Abstract

This empirical analysis of the use of interpretive arguments by ad hoc tribunals of the International Centre for the Settlement of Investment Disputes covers almost 100 cases decided during the past 10 years. The cases are analysed with a view to determining which arguments the tribunals use and how the arguments are used in light of Articles 31 and 32 of the Vienna Convention on the Law of Treaties. The analysis provides a basis for addressing the extent to which ICSID tribunals contribute to creating a predictable legal framework in which the interests of investors, states, and third parties are taken properly into account; the extent to which ICSID tribunals contribute to a coherent development of international investment law; and whether ICSID tribunals contribute to a ‘fragmentation’ of international law. Despite ICSID tribunals being ad hoc tribunals that solve legal disputes on the basis of heterogeneous legal sources, the article indicates that there is a tendency among ICSID tribunals to contribute to a homogeneous development of the methodology of international law. Nevertheless, the article concludes that ICSID tribunals could do significantly more to align their approaches to interpretive arguments with those of other international tribunals.

Introduction

This article examines the use of interpretive arguments by ad hoc tribunals of the International Centre for the Settlement of Investment Disputes (ICSID). Tribunals use interpretive arguments in the process of determining the content of rules to be applied in specific cases. Such interpretive arguments can be distinguished from the sources of law listed in Article 38(a)–(c) of the Statute of the International Court of Justice (ICJ). The latter are the starting point or general framework for establishing the rules, while interpretive arguments are ‘subsidiary means for the determination of rules of law’: see Article 38(d) of the Statute of the ICJ.1

The empirical approach to ICSID tribunals’ use of interpretive arguments applied in the following consists in determining which arguments are explicitly used in the decisions of the tribunals,2 how the arguments were used, and how they relate to each other. The article will not analyse why certain arguments were preferred over other arguments or why arguments were used in one way and not in a different way in individual cases. The article will, however, analyse general trends in the use of arguments and explore possible reasons why the trends occur.3

On the basis of the analysis the article aims at responding to the following general questions: (1) to what extent do ICSID tribunals contribute to creating a predictable legal framework in which the interests of investors, states, and third parties are taken properly into account; (2) to what extent do ICSID tribunals contribute to a coherent development of international investment law; (3) do ICSID tribunals contribute to a ‘fragmentation’4 of international law?

The literature on international tribunals focuses in general on describing and analysing the main features of the tribunals, in particular their establishment, jurisdiction, procedure, the effects of their decisions, and their relationship to other international tribunals.5 Few authors have focused on the way in which tribunals use interpretive arguments in their legal reasoning. Those studies that address the reasoning of tribunals often do this in the form of a series of case studies,6 and through extensive analyses of the ‘real’ reasons for the tribunals’ interpretation.7 This study differs from such studies by identifying categories of interpretive arguments and by using these categories as a basis for investigating the extent to which arguments are used and how they are used.8

Since ICSID tribunals deal with disputes between investors and states, they may be faced with interpretive issues in relation to the following seven sets of rules (‘sources of law’): the ICSID Convention,9 multilateral investment treaties,10 bilateral investment treaties (including separate investment chapters in ‘economic integration agreements’),11 customary international law, general principles of law, specific agreements or decisions,12 and national legislation.13 The instrument referring the case to ICSID arbitration often determines the rules to be applied.14 Otherwise, Article 42(1) of the ICSID Convention applies: ‘the Tribunal shall apply the law of the Contracting State party to the dispute (including its rules on the conflict of laws) and such rules of international law as may be applicable’.15

The use of interpretive materials may differ significantly according to the category of rules to be interpreted. The focus of this article is on international law. Hence, it will not explore the approaches of tribunals when they interpret domestic legislation, administrative decisions, or contracts. Moreover, the article will not analyse the approaches of tribunals when they determine facts or apply the relevant rules to the facts.16

The analysis is limited to decisions by ICSID tribunals in the period between 1 January 1998 and 31 December 2006. There are three main reasons why the study is limited to this period. The first is that these decisions will give an updated picture of the interpretive approach taken by tribunals in the light of recent agreements referring cases to the ICSID, in particular the NAFTA and bilateral investment treaties, and in the light of the fact that some interpretive arguments, such as in particular decisions of other ICSID tribunals, will only gradually become available.17 Secondly, it is only recently that a high number of cases have been referred to dispute settlement before ICSID. Finally, to the extent that harmonization of the use of interpretive materials occurs under ICSID, we may see the result of such harmonization in recent cases rather than in early cases.

This article analyses 98 decisions in 72 different cases.18 Some decisions and cases during the period have been left out of the study, mainly because they were not made public, they essentially concerned issues of domestic law, or they did not raise issues of interpretation of international law.19 The 98 decisions examined in this study can be sorted into the following categories: 28 decisions on the merits, which in many cases also include decisions on jurisdictional and procedural issues, 54 decisions on jurisdiction,20 seven decisions on provisional measures,21 four decisions on other preliminary issues,22 one decision on interpretation of an award,23 and four decisions on annulment of awards.24

The ICSID Convention and related Rules set out a general framework for the tribunals’ decisions. According to Article 48(3) of the Convention and Rule 47(1)(i) of the Arbitration Rules, a decision shall deal with ‘every question submitted’ to the tribunal and shall set out ‘the reasons upon which’ the decision is based. Article 42 of the Convention states that a tribunal shall ‘decide a dispute in accordance with such rules of law as may be agreed by the parties’ to the dispute.25 These general statements do not give tribunals much guidance on how to approach interpretive issues. In light of the differences between the seven categories of rules that potentially may be applicable in individual cases, it is understandable that the drafters of the ICSID Convention preferred not to determine tribunals’ use of interpretive materials in more detail.

In principle, the rule or act referring a case to ICSID may set out how it shall use interpretive arguments, including whether the case shall be resolved ex aequo et bono.26 In practice, most such rules or acts do not set out how interpretive materials are to be used. However, some limited rules can be found in Article 102(2) of the NAFTA,27 Article 4 and Annex D of the ECT,28 and exceptionally in bilateral agreements.29

In light of the question concerning coherence of ICSID case law, it can be noted that decisions of ICSID tribunals are final, unless they are set aside by an annulment decision. Annulment is available only in exceptional cases: see Article 52 of the ICSID Convention and Arbitration Rule 50(1)(c)(iii). Case law has set a high threshold for annulment, and it has been emphasized that the annulment procedure should not be confused with an appeals procedure.30 Although there are some examples of decisions being annulled,31 the annulment procedure must be regarded only as an avenue for eliminating decisions based on clear violations of relevant rules, and not as an instrument for harmonizing the practice of ICSID tribunals. In contrast, some recent bilateral agreements contain clauses on the establishment of appeals procedures.32 It is not clear how such initiatives can be reconciled with the ICSID Convention: see Article 53(1).

Little has been written on the use of interpretive materials in ICSID tribunals. The main article is Schreuer's on ‘Diversity and Harmonization of Treaty Interpretation in Investment Arbitration’.33 Schreuer raises various issues relating to dispute settlement and discusses them in the light of a selection of cases. The present article goes through all relevant decisions during a limited time period and identifies how the tribunals resolve interpretive problems relating to all issues raised by the tribunals. Hence, while Schreuer's approach tends toward a selective and qualitative analysis, the approach here is more comprehensive and quantitative.

The Approach of the Study

In order to respond to the three general questions identified above, the following specific questions were asked relative to individual decisions:

  • 1 In relation to the decision generally:

    • a. To what extent did the tribunal follow the general rules on treaty interpretation set out in the Vienna Convention on the Law of Treaties (VCLT, 1969)?

    • b. Which general approach did the tribunal take to interpretive issues (i.e., did it follow an objective, subjective, or teleological approach, and did it make use of the principle of restrictive or effective treaty interpretation)?

  • 2 In relation to each interpretive issue raised in the decision:

    • a. Which interpretive arguments did the tribunal use?

    • b. What importance did the tribunal attribute to the argument?

