, in its cost decision, dissatisfaction with the litigation conduct of claimant. It noted that the costs significantly exceeded what the tribunal estimated for the security deposit: "due principally to the unsolicited further submission of Claimant" (para. 96). This reference suggests that the 22200 US $ contribution reflected an unnecessary increase of the arbitration cost due to unreasonable litigation conduct by claimant. The case, therefore, rather supports the principle that attorney costs of prevailing respondent can only be allocated to the unsuccessful claimant if there was in the tribunal's judgment evidence of unreasonable, cost-enhancing, conduct. E contrarlo, the Link Trading award therefore follows the rule that Art. 40 (2) of the Uncitral rules has to be applied in Investment disputes so that the losing claimant does not have to pay the respondent's costs of legal representation. Since there is no such evidence or Indication In the Thunderbird award of any professional cost-inflating or otherwise "bad faith" misconduct, the Link Trading case provides another example of a "jurisprudence constante": Each party, in particular respondent, have to bear their own attorney costs.

The only NAFTA/ICSID case¹⁹⁷ where the parties' legal costs were awarded to the prevailing government¹⁹⁸ is the Methanex v US case that came out in August 2005¹⁹⁹. The tribunal here recognizes, based on a survey of 1991 - i.e. before Investment cases became widespread and at a date where modern NAFTA/ICSID jurisprudence on costs did not exist (as it does now):

"Certain tribunals are reluctant to order the unsuccessful party to pay the costs of the successful party's legal representation unless the successful party has prevailed over a manifestly spurious position taken by the unsuccessful party."

The Methanex award mentions that "other tribunals consider that the successful party should not normally be left out of pocket in respect of the legal costs reasonably Incurred in enforcing or defending its legal rights" - but it does not mention such cases, in particular Investment claims where the government has prevailed. It also does not discuss the contrast between arbitration and legal costs as is set up between section 1 and section 2 of Art. 40 of the Uncitral rules nor does it give any more detailed reasoning. But from the procedural history of the case, one can Infer that the tribunal was not satisfied with the way claimant conducted its claim, In particular with respect to evidence²⁰⁰. Given this situation, the Methanex case should be seen as awarding attorney costs to prevailing respondent for the well-established reasons (even mentioned in the award) at para. 9), that is either frivolous claims or claims where the tribunals have developed a practice of penalizing unprofessional conduct by the party against which the cost determination is made. To quote:

"The tribunal decided that this documentation was procured by Methanex unlawfully... in violation of a general duty of good faith Imposed by Uncitral rules and, indeed, Incumbent on all who participate in international arbitration"²⁰¹.

From this survey, it is safe to infer that it is by now a standard principle of International Investment law, In particular In the NAFTA context, but also in other ICSID cases, that in principle, each party bears its own legal costs and the costs of the arbitration are shared. Exceptions to this rule have occurred very rarely with respect to the arbitration expenditures and in favour of the winning claimant, but never - except for a limited "contribution" to the government's legal expenses in the Generation Ukraine and Link Trading v Moldovia case -in the case of a losing claimant. The only concept under which this so far well-established rule has not been observed or a different treatment suggested is for "manifestly spurious or unmeritorious" positions taken by the loser, unprofessional conduct and significant breach of good-faith in arbitration²⁰². There are a good reasons for this approach which has so far been intuitively, but not yet explicitly appreciated by tribunals in thrall to the attitudes prevalent in commercial arbitration: Investment arbitration is not a reciprocally agreed and structured method of dispute resolution. It is a unilateral right of Investors - not mirrored by a reciprocal government right - to claim against alleged misconduct by governments under an Investment treaty²⁰³. It is in substance comparable at most to national and international judicial review of administrative conduct - rather than to the reciprocal "contract" model of commercial arbitration. Governments can not sue Investors because Investors can not breach the treaty disciplines such as "expropriation", discrimination or fair and equitable treatment. They focus exclusively on governmental action targeting foreign Investors. Governments have made this asymmetric right available because it helps them to attract capital and Improves their Internal governance, and the perception of their governance quality Internationally.