When determining the extent to which ICSID tribunals contribute to creating a predictable legal framework and to developing international investment law, one may envisage a continuum between two extremes with regard to the use of interpretive arguments. At one extreme is a tribunal that strictly and solely focuses on solving the dispute. Such a tribunal would be interested only in the relationship between the parties to the dispute (a ‘dispute-oriented’ tribunal). At the other extreme is a tribunal that sees its role as comparable to that of a legislator (a ‘legislator-oriented’ tribunal). Tribunals may be placed on this continuum depending on the extent to which they take into account factors other than those strictly relevant to the relationship between the parties to the dispute. Hence, the more a tribunal, when interpreting relevant rules, takes into account factors such as those relating to the interests of third parties, the general functioning of the ICSID system, the potential impact of its reasoning or conclusions for future cases, the general need to clarify issues of law, or the need to prevent future disputes, the closer that tribunal can be placed toward the ‘legislator-oriented’ end of the continuum. Against this background, we may identify the following main factors which will be explored further in the analysis of the decisions:One objective in the following is to assess where ICSID tribunals in general can be placed on the continuum between ‘dispute-oriented tribunals’ and ‘legislator-oriented tribunals’. Already at this point, however, we may observe that the approaches of individual ICSID tribunals differ significantly.34

  • 1 Which interpretive arguments tribunals use: we may distinguish between tribunals that refer only to arguments presented by the parties to the dispute and tribunals that make a more independent assessment of available interpretive arguments. The more a tribunal restricts its arguments to those presented by the parties to the dispute, the more it can be regarded as ‘dispute-oriented’.

  • 2 How tribunals use the interpretive arguments: we may distinguish between tribunals that present only those arguments that support their conclusion and tribunals that present all relevant arguments and subsequently draw a conclusion on the basis of an overall assessment. The more a tribunal restricts its arguments to those strictly necessary to justify its conclusion, the more it can be regarded as ‘dispute-oriented’.

  • 3 The extent to which tribunals distinguish clearly between clarification of facts, interpretation of rules, and application of the rules to the facts: We may identify tribunals that make clear distinctions between these elements of the decision and tribunals that build on a more integrated and overall assessment. It can be argued that the more a tribunal bases its decision on an integrated and overall assessment, the more it can be regarded as ‘dispute-oriented’, while a tribunal that isolates the interpretive issues and deals with these independently of the facts of the case would be more ‘legislator-oriented’.

ICSID tribunals are ad hoc tribunals. There is thus a diversity of tribunals. Moreover, there are few common norms to guide their use of interpretive arguments. This means that it will be difficult to identify a general approach to interpretive materials that is sufficiently based on existing case law unless one assesses and takes into account all relevant cases. Hence, it can be argued that an analysis based on a quantitative approach has advantages over an analysis based on a qualitative assessment of a limited number of decisions, since the latter may easily be biased in favour of certain approaches to interpretive arguments. An analysis of all interpretive issues raised in all relevant decisions during a period of time will give a more nuanced and reliable contribution to the understanding of how ICSID tribunals have approached interpretive issues. Such an analysis will also provide a broad basis for assessing whether and how ICSID tribunals contribute to a predictable legal framework, to the development of international investment law, and to a ‘fragmentation’ of international law.

Contrary to the above considerations, it can be argued that tribunals may not always state all interpretive arguments that they take into account and how they use their arguments. While this is likely to be the case to some extent, every ICSID tribunal is under the obligation to ‘state the reasons upon which [its decision] is based’: see Article 48(3) of the ICSID Convention. Moreover, a decision may be annulled if it ‘has failed to state the reasons on which it is based’: see Article 52(1)(e). Hence, even if tribunals have a broad margin of appreciation with regard to how they use interpretive materials, they are under a clear obligation to state their reasons. Moreover, interpretive materials have been made easily accessible through numerous publications, and ICSID tribunals are assisted by experienced secretaries from the ICSID Secretariat. These factors make it easy for tribunals to use a broad range of interpretive arguments independently of how the parties to the dispute have presented their arguments.

The starting points for the classification of interpretive arguments in this article have been Article 38(d) of the Statute of the ICJ and Articles 31–33 of the VCLT. These starting points have been adjusted in order to reflect the categories of interpretive arguments actually used in the decisions of ICSID tribunals. Against this background, the following main categories of interpretive arguments will be analysed: the wording of the provision (section 5), the context (section 6), the object and purpose (section 7), customary international law (section 8), general principles of law (section 9), analogies and a contrario arguments (section 10), agreements between the parties (section 11), case law (section 12), state practice (section 13), preparatory work (section 14), legal doctrine (section 15), and reasonable results (section 16).35

The main focus of this article is on how ICSID tribunals address issues of treaty interpretation. As indicated above, customary international law and general principles of law may be used as interpretive arguments in relation to treaty provisions. Moreover, ICSID tribunals may apply customary international law and general principles of law directly as independent sources of law. How ICSID tribunals approach customary international law and general principles of law will therefore be addressed briefly in section 3. Thereafter in section 4 there follows an analysis of the general approaches ICSID tribunals have taken to treaty interpretation.

The large number of issues to be addressed and the high number of decisions to be analysed pose a problem concerning how the findings can be documented without making the article too long. I have decided to proceed as follows: all cases are referred to with short names and complete references can be found in the Annex; long lists of cases in footnotes are omitted while shorter lists of cases are included due to an assumption that the latter will illustrate deviations from general trends or exceptions to general rules;36 long lists of cases are included where such lists are considered to be of particular interest; footnotes are omitted where relevant information can easily be obtained through footnotes in the neighbouring text.

Approaches to Customary International Law and General Principles of Law

Introductory Remarks

This section presents an overview of the approaches of ICSID tribunals to customary international law and general principles of law. The objective is to analyse which rules of customary international law or general principles of law the tribunals use, and how they argue with regard to the existence and content of such rules.

ICSID tribunals may be faced with the application of customary international law and general principles of law where there is a reference to such rules in a relevant treaty,37 where the treaty does not address the issue in question, i.e., where there is a legal lacuna in the treaty,38 and where customary international law replaces a clause in a treaty (i.e., it develops after the treaty has been concluded).39 The only reference to customary international law and general principles of law under the ICSID Convention is the instruction that tribunals shall apply ‘such rules of international law as may be applicable’ where the parties to the dispute have failed to agree on the rules to be applied: see Article 42(1).40 It is quite common to include clauses referring to customary international law in relevant treaties.41

There is no generally established distinction between customary international law and general principles of law. In the following, the concept ‘general principles of law’ will be used as referring to principles of law derived from national legal systems, and which do not have the status of customary international law.42 As it is often difficult to determine whether a norm has attained the status of customary international law, the distinction will be based on how tribunals have classified the norms in question. However, there are examples where one tribunal has characterized a norm as customary international law while another tribunal has characterized the same or a very similar norm as a general principle of law.43

Approaches to Customary International Law

The tribunals used customary international law as a separate legal basis in 34 of the 98 decisions. Their use of such rules seemed to depend heavily on the arguments of the parties to the dispute. Customary law was discussed in relation to a broad range of issues, including jurisdictional issues,44 procedural issues,45 substantive issues,46 and issues concerning so-called ‘secondary rules of international law’.47

ICSID tribunals generally based their findings with regard to the existence and content of rules of customary international law on references to case law from the ICJ, the Permanent Court of International Justice, and arbitral tribunals, references to treaties, in particular the VCLT, references to documents adopted by the International Law Commission, and references to the legal doctrine. No tribunal made its own assessment of whether a rule of customary international law existed, and only exceptionally did tribunals explicitly address questions concerning opinio juris.48 Some tribunals analysed the content of rules of customary international law based on several sources.49

Where there is a reference to customary international law in a treaty, it can be asked whether tribunals are to apply customary international law as it was at the time of conclusion of the treaty or at the time of the dispute. Statements in case law seem to indicate that tribunals generally preferred customary international law at the time of the dispute. In Técnicas v. Mexico, where the tribunal used customary international law to clarify the concept of indirect expropriation, it applied ‘customary international law, not as frozen in time, but in [its] evolution’.50 It thus seems appropriate to assume that unless there is a clear indication that one shall apply customary law as it was at the time of the adoption of the treaty, for example in the preparatory work, a tribunal will apply current customary international law.

Approaches to General Principles of Law

General principles of law are a source of law that plays a marginal role in most areas of public international law. However, such principles could be expected to play a significant role in international investment law. One reason is that there is a close substantive relationship between public international law, private international law, and domestic law in relation to international investments. Moreover, ICSID tribunals often have competence to make decisions in accordance with international law, domestic law, and contractual obligations simultaneously. It may be difficult or even impossible to distinguish clearly between these legal bases in a given decision.51 General principles of law can thus be a ‘common denominator’ for and function as a bridge between these three sources of law.

Contrary to what was expected, the tribunals examined general principles of law as a separate legal basis in only eight of the 98 decisions. Their use of such principles seemed to depend entirely on the arguments of the parties to the dispute. General principles were discussed in relation to jurisdictional issues,52 procedural issues,53 and issues concerning ‘secondary rules of international law’.54 Tribunals based their arguments concerning the existence and content of the principles only on general references to a principle in three decisions, references to a previous ICSID tribunal in three decisions, and broader assessments of the existence and content of the principles in three decisions.55 Hence, there was no discussion of the existence or content of the principle in most cases.

Some Concluding Remarks

ICSID tribunals applied customary international law and general principles of law quite frequently in their decisions. In general, the arguments presented did not contribute substantially to solving questions concerning the existence of such rules. The main contribution of the decisions was to clarify the content of the rules. However, few tribunals took upon themselves the task of contributing to such clarification. In general, tribunals made use of a limited range of arguments concerning the existence and content of customary international law and general principles of law. These findings show that ICSID tribunals in general have a significant potential to improve their reasoning relative to customary international law and general principles of law.

General Rules on Treaty Interpretation

This section will address how ICSID tribunals make use of the Statute of the ICJ and the VCLT when addressing interpretive issues. The focus will in particular be on the extent to which ICSID tribunals follow general rules on treaty interpretation as set out in the VCLT and applied by the ICJ, or whether they rather follow a different approach to treaty interpretation, i.e., an approach that moves them in the direction of a ‘self-contained regime’.56

The main argument that can be advanced in favour of the position that ICSID tribunals are likely to pursue a ‘self-contained regime’ approach would be that investment treaties have certain characteristics that distinguish them from other treaties. Such characteristics could be:

  • 1 the investor–state dispute settlement mechanism (the mechanism is available both to investors and to states57),

  • 2 the treaties essentially concern issues related to domestic decision-making,

  • 3 the treaties are of a ‘contractual’ nature (as opposed to ‘law-making treaties’), and

  • 4 ICSID tribunals are established ad hoc.

Each of these characteristics is, however, shared with other international regimes. The investor–state dispute settlement and the relationship to domestic decision-making are shared with human rights treaties, while the contractual nature of the treaties and ad hoc establishment of tribunals are shared with the WTO Agreement. Nevertheless, it can be argued that no other regime shares all the characteristics of ICSID (except, of course, other regimes for the settlement of investment disputes). Against this background, one could perhaps expect ICSID to develop in the direction of a self-contained regime. However, the ICSID Convention itself indicates that the negotiators envisaged ICSID as a part of the general system of public international law, since it includes a right for a state to refer differences with other states concerning interpretation of the ICSID Convention to the ICJ: see Article 64.58

References to Articles 31–33 of the VCLT were found in 35 of the 98 decisions.59 There is an increase in references toward the end of the period examined.60 A clear majority of the references were limited to Article 31(1) of the VCLT. Only in 16 decisions did tribunals extend their references beyond this provision. References to Article 38 of the Statute of the ICJ were found in only six decisions.

If we look beyond the mere references to the VCLT and the ICJ Statute and assess whether the tribunals actually made active use of the instruments in their reasoning, we can find elements of such application of Articles 31–33 of the VCLT in 20 decisions,61 and of Article 38 of the ICJ Statute in three decisions.62 The application of the instruments was in general very brief and they were used only as general arguments in support of the tribunals’ approaches in almost all decisions. Only in exceptional decisions did tribunals integrate the VCLT into their reasoning beyond general references63 or make any link between the VCLT and case law of the ICJ.64

The above findings indicate that a conclusion that ICSID tribunals in general follow an approach that includes important elements of a ‘self-contained regime’ cannot be rejected. ICSID decisions can thus be distinguished from panel and Appellate Body decisions under the WTO, since the latter to a significant degree integrate the VCLT and relevant ICJ decisions. However, on the basis of the above findings it is not possible to conclude firmly as to the degree to which ICSID should be classified as a self-contained regime. Hence, any such conclusion will have to await the analysis below of how ICSID tribunals use interpretive arguments in practice.

General Approaches to Treaty Interpretation

The Use of Obiter Dicta

ICSID tribunals have a general obligation to ‘deal with every question submitted’ to them and to ‘state the reasons upon which’ they base their decisions: see Article 48(3) of the ICSID Convention. The general picture is that ICSID tribunals made an effort to address all arguments raised by the parties to the dispute. But this does not prevent tribunals from exercising judicial restraint by avoiding dealing with issues that can be left aside as a consequence of conclusions on other issues.

ICSID tribunals differed quite significantly as to their willingness to pronounce obiter dicta. While many tribunals made general statements concerning issues of law that in many cases were unnecessary or weakly related to the facts of the case,65 a few tribunals were reluctant to pronounce on ‘hypothetical’ issues.66 Tribunals frequently responded explicitly and in detail to all arguments made by the parties to the dispute, and many tribunals seemed unwilling to let go of an opportunity to contribute to the development of international investment law.

Despite the general willingness to pronounce obiter dicta, ICSID tribunals made explicit statements concerning their general approach to treaty interpretation in a few cases. When they did express themselves, they normally addressed issues relating to the application of the VCLT or the Statute of the ICJ. Only in exceptional cases did tribunals base their approaches to interpretive issues on references to the legal doctrine.67 The assessment below of whether tribunals favoured objective, subjective, or teleological approaches to treaty interpretation and how they approached the principles of effective and restrictive interpretation of treaties is therefore in essence based on how tribunals applied interpretive arguments in practice.

Objective, Subjective, and Teleological Approaches to Treaty Interpretation

It is common to distinguish between objective, subjective, and teleological approaches to treaty interpretation.68 Recourse to one approach does not rule out the use of other approaches. Indeed, as indicated in Article 31(1) of the VCLT, tribunals may choose to apply a combination of the three approaches.69 What is of interest here is whether we can see a general tendency in favour of one of the approaches in ICSID case law.

ICSID tribunals in general indicated a clear preference for the objective approach to treaty interpretation. In almost all cases where a preference for one of the approaches could be traced, the preference was in favour of the objective approach.70 In many cases, the objective approach was based on a reference to Article 31(1) of the VCLT.

In some cases, the objective approach was implemented by mentioning different language versions of the provision to be interpreted71 or by using arguments based on the linguistic meaning of terms according to dictionaries.72 However, relatively few tribunals made active use of the wording in their reasoning in the sense that they identified possible alternative interpretations. Hence, even if tribunals in general indicated a clear preference for the objective approach, this was only to a limited extent reflected in practice as an analysis of the exact wording of provisions was infrequent in the cases examined.

Some tribunals made use of a teleological approach. This was generally done in relation to specific questions of interpretation by invoking ‘object and purpose’ as an interpretive argument, and in some cases after noting that no clear conclusion could be made on the basis of the text itself.73 Section 7 below will investigate in more detail the use of ‘object and purpose’ as an interpretive argument. The general impression is that the teleological approach was subsidiary to the objective approach.74

Many authors emphasize the importance of identifying the intention of the parties in the process of treaty interpretation.75 Moreover, it can be argued that international investment law, due to its emphasis on bilateralism and contractual issues, is an area in which subjective approaches to treaty interpretation can be expected to be of particular importance. However, ICSID tribunals had a focus on identifying the ‘intentions’ of the parties to the treaties in only 15 of the 98 decisions.76 It was often difficult to distinguish clearly between instances in which tribunals opted for a teleological approach and instances where they preferred a subjective approach.77 The intention of the parties was most often used as a specific argument against a proposed interpretation, for example by stating that ‘it cannot have been the intention of the parties’. To the extent that tribunals justified their statements concerning the intention of parties, they mostly referred to the preparatory work.

Even if a recent decision may seem to indicate a preference for a subjective approach,78 the general impression is that the subjective approach is subsidiary to the objective and often difficult to distinguish clearly from the teleological approach. Hence, the conclusion is that ICSID tribunals in general preferred an objective approach to treaty interpretation. Subjective and teleological approaches were subsidiary to the objective approach, and in general limited to specific interpretive issues.

The Principles of Effectiveness and Restrictiveness79

The distinction between the principles of ‘effectiveness’ and ‘restrictiveness’ is common in the context of treaty interpretation. The principle of effectiveness is closely related to the teleological approach to treaty interpretation in the sense that the principle can be formulated as favouring the interpretation that would most effectively fulfil the objectives of a provision or a treaty. The principle of restrictiveness would favour the interpretation that best protects the sovereignty of the parties to the treaty.

We may distinguish between two versions of the principle of effectiveness. One version is the broad version formulated above. The other is a narrow and specific version related to the use of effectiveness as a specific argument based on the presumption that a provision of a treaty shall not be interpreted in a way that makes other provisions superfluous or meaningless. This latter version of the principle is often referred to as effet utile,80 and this term will be used in the following.