This principle of cost allocation In International judicial review of government conduct is also applied in GATT litigation. There has been a formal proposal to award litigation costs to winning developing countries because of the prohibitively high costs of WTO litigation; one can see this as a similar concept to the idea that Investors - in particularly smaller companies - should not be penalized for complaining about host state breach of treaty Investment protection obligations; the common Idea is that for under-resourced claimants access to justice is Illusionary if the cost (and risk) of litigation is prohibitive²⁰⁴. Costs of the winning respondent have also not been awarded in UN Compensation Commission cases; in the Iran-US Claims Tribunal cases, in a small number of cases where Iran prevailed, it was awarded very modest costs (a few thousand dollars)²⁰⁵. It is worth noting that the Iran-US Claims Tribunal relied, in order to reject award of attorney costs, inter alia, on the fact that in the US each party bears its own attorney fees. its approach of avoiding generalization of a particular legal-culture approach to the cost issue makes sense to me.

The judicial practice most comparable to treaty-based investor-state arbitration is the judicial recourse available to Individuals against states under the European Convention on Human Rights; again, states have to defray their own legal representation expenditures, even if they prevail.

Imposing the risk of government attorney costs on losing Investors in effect undermines the very purpose of such treaties; it raises the litigation risk In factual situations which are as a rule ambiguous, confused and contradictory to a prohibitive level, in particularly for smaller companies for whom litigation risk is high and where a government enjoys significant superiority in terms of expertise, experience and resources available for defense against NAFTA arbitration. In this particular case, we have had a well integrated, highly competent, coordinated and seamlessly functioning government defense team consisting of over 7 or 8 Mexican and International lawyers, including from two expert law firms, hardened by many rounds of NAFTA Chapter XI litigation facing a small entrepreneurial company with (equally competent I should add) two outside singlepractitioner lawyers. The tribunal's break-out from - for good reasons

- established NAFTA and BIT/ICSID jurisprudence on cost allocation can only be seen as foreclosing access to this type of justice for smaller companies. But that is not the objective of the NAFTA treaty which, to promote trade, new employment opportunities (note the Preamble of the NAFTA), Increase Investment opportunities and eliminate barriers to trade in, and facilitate cross-border movement of goods and services (Art. 102) is not intended to provide access to Investment arbitration only to major US and Canadian multinational companies. The approach to costs in this award suggests that Investment arbitration is only for the very large companies, leaving out entrepreneurs with Initiative, willingness to take (sometimes perhaps recklessly) risk and who may not have the same "International corporate style" appeal of the "men In dark suits". But there is no indication that this was the Intention of the negotiators of such treaties. The highly unusual cost award thus casts a "chill" over attempts by junior companies to rely on the NAFTA's Investment protection regime and makes that recourse - very high-risk anyway²⁰⁶

- doubly prohibitive because of the now added cost risk. In effect and in practice, it makes recourse to independent justice for smaller companies prohibitive.

The only reason possible to use the reversal of standard NAFTA and BIT cost jurisprudence is legitimate reliance on the principle of "manifestly spurious claims" and grave professional misconduct by claimant and its advocates²⁰⁷. But it is hard to find such evidence in the Thunderbird - Mexico case. The tribunal itself has been explicitly positive about the professional conduct of the arbitration by claimant (para 213). if the claim had been "manifestly spurious", the tribunal could have accepted Mexico's request for bifurcation and in a preliminary decision rejected jurisdiction and admissibility. The implication of ICJ jurisprudence on the requirement that claims be founded prima facie in law and in fact, suggests that manifestly spurious claims should be rejected in a preliminary phase²⁰⁸. There is also the practice of the Iran-US claims tribunal to dismiss a claim that is manifestly without merit²⁰⁹. But this tribunal rejected the Mexican request for a preliminary phase focusing on jurisdiction and admissibility issues only. This conduct of the tribunal hardly suggests that the claim was manifestly spurious. The fact that it had to rely on a painstaking ex-post analysis of a convoluted and ambivalent Mexican "comfort letter " ("oficio"), arrived at after years of in-depth argument Involving dozens of fine-comb-handling lawyers and evidence from both sides, does not help to establish a "manifestly spurious claim".

Similarly, I do not share the tribunal's statement that, in order to distinguish Itself from the Azinian award (NAFTA protection was sought for a fraudulently obtained concession contract with no Indication of governmental misconduct), that NAFTA jurisprudence was no longer "novel". It may be useful to cite here the approach of the Vivendi v. Argentina annulment committee commenting on and endorsing the cost decision of the original Vivendi tribunal (para 117, 118) which made each party bear its own expenditures:

"It observed that the dispute raised "a set of novel and complex issues not previously addressed in International arbitral precedent....... Moreover, Argentina was entitled to take the position it took, which Itself raised a difficult and novel question of public importance concerning ICSID and the operation of Investment protection agreements on the model of the BIT."