There were few tribunals that made an explicit or implicit link between the general principle of effectiveness and a teleological approach to treaty interpretation.81 In contrast, there were examples where tribunals refused to use the principle.82 Hence, there was not much support among the tribunals for using the principle of effectiveness as an extension or reinforcement of a teleological approach to treaty interpretation.

ICSID tribunals made quite frequent use of effet utile arguments. They almost always used such arguments in order to reject interpretations that would make specific provisions or the treaty useless.83 Hence, there was support in case law for effet utile arguments, and such arguments were given significant weight in a number of cases.84

Case law seemed to be somewhat divided with regard to the principle of restrictiveness. While some tribunals made statements in favour of the principle, in particular in the context of interpreting exceptions to main rules,85 other tribunals rejected arguments based on the principle.86 Arguments based on restrictiveness were regarded as relevant only when tribunals were interpreting exceptions. However, it may frequently be unclear whether a narrow interpretation of an exception results from a principle of restrictiveness or a principle of effectiveness, since the result, while favouring a restrictive interpretation of the exception, will in general also favour an ‘effective’ application of the main rule. In sum, the decisions did not in general favour the principle of restrictiveness.

Finally, it may be asked whether there is a trend to be identified in the use of the principles of effectiveness and restrictiveness. In addition to the specific application of the principles as described above, there are some recent decisions that express the general position that they do not support any of the principles or that they would favour a ‘balanced approach’.87 The general tendency seems to be in the direction of disregarding arguments based on the principles, while retaining the possibility of using arguments based on effet utile in specific cases.

The Context

Article 31(2) of the VCLT defines the ‘context’ as follows:

The context for the purpose of the interpretation of a treaty shall comprise, in addition to the text, including its preamble and annexes:

  • (a) any agreement relating to the treaty which was made between all the parties in connection with the conclusion of the treaty;

  • (b) any instrument which was made by one or more parties in connection with the conclusion of the treaty and accepted by the other parties as an instrument related to the treaty.

This is a narrow definition of the context, and it can be regarded as a ‘lowest common denominator’ for the core contextual arguments. In the following, we will assess both the extent to which ICSID tribunals made use of contextual arguments that fell within the scope of the VCLT definition, and to what extent they made use of other contextual arguments.

The main reasons for using the context in which a phrase occurs as an interpretive argument are to avoid inconsistencies in the text and to ensure that provisions are mutually supportive.88 Contextual arguments are thus closely related to interpretive arguments based on object and purpose (see infra, section 7), since both arguments generally aim at making provisions mutually supportive in order to fulfil the same objects and purposes. Contextual arguments are also closely related to arguments based on preserving the effet utile of a provision (see supra, section 5C) and a contrario arguments (see infra, section 10)89 as these arguments generally serve to ensure consistency and effectiveness within the treaty.

In the following, a distinction will be drawn between contextual arguments and the three arguments mentioned above. The most important consequence is that arguments explicitly based on the object and purpose will not be addressed here, but in section 7. Contextual arguments must also be distinguished from a contextual application of provisions of a treaty. Hence, the application of definitions and explicit references between provisions of a treaty will not to be regarded as contextual arguments. Contextual arguments were in general closely related to a broad range of other interpretive arguments, and it was often difficult to distinguish contextual from other arguments.

A phrase to be interpreted is surrounded by ‘layers’ which constitute the context of the phrase. The innermost layer is the specific rule to which the phrase belongs. There were only four decisions in which ICSID tribunals made use of this layer of contextual arguments.90 A second layer is the remainder of the provision within which the rule occurs. There were five decisions in which such contextual arguments were made.91 Contextual arguments were, not surprisingly, most common in the third layer of contextual arguments, namely other provisions of the treaty, where such arguments were found in approximately one third of the 98 decisions examined. The arguments made ranged from invoking the specific relationship between two provisions92 to arguments based on the location of a provision in the treaty93 and consistency in the use of terminology throughout the treaty.94 Hence, a broad range of arguments was made within this layer of contextual arguments, ranging from arguments that clear inconsistencies should be avoided to assumptions that a specific logic was followed when the treaty was drafted95 or that certain provisions indicated a trend, for example in the direction of increased transparency.96 If we go beyond the main body of the treaty to a fourth contextual layer, other parts of the treaty, 12 decisions were identified in which contextual arguments were based on the preamble (i.e., exclusive references to the object and purpose), annexes, or appendices to the treaties, or directly related rules and regulations97 (such as the Arbitration Rules: see Article 31(2)(b) of the VCLT).98

If we broaden the context beyond the definition in Article 31(2) of the VCLT, we may identify contextual arguments based on the relationship between the ICSID Convention and relevant BITs. However, contextual arguments were generally insignificant in the three decisions in which such arguments were made.99 There are also examples of contextual arguments that are based on the relationship between ICSID and alternative arbitration mechanisms such as UNCITRAL.100 Contextual arguments beyond those mentioned above will be addressed where relevant below.101

Altogether, contextual arguments were identified in 49 of the 98 decisions. In order further to analyse the use of the arguments, we may broadly distinguish between three groups of cases, namely those where:

  • 1 contextual arguments were used as a starting point for the subsequent legal reasoning,

  • 2 contextual arguments were used as an essential interpretive argument, and

  • 3 contextual arguments were used as non-essential interpretive arguments.

These distinctions will be used throughout the rest of this article. Here, ‘starting point’ is defined as general statements at the start of a legal reasoning that set the framework for the subsequent analysis.102 ‘Essential arguments’ refers to arguments that constitute an important factor in the subsequent analysis.103 An essential argument is not necessarily in support of the final conclusion. ‘Non-essential arguments’ refers to arguments that are added after it is clear what the conclusion will be, whether in support of the conclusion, to confirm the conclusion, or to indicate why other conclusions should be rejected.104 It has not been necessary for the purpose of this article to establish more precise definitions. The purpose of the distinctions is to serve as a tool to establish general assessments of how arguments have been used.

Contextual arguments were used as general starting points for the subsequent analysis in nine decisions, as essential arguments in 38 decisions, and as non-essential arguments in seven decisions. This means that clearly the most important function of contextual arguments was as essential arguments, a finding that is in harmony with the general role envisaged for such arguments in Article 31(2) of the VCLT. It also corresponds well with the finding that most of the contextual arguments were derived from relevant provisions of the treaty in question, and that there were few instances where contextual arguments were derived from other parts of the treaty or other relevant instruments.

In conclusion, contextual arguments were used quite frequently, but not as frequently as could be expected when taking into account the role attributed to such arguments in Article 31 of the VCLT. This is in line with the finding above that relatively few ICSID tribunals made active use of the wording in their reasoning. Hence, there is in general a potential for ICSID tribunals to make more active use of the treaties to be applied, both through analyses of the wording of the rules to be applied and through more thorough analyses of their context.

Object and Purpose

Section 5B concluded that even if ICSID tribunals did not in general indicate a preference for a teleological approach to treaty interpretation, they frequently referred to object and purpose as an interpretive argument. Object and purpose was used as an interpretive argument in 48 of the 98 decisions. The following three questions will be addressed in order further to clarify ICSID tribunals’ use of object and purpose: (1) from where did tribunals derive the object and purpose; (2) how did tribunals define the object and purpose; and (3) how was object and purpose used as an interpretive argument?

(1) Object and purpose can be derived from a number of sources. Schreuer found that ‘[t]he most frequent way to find a treaty's object and purpose was to look at the preamble’.105 However, this was true only to the extent that tribunals actually indicated whence they derived object and purpose. In a clear majority of decisions, the tribunals did not refer to any source for their statements concerning the object and purpose.106 The lack of identification of sources may indicate that tribunals were of the opinion that the object and purpose were evident and that no reference was needed, or that tribunals merely based their arguments on their own opinion concerning the object and purpose.107 It was not possible to identify which of these causes was most important.

Another important source for identifying the object and purpose was provisions of the treaties.108 Some tribunals also referred to literature, to the title of the treaty, and to case law in order to determine the object and purpose.109 Only three decisions referred to the preparatory work. All these references were to the preparatory work of the ICSID Convention. Hence, even if the preparatory work of the ICSID Convention has been easily accessible, it was invoked in only exceptional cases, and the preparatory work of BITs or the NAFTA was not mentioned when determining the object and purpose of these treaties.

(2) ICSID tribunals relied on the object and purpose of the treaty as such in a clear majority of decisions.110 In significantly fewer instances they referred to the more specific object or purpose of a chapter of a treaty, a selection of provisions of a treaty, or a specific provision.111 Some tribunals referred to a broader version of object and purpose, for example by drawing on the object and purpose of the ICSID Convention when interpreting provisions of a BIT.112

(3) It was difficult to distinguish clearly between the different ways in which tribunals made use of the object and purpose. The main reason was that tribunals often simply referred to the argument without explaining explicitly how it was used. Only in exceptional instances did tribunals address in more general terms the importance they attributed to the object and purpose.113 Many tribunals referred to the object and purpose as a general starting point for their interpretive process.114 Where tribunals actively used object and purpose as an interpretive argument, they used it as an essential argument in a clear majority of the decisions.115

There were a few decisions in which tribunals used object and purpose indirectly in their argument. Most of these were decisions in which tribunals made arguments based on what clearly fell outside the purpose of the treaty.116 In addition, tribunals made assessments of whether their findings would be contrary to the object and purpose of the treaty.117 Arguments that a provision's or a treaty's effet utile must be preserved were addressed in section 5C above.

The above analysis indicates that tribunals that made use of object and purpose as an interpretive argument were inclined to make specific use of it as an essential and sometimes decisive interpretive argument. One tribunal indicated that the object and purpose could prevail over the wording of the treaty.118 However, this was an exceptional case where other factors pointed in the same direction. The object and purpose would, under normal circumstances, be an important interpretive argument where the wording of the relevant treaty provision is unclear. Hence, the present findings confirm the conclusion in section 5B that ICSID tribunals primarily use an objective approach to treaty interpretation, and that a teleological approach is generally used as a supplement. The extent to which ICSID tribunals found it unnecessary to indicate how they established the object and purpose is remarkable. It is the opinion of this author that ICSID tribunals should be expected to indicate how they establish the object and purpose they invoke.

Customary International Law

In the following, we will focus on how customary law has been used as an interpretive argument.119 We must distinguish between interpretive arguments based on customary international law as such and those based on the constitutive elements of customary international law, in particular various forms of state practice. To the extent that interpretive arguments are based on the constitutive elements of customary international law, these will be addressed where relevant below.120

Customary international law can in most cases be regarded as general law which countries may codify, specify, or derogate from through treaties. Consequently, arguments based on customary international law would typically be available where a treaty makes use of general concepts or principles, in areas where customary international law is well established and defined, and with a view to avoiding or minimizing possible inconsistencies between customary international law and the treaty.121 In most treaties, there is no specific rule on the relationship between the treaty and customary international law. However, the concept ‘international law’ as used in Article 31(3)(c) of the VCLT must be assumed to cover customary international law. Moreover, Article 102(2) of NAFTA requires the parties to interpret and apply the provisions of NAFTA ‘in accordance with applicable rules of international law’.122

ICSID tribunals resorted to customary international law as an interpretive argument in 24 of the 98 decisions examined.123 Such arguments were quite frequently used in the context of both jurisdictional and procedural issues124 and substantive issues.125 The tribunals based their arguments on a thorough analysis of the status or content of the customary international law in only seven decisions. They made a summary analysis of the rules in eight decisions, and included no analysis of the rules in 13 decisions. Customary international law was used as a general starting point for the subsequent reasoning in six decisions, it was used as an essential argument in 14 decisions,126 and as a non-essential argument in 11 decisions.

In addition to general references to rules of customary international law, ICSID tribunals also referred to such rules as they appear in the VCLT and the Statute of the ICJ. The use of the VCLT and the Statute of the ICJ as interpretive arguments comes in addition to the use of the instruments when determining the general approach to the use of interpretive arguments: see section 4 above. Provisions of the VCLT were used as a basis for interpretive arguments in seven decisions,127 while provisions of the Statute of the ICJ were used in three decisions.128

When taken together with the use of customary international law as a separate legal basis,129 customary international law emerges as a significant element in the decisions of ICSID tribunals. It is also worth noting the broad variety of legal issues for which customary international law has been regarded as relevant and important by ICSID tribunals. On the other hand, their use of customary international law as an interpretive argument supports the above conclusion that ICSID tribunals have significant potential to improve their reasoning relative to clarifying the status and content of customary international law.

General Principles of Law Recognized by Civilized Nations

Tribunals made use of general principles of law130 as an interpretive argument in only four decisions.131 Hence, even if international investment law is an area of law where there are close links between domestic and international law and where general principles derived from domestic legal systems may be expected to be of importance to the interpretation and application of international law, there were few instances where such principles had practical significance. Moreover, there was no thorough analysis of the content of the principles in any of the decisions, and the principles were not used as essential interpretive arguments, but rather as non-essential arguments or as general starting points for the subsequent analysis. In the light of the general expectations identified in section 3C above, it is surprising that ICSID tribunals paid so little attention to general principles of law as an interpretive argument.

Analogies and A Contrario Arguments

Two interpretive arguments found in many domestic legal systems are analogies and a contrario arguments. Direct references to analogies are not very common among ICSID tribunals.132 The main reason seems to be that arguments which could be regarded as analogies in domestic law would in most cases be classified differently in international law, in particular as arguments based on state practice or case law from other regimes. Examples are arguments based on how an issue has been dealt with in other BITs, or on how the WTO dispute settlement mechanism has dealt with an issue. Hence, analogies will be addressed under the relevant headings below, in particular in sections 12C and 13B.

A contrario arguments,133 on the other hand, are arguments that are not easily classified under other headings. Hence, these will be addressed here. When discussing the use of a contrario arguments in investment disputes, Schreuer observed that:134

The problem with the expressio unius principle is not so much a lack of consistency of the tribunals but its limited usefulness. Whether the mention of one item or a list of items in a provision really excludes the relevance of other items depends very much on the particular circumstances and cannot be answered in a generalized way. Similarly, the question whether a provision in one treaty may be taken as proof that another treaty that lacks such a provision was meant to exclude the effects of the provision is difficult to answer in a generalized way with the tools of abstract logic.

This scepticism of a contrario arguments received some support in Plama v. Bulgaria (award) where the tribunal refused to accept such an argument.135 However, there were only two more tribunals that explicitly rejected a contrario arguments.136 Active use of such arguments could be found in 21 of the 98 decisions. One tribunal noted:137

This a contrario interpretation of one of the items on an enumerative list, even one that is not exhaustive, is fully in keeping with the logic and the spirit of a BIT and is equivalent to an interpretation ‘in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose,’ in accordance with Article 31(1) of the Vienna Convention on the Law of Treaties of 1969, which codifies a rule of customary international law.

Two factors seem to be important for ICSID tribunals’ use of a contrario arguments, namely the extent to which the provision to be interpreted sets out a specific and detailed rule,138 and whether there are circumstances indicating that the negotiators considered whether or not to include a specific issue in the rule in question.139 While five tribunals related their a contrario arguments to items not being included in exhaustive or non-exhaustive lists, a majority of the tribunals related their argument to other issues, such as the choice of wording and solutions found in other treaties.140 The tribunals used a contrario arguments as essential arguments in 11 decisions,141 and as non-essential arguments in 10 decisions.142 Against this background, it can be concluded that a contrario arguments have been significant for ICSID tribunals despite the arguments that have been raised against their use. This may be a result of the close link between a contrario arguments, the objective approach to treaty interpretation, and contextual arguments, the latter having a strong standing among ICSID tribunals.

Agreement between the Parties to Treaties

Introduction

Article 31 of the VCLT distinguishes between two forms of agreement between the parties to a treaty, namely agreement made in connection with the conclusion of the treaty, which is part of the ‘context’ (paragraph 2),143 and subsequent agreement regarding the interpretation of the treaty, which come in addition to the ‘context’ (paragraph 3(a)). It can be asked whether and to what extent these provisions cover agreements other than those that are legally binding. The wording of the provisions is unclear, since ‘agreement’ can be interpreted narrowly as only legally binding instruments or broadly as covering also non-binding instruments, including ‘soft law’ instruments, adopted by the parties. The International Law Commission stated in its commentaries that:144

it is well settled that when an agreement as to the interpretation of a provision is established as having been reached before or at the time of the conclusion of the treaty, it is to be regarded as forming part of the treaty. Thus, in the Ambatielos case the Court said: ‘… the provisions of the Declaration are in the nature of an interpretation clause, and, as such, should be regarded as an integral part of the Treaty …’. Similarly, an agreement as to the interpretation of a provision reached after the conclusion of the treaty represents an authentic interpretation by the parties which must be read into the treaty for purposes of its interpretation.