In the light of the importance of the arguments advanced by the parties in connection with this case, the Committee considers it appropriate that each party bear its own expenses Incurred..."

The issues at stake here - allocation of the risk of ambiguity of a governmental assurance, the scope of disclosure to government before obtaining a comfort letter - raise hitherto in BIT jurisprudence unsettled questions of how to apply the principle of "legitimate expectation" under Art. 1105 of the NAFTA. While there have been recently several awards on the question, no easily applicable doctrine of legitimate expectation under Art. 1105 of the NAFTA has so far evolved. It appears to me far from settled that one can dismiss easily a national treatment claim In a situation where the foreign Investor -having pursued a cooperative approach - is closed down Immediately and effectively, with no luck at all with domestic courts, while government enforcement is never successful with a politically well-connected domestic competitor. Also, the evidence of the government counsel huddling for over 13 hours discreetly with a senior judge in the case does at least raise questions of due process under International, if not national, standards. The tribunal may not have felt such procedural flaws in their aggregate reached the threshold of a breach of Art. 1105, but that is not to say that claimant made a frivolous claim In raising the issue of administrative denial of justice. I doubt that a reasonable, objective observer would have come at the outset to the conclusion that Thunderbird's claim was frivolous and manifestly unjustified.

Under these circumstances, I must conclude that there is no evidence, so far and on the record of the case, of a "manifestly spurious" claim. One could have thought to define "winning" by a relation of the Investor's claim (about 100+ M US $) to the outcome -zero in the award, about 500K US $ in my reckoning). But that would be a practice of civil law litigation not applicable here. This was a North-American Investor, Incorporated in Canada but largely run from the US and with a US approach. It (and its counsel) followed what seems to be the standard US practice of making arguably "excessive" monetary claims. But International tribunals have to be careful with cultural prejudices; NAFTA litigation is not meant to penalize claimant and its counsel for what is normal in their own jurisdiction, but which might be seen In other jurisdictions not as acceptable. A measure of cultural tolerance is required in transnational dispute resolution.

Another reason one could think of would be that the claim related to gambling, an Industry that is seen In many religious quarters as not very salubrious and provoking a moral opprobrium. But gambling, while usually heavily regulated, of the type at issue here (slot machines) is an entertainment activity that is widely practiced around the world²¹⁰ and can not be per se condemned as not worthy of Investment treaty protection. The next reason I can think of as providing a justification for this decision deviating significantly from standard Investment arbitration practice is the suspicion raised by the payment of the success fee - i.e. that possibly the two lawyers who arranged for the comfort letter may have shared that fee with Mexican officials. If it were so, I would have no hesitation whatsoever to penalize the claimant for daring to use an Investment treaty to protect the fruits of unethical behaviour. As I have argued earlier, I would not require full proof, but rather enough corroborating Indicators leading to a reversal of the burden of proof on claimant. But Mexico has not raised this suspicion openly, with substantiated argument and evidence and with a credible effort to bring its own officials who were involved in the comfort letter to the tribunal. In that situation, I consider that we have to disregard Insinuations of bribery if they are not properly raised, substantiated and open to a fair hearing. Otherwise, and with this award, a signal is sent out to respondent governments to Insinuate corruption as a standard defense technique; it is persuasive and effective, without having to stand up to the proper scrutiny of a full and proper litigation debate. The negative proof of non-corruption, is rarely possible. There is no foreign Investment where a close interaction, mostly with use of local consultants, with the government can be avoided. The whiff of corruption can therefore be made to appear In virtually any foreign Investment project.

To conclude: While I have sympathized with my colleagues' view not to find a breach of the principle of "legitimate expectation" under Art. 1105 of the NAFTA, I find no reason to reverse a well established NAFTA and ICSID jurisprudence consisting In letting each party, winning or losing, bear its own legal expenses and share the costs of arbitration short of clear evidence of either gross professional misconduct on the part of a party in arbitration or a manifestly spurious claim.

Annex: Decision on Costs in Arbitration rendered against the investor (investor-state) arbitration²¹¹