This statement indicates that the intention of the International Law Commission was that the term ‘agreement’ should be read quite narrowly. Subsequent commentators seem to confirm such a reading of the term.145 Hence, while a document which is agreed among the parties to the treaty and is intended as an authoritative interpretation of the treaty must be regarded as an ‘agreement’, a guideline adopted in order only to facilitate the implementation of a treaty would not be covered. To the extent that agreements fall within the scope of Article 31, they are regarded as primary interpretive arguments. According to the International Law Commission, ‘these categories of documents should not be treated as mere evidence to which recourse may be had for the purpose of resolving an ambiguity or obscurity, but as part of the context for the purpose of arriving at the ordinary meaning of the terms of the treaty’.146 Instruments that fall outside the scope of Article 31(2) and (3)(a) will be addressed in section 13 below on state practice.

The extent to which there are institutional mechanisms relating to treaties is significant for the likelihood that agreements will be adopted. Moreover, it will normally be easier to adopt agreements the fewer contracting parties are involved.147 These two factors can be expected to play an important role for the extent to which ‘agreements’ in the sense of Article 31(2) and (3) are adopted under the treaties of interest here.

In order for there to be an agreement, there must be evidence of communication between the parties of a nature that makes it possible to conclude that there is a mutual agreement between them. Hence, the coincidence of unilateral statements does not constitute an agreement unless the statements are related to each other and there is evidence that the parties intended their statements to constitute an agreement.148

In the following, we will examine which instruments may qualify as agreements under Article 31 of the VCLT and how they have been used in practice. We will distinguish according to the treaty to be applied, i.e., the ICSID Convention, bilateral investment treaties, and regional instruments (the NAFTA and the ECT).

The ICSID Convention

ICSID's decision-making is regulated in Articles 6 and 7 of the ICSID Convention. The extent to which there are procedural requirements that must be fulfilled before an ‘agreement’ is established under ICSID remains unclear.149 Some documents that have been adopted in the context of the ICSID Convention and that may have an impact on the interpretation of the Convention will not be addressed in the following, in particular the ICSID Rules and Regulations adopted in accordance with Article 6(1) of the ICSID Convention.150

One document that may possibly constitute an agreement in accordance with Article 31(2)(a) of the VCLT is the Report of the Executive Directors on the Convention on the Settlement of Investment Disputes between States and Nationals of Other States (Report of the Executive Directors).151 This Report was presented to governments by the negotiators together with the result of the negotiations. Paragraph 14 of the Report sets out that the ‘provisions of the attached Convention are for the most part self-explanatory. Brief comment on a few principal features may, however, be useful to member governments in their consideration of the Convention.’ The comments can thus be regarded as representing the common understanding of the negotiators. The Report could accordingly be regarded either as part of the preparatory work of the Convention or as an ‘agreement’ in accordance with Article 31(2)(a) of the VCLT. Whether it is regarded as one or the other is important for its role as interpretive argument.

The way in which the report was established indicates that it should be regarded as part of the preparatory work to the ICSID Convention. Although it is not common for negotiators to present a set of comments together with the results of the negotiations, it is not unheard of.152 Moreover, the fact that the Report was not subject to formal adoption by the parties to the Convention points in the same direction. Even if the Report was issued as an explanation to governments when they considered whether to join the Convention, and as such is closely related to the document to which states consented, it does not seem appropriate to expect states who were opposed to statements in the Report to enter formal reservations or to work for a general understanding that the Report should not be regarded as generally accepted.153 Although parts of the Report set out definitions or explanations of terms and rules of the Convention,154 most of the Report does not include statements of relevance to the interpretation of the Convention.

On the other hand, the Report does not give an account of what took place during the negotiations. It is also noted that ICSID presents the Report as if it were part of the documents to which parties agreed.155 Moreover, the way that parts of the Report were used by ICSID tribunals indicates that it was not regarded by them as preparatory work in the sense of Article 32 of the VCLT. The Report was used as an interpretive argument in 26 of the 98 decisions, and none of the tribunals dealt with the Report as a ‘supplementary means of interpretation’.156 This indicates that the Report has attained a status that differs from ordinary preparatory work.157 Against this background, even if the Report as a whole cannot be regarded as an ‘agreement’ under Article 31 of the VCLT, certain parts of it have been used by tribunals in such a way that it seems more appropriate to deal with them here than in the context of the preparatory work. Hence, for the purpose of this analysis, the Report is regarded as an ‘agreement’ in accordance with Article 31(2)(a) of the VCLT.

Another document to be considered here is the ICSID Model Clauses.158 These Clauses include a number of comments that may be invoked as interpretive arguments.159 The question is whether these Clauses and their commentaries can be seen as a ‘subsequent agreement between the parties regarding the interpretation of the treaty or the application of its provisions’: see Article 31(3)(a) of the VCLT. It can be argued that, while the Clauses were not adopted as interpretations of the ICSID Convention, they were at least to some extent adopted with a view to facilitating the application of its provisions. However, despite the fact that ICSID tribunals have made use of a broad range of interpretive arguments, there was no decision that referred to the Model Clauses. Hence, even if it might be possible to argue that they should be regarded as an ‘agreement’ in accordance with Article 31(3)(a) of the VCLT, it is unlikely that they will be significant for the interpretation of the ICSID Convention.

Finally, it may be mentioned that the ICSID Secretariat at one point prepared explanatory notes that were considered by the Council and regarded as sufficiently useful to be published together with the text of the rules. However, these explanatory notes were not formally adopted and have not played any significant role in case law.160 Moreover, they are no longer published together with the text of the rules. Hence, they cannot be regarded as a subsequent agreement in accordance with Article 31(3)(a) of the VCLT.

Bilateral Investment Treaties

A number of bilateral treaties include a variety of instruments regarding the interpretation of provisions of the treaties. Most of these instruments form integral parts of the agreement, such as footnotes to provisions, annexes, protocols, and even exchange of letters,161 and would thus not be of relevance here.162 Other instruments are adopted as memoranda of understanding and decisions. Most of these instruments are adopted simultaneously with the treaty: see Article 31(2) of the VCLT. However, some instruments are adopted subsequently and may thus fall under Article 31(3)(a).163 Some treaties, in particular ‘economic integration agreements’, set up institutional mechanisms that may facilitate the adoption of such instruments.164 Moreover, in many cases it would be easy to establish common agreement between the parties concerning the interpretation of relevant provisions due to the limited number of parties to bilateral treaties.

One concern with subsequent agreements is that they may reduce the transparency of the treaties, in the sense that investors cannot trust that the treaty constitutes the final regulatory framework. Another concern is that such agreements may undermine domestic constitutional and democratic procedures. Hence, if an agreement were to go against the ordinary meaning of the terms used in the treaty, or would otherwise add to or detract from the obligations set out in the treaty, it can be argued that the appropriate approach would be to negotiate amendments to the treaty.165 While such an option may not be available in practice under multilateral treaties, it would normally be easily available under bilateral treaties.

There was only one case in which agreement between the parties was discussed as an interpretive argument.166 Hence, while some such agreements may exist, their numbers seem to be low and they have been insignificant in case law. This situation may change due to developments in case law under the agreements and the recent proliferation of institutional mechanisms under BITs and economic integration agreements.

Multilateral Investment Treaties

It is common to adopt instruments in the context of multilateral treaties, and many such instruments could qualify as ‘agreements’ under Article 31(2)(a) and (3)(a) of the VCLT. Examples of agreements adopted in connection with the conclusion of a treaty include the decisions adopted at the European Energy Charter Conference.167 An example of a subsequent agreement can be found in the Notes of Interpretation issued by the Free Trade Commission (FTC) under the NAFTA.168 The latter addressed, inter alia, the interpretation of Article 1105 of NAFTA in the light of minimum standards of treatment under customary international law. This interpretation was extensively discussed in subsequent case law, in particular as regards the applicability of the interpretation to disputes that were initiated before it was adopted.169 There were no other references to agreements under multilateral treaties in the decisions examined.

Concluding Remarks

Against this background, we may conclude that while there is significant potential for establishing agreements under investment treaties, few such agreements have been established in practice. The agreements that have been established have in general played a limited role in case law. The party invoking an agreement will have the burden of proving the existence of a relevant agreement, and it seems that tribunals are likely to require formal proof. Once it is established that an agreement exists, it is likely to be regarded as a decisive, or at least as an essential, interpretive argument.

Case Law

General Issues

The ICSID Convention and the Arbitration Rules contain no provision on the relationship between decisions of different tribunals or on the use of case law as an interpretive argument. Article 38(1)(d) of the Statute of the ICJ identifies ‘judicial decisions’, together with ‘the teachings of the most highly qualified publicists of the various nations’ as ‘subsidiary means for the determination of rules of law’. According to Article 59 of the Statute, ICJ decisions are binding only in the relation between the parties and only in the case in question. Even if the Statute of the ICJ is not directly applicable to ICSID tribunals and has not received much attention in case law under ICSID,170 these rules have a basic standing in international law and should therefore serve as a starting point for this analysis.

The main purpose of giving international investors access to ICSID tribunals is to offer them a dispute settlement system that is more effective and predictable than the courts they would otherwise be faced with in host countries. There is no way to appeal decisions of ICSID tribunals. Hence, from the perspective of investors, a main priority for ICSID tribunals should be to ensure consistency and predictability. As Schreuer has pointed out:171

Reliance on past decisions is a fundamental feature of any orderly decision process. Drawing on the experience of past decisions plays an important role in securing the necessary uniformity and stability of the law. The need for a coherent case law is evident. It strengthens the predictability of decisions …

Against this background, one objective is to assess the extent to which consistency and predictability have been ensured in the case law of the ICSID. It will also be of interest to examine to what extent ICSID tribunals follow the lead of other international tribunals, or whether they rather contribute to a ‘fragmentation’ of international tribunals.

There is no support in international law for asserting that decisions of tribunals have direct legal effects beyond the cases in question: see Article 59 of the Statute of the ICJ. A number of ICSID decisions contain general statements with regard to their use of case law as interpretive arguments. One representative example can be found in ADC v. Hungary:172

The Parties to the present case have also debated the relevance of international case law relating to expropriation. It is true that arbitral awards do not constitute binding precedent. It is also true that a number of cases are fact-driven and that the findings in those cases cannot be transposed in and of themselves to other cases. It is further true that a number of cases are based on treaties that differ from the present BIT in certain respects. However, cautious reliance on certain principles developed in a number of those cases, as persuasive authority, may advance the body of law, which in turn may serve predictability in the interest of both investors and host States.

Nevertheless, many ICSID tribunals have pointed out that the issues under each case must be determined on their own merits and that the tribunals remain free to deviate from previous case law.173 One tribunal has argued that there is no good reason for allowing the first tribunal in time to resolve issues for later tribunals.174 However, this argument has not received explicit support in subsequent case law.175

The tribunals specified the cases from which an interpretive argument was derived in almost all instances.176 Tribunals generally referred only to decisions that were publicly available.177 Tribunals conducted extensive analyses of available case law in some instances.178 However, this was the exception rather than the general rule. Tribunals in general did not distinguish explicitly according to whether statements in earlier cases were ratio decidendi or obiter dicta. Even in cases where they noted the distinction and classified a statement as an obiter dictum, they did not seem to attach any particular significance to this finding.179 Hence, ICSID decisions seem in general to attach the same significance to obiter dicta as to ratio decidendi.

Case law was used as an interpretive argument in 92 out of 98 decisions.180 Hence, the focus will not be on the extent to which ICSID tribunals made use of case law, but rather on what kind of case law was used and how it was used. Section 12B below will address how ICSID tribunals took into account the decisions of other ICSID tribunals. There follows an examination of the extent to which and how ICSID tribunals used decisions from other investment tribunals (section 12C), and other international courts and tribunals (section 12D). The relationship between ICSID tribunals and domestic courts will be examined as part of state practice in section 13C.

ICSID Decisions

Case law from ICSID was used as an interpretive argument in 90 of the 98 decisions. Hence, this was by far the most widely used and most important interpretive argument. Due to the high number of instances in which tribunals referred to ICSID case law, most decisions contained references to ICSID case law in relation to several interpretive issues, a quantitative approach such as the one used for other interpretive arguments has been inappropriate. Hence, a more qualitative analysis has been applied to this category of arguments.

The first issue to be addressed is how ICSID tribunals used such decisions in the process of interpretation. The following uses of case law could be identified in the decisions:

  • references to how other tribunals approached similar issues (e.g., a ‘test’ applied by a former tribunal),181

  • references to how other tribunals used interpretive arguments (e.g., a reference to preparatory work may be justified by previous tribunals having used the same documents in a similar manner),182

  • references to the reasoning of other tribunals,183

  • references to the conclusions of other tribunals (e.g., the preference of one interpretation over an alternative interpretation may be justified by referring to the fact that other tribunals have expressed the same preference),184

  • references to case law in general as supportive arguments (i.e., without specifying which elements of the cases or even which cases),185

  • references to case law in order to establish a general starting point for the subsequent reasoning,186

  • references to case law as an a contrario argument (in particular in the form of ‘distinguishing the case’),187 and

  • references to case law as an analogy argument (i.e., the general argument that similar cases should have the same conclusion).188

These uses corresponded in essence with what could be expected and with what can generally be observed in other tribunals. The most interesting aspect was the diversity of approaches found in the decisions. Such diversity was hardly surprising given the ad hoc nature of ICSID tribunals.

The importance attributed to ICSID case law varied from using it as one of several supportive arguments after the tribunal has reached its conclusion189 to being a decisive argument for reaching the conclusion.190 It was quite common for tribunals to use case law as a means to establish a presumption in favour of one result, and thus for placing a burden of proof on one of the parties.191

The importance of ICSID decisions as an interpretive argument depends on a broad range of factors, of which some are case specific (e.g., differences and similarities between the facts of the cases) and some are of a more general nature (e.g., the availability of other interpretive arguments). Hence, the importance has to be determined on a case-by-case basis. Against this background, it is of interest to identify the general factors that are significant for determining the importance of ICSID case law as an interpretive argument.

It can be asked whether certain ICSID decisions have been identified as ‘leading’ cases and thus been given a general status as particularly important interpretive arguments.192 A decision may become a leading case through references by other tribunals, by being frequently invoked by parties to disputes, by being frequently referred to by states or international institutions, through references in the legal doctrine, and by being the first case to deal with a specific issue.193 There are several examples where tribunals emphasize the quality of a decision when determining whether or not to make use of it as an interpretive argument. This is done both when a tribunal finds another tribunal's reasoning particularly convincing,194 and when a tribunal finds that a case cannot be taken into account due to omissions in the reasoning.195 The general impression from ICSID case law is that it is difficult to identify cases that have generally been regarded as leading cases. In most instances, it is possible to find tribunals that have disregarded or criticized decisions which other tribunals have regarded as particularly important.196

It can also be asked how tribunals act where longstanding and consistent case law favours one interpretation over another. On the one hand, some interpretive issues that have been raised have found their solution through consistent case law, and are in general no longer raised by the parties to the disputes. If a party to a dispute chooses to argue that the interpretive issue must be solved differently in such instances, a tribunal must be expected to demand clear and convincing arguments for distinguishing the case. Hence, most tribunals accept a strong presumption in favour of following longstanding and consistent case law.197 On the other hand, certain interpretive issues remain controversial despite being addressed by numerous tribunals. The interpretation of most favoured nation and umbrella clauses is the main example. The divergence in case law on such issues is illustrative of the diversity of approaches that can be found in ICSID case law.198

Between these two extremes there are a number of situations in which tribunals may face arguments in favour of complying with or deviating from previous case law. As there is a limited system for review and annulment of decisions of ICSID tribunals, see Section 5 of the ICSID Convention,199 one may expect a need for flexibility to deviate from previous case law. This argument is reinforced by the ad hoc nature of ICSID tribunals, as ad hoc tribunals’ approaches to interpretive issues are likely to vary more than those of permanent courts.200 Against this background, it was not surprising that some tribunals emphasized their general independence from previous case law.201 Moreover, many ICSID tribunals were willing to deviate explicitly from previous case law. Schreuer has identified four specific situations in which he found that ‘tribunals sitting in different cases have come to conflicting conclusions on identical questions’.202 That many ICSID tribunals explicitly deviated from previous case law was of particular interest in light of the fact that most tribunals may instead choose to ‘distinguish the case’. Even if tribunals often made an effort to distinguish their cases from previous cases,203 the extent to which ICSID tribunals in general felt free to criticize and deviate from the findings in previous case law was remarkable.

Another issue is whether there were differences as to how ICSID tribunals used previous case law depending on the rules subject to interpretation. First, it can be asked whether there were differences where decisions concerned jurisdictional or procedural issues and where they concerned substantive issues. It can be argued that since the ICSID Convention is the main basis for decisions on jurisdictional and procedural issues, while a broad range of other rules are the basis for the substantive issues, one might expect previous case law to be most prominent as an interpretive argument in relation to jurisdictional and procedural issues.204 Taken together with the broad margin of appreciation applied by ICSID tribunals when determining the importance of previous case law, this could indicate that we should expect a higher degree of consistency for jurisdictional and procedural issues than for substantive issues. However, the case law gave no basis for pointing to any significant differences in this respect.

Secondly, it can be asked whether there were differences according to the substantive rules to be applied. It could be argued that, due to the inherent institutional and regulatory differences between the ICSID Convention, BITs, and multilateral investment agreements (the NAFTA and the ECT), one could expect differences in how existing case law was used. The only significant difference found was the frequent use of UNCITRAL arbitration under NAFTA: see section 12C below. Hence, ICSID case law may seem to be somewhat less significant in relative terms under NAFTA than under the ICSID Convention and BITs.205

Investment Tribunals

For the purposes of this article, it is useful to divide investment tribunals other than ICSID tribunals into three main groups, namely UNCITRAL tribunals,206 the Iran–US Claims Tribunal,207 and other investment tribunals. The ‘other investment tribunals’ category covers a broad range of tribunals, including a number of claims commissions. UNCITRAL decisions were used as interpretive arguments in 30 decisions, Iran–US Claims Tribunal decisions were used in 22 decisions, and other investment tribunals’ decisions were used in 30 decisions. Many of the other investment tribunals’ decisions were old cases dating back to the first half of the 20th century. ICSID tribunals in general did not seem to differentiate between cases according to their age.

Figure 1 distinguishes between instances in which tribunals used the case law as a starting point for their interpretation, as a non-essential interpretive argument, and as an essential interpretive argument. These categories are not precisely defined, and are thus used only to generate a general picture of whether the use of the three categories of case law differed significantly.208 The figure shows the percentage of decisions209 that used the case law as indicated.

Figure 1.

Case law as interpretive argument.

Figure 1.

Case law as interpretive argument.

Case law from these tribunals was in general more often used as a non-essential interpretive argument than as an essential one. Moreover, the only significant differences between the categories was that tribunals less often used case law of ‘other investment tribunals’ as non-essential arguments, and that they more often used such case law as starting points for their interpretation. As these cases were often old, the findings indicate that ICSID tribunals quite frequently used an approach to interpretive issues based on the historical development of international investment law.

It is also of interest to examine for which purposes the tribunals used the case law. The categories used here are interpretation of jurisdictional and procedural rules, interpretation of substantive rules, and interpretive methodology. While it is difficult to draw an exact distinction between these categories, the findings may nevertheless indicate certain general tendencies. Figure 2 shows the percentage of decisions that used case law in relation to the topics as indicated.

Figure 2.

Interpretive issues clarified through references to case law.

Figure 2.

Interpretive issues clarified through references to case law.

Hence, ICSID tribunals almost never referred to case law from other investment tribunals when determining their interpretive methodology.210 Moreover, the figure indicates that ICSID tribunals to a large extent made use of UNCITRAL case law in relation to procedural and jurisdictional issues, and mainly used case law from the Iran–US Claims Tribunal and other investment tribunals in relation to substantive issues. The frequent use of UNCITRAL case law in the context of jurisdictional and procedural matters can be explained by the similar functions and procedures of UNCITRAL and ICSID tribunals.

ICSID tribunals referred to case law under ICSID far more often than case law from other tribunals. However, there were no decisions in which they explicitly distinguished in favour of case law originating in ICSID tribunals, and there were few decisions that implicitly indicated a preference for ICSID case law.211 Hence, ICSID tribunals were most likely to use interpretive arguments from ICSID case law, but when they used interpretive arguments from case law of other investment tribunals they attributed the same importance to such case law as to case law from ICSID tribunals.212

Other International Courts and Tribunals

Many courts and tribunals other than those mentioned above occasionally deal with investment issues. Hence, their case law may be directly relevant for interpretive issues addressed in ICSID cases. Moreover, case law from such courts and tribunals may be of relevance in order to establish the appropriate interpretive methodology to be applied by ICSID tribunals. Finally, when ICSID tribunals establish the content of principles or rules of customary international law, they may rely on decisions from such other courts and tribunals.

References to decisions of the ICJ and its predecessor, the Permanent Court of International Justice, were found in 46 decisions.213 These decisions were used to clarify jurisdictional and procedural issues in 30 decisions, substantive issues in 19 decisions, and issues of interpretive methodology in six decisions (including three decisions relating to burden of proof). This distribution is strikingly similar to ICSID tribunals’ use of UNCITRAL case law.214 Hence, ICSID tribunals used references to ICJ and UNCITRAL case law in relation to the same general issues. However, while UNCITRAL case law was mostly used to interpret treaty provisions, ICJ case law was often used to determine rules of customary international law.

Moreover, it is remarkable that there were so few references to ICJ case law in relation to interpretive methodology. This reinforces the above conclusion that ICSID tribunals in general have not entered into detailed discussions of their approaches to treaty interpretation: see above, sections 4 and 5.

ICJ case law was used as an essential argument in 31 decisions,215 as a non-essential argument in 16 decisions, and as a general starting point for interpretation in 23 decisions. The tribunals did not take into account whether the decisions were unanimous or contained dissenting opinions.216 In some instances, ICSID tribunals found it necessary to distinguish their cases from the case law of the ICJ.217 Against this background, we may observe that ICSID tribunals generally attributed more importance to ICJ case law than to case law from other tribunals, and that they much more frequently used ICJ case law as a starting point for interpretation.

In some cases where the ICJ and ICSID tribunals dealt with the same issues, ICSID tribunals took ICJ decisions as the starting point, and followed up with an analysis of the case law of ICSID tribunals. The main examples were issues that are not regulated in detail in the ICSID Convention or investment treaties, in particular general jurisdictional issues and state responsibility.218 In some exceptional cases ICSID tribunals justified their reliance on ICJ jurisprudence on parallels between the ICSID Convention and the Statute of the ICJ.219

ICSID tribunals rarely referred to case law of the WTO. Such references appeared in only five decisions, mainly in relation to procedural issues.220 Case law from human rights tribunals was mainly, but not exclusively, employed in the context of expropriation and compensation.221 ICSID tribunals used WTO and human rights case law as both essential and non-essential arguments. ICSID tribunals also made individual references to case law from the European Court of Justice, the International Tribunal of the Law of the Sea, and the Tribunal of Justice for Andean Community of Nations.222 In general, it seems that ICSID tribunals’ use of such case law mainly depended on whether the parties to the disputes invoked it. Hence, there was no indication that ICSID tribunals systematically referred to case law from these tribunals in order to ensure consistency or predictability. In general, it seemed that ICSID tribunals were willing to consider any decision invoked by parties to a dispute, regardless of which international court made the decision. If the decision was relevant, the tribunal would take it into account.

Conclusions

The above analysis makes it possible to suggest where ICSID tribunals should be placed on the continuum between ‘dispute-oriented’ and ‘legislator-oriented’ tribunals: see Section 2. Factors indicating that ICSID tribunals should be placed close to dispute-oriented tribunals include the reliance of many ICSID tribunals on case law invoked by the parties to the dispute223 and the willingness of tribunals to deviate from previous case law. Factors indicating that ICSID tribunals should be placed close to legislator-oriented tribunals include the extensive use of case law from a broad range of tribunals,224 the fact that many tribunals discussed case law that did not support their conclusion, and the importance attributed to case law from the ICJ. ICSID tribunals’ use of previous case law in their reasoning did in general contribute to consistency and predictability, and thus to the development of international investment law, even if a number of tribunals stressed their independence from previous case law. Moreover, the extent to which ICSID tribunals considered and took into account case law from other tribunals, in particular the ICJ, indicates that most ICSID tribunals showed a general willingness to follow the lead of other international tribunals and to contribute to the general development of international law.

Against this background, it seems appropriate to conclude that ICSID tribunals’ use of case law indicates that ICSID tribunals are located closer to ‘legislator-oriented’ tribunals than to ‘dispute-oriented’ tribunals. Moreover, even if there is a significant minority of ICSID tribunals that made limited or no use of case law from other tribunals, it does not seem appropriate to conclude that ICSID tribunals in general contribute to a ‘fragmentation’ of international law or international tribunals. These findings indicate that ICSID tribunals are far from establishing a ‘self-contained regime’.

Practice of States and International Institutions

General Issues

The term ‘state practice’ was rarely used by the ICSID tribunals,225 and they have thus not established a standard use of the term. ‘State practice’ may cover a broad range of interpretive arguments. If taken in the meaning used when referring to ‘general practice’ as one of the constitutive elements of customary international law, such as in Article 38(b) of the Statute of the ICJ, it refers to factors ranging from domestic court and administrative decisions to provisions in bilateral, multilateral, and global treaties. In addition, ‘state practice’ may refer to analogies or a contrario arguments based on individual acts and to acts that indicate the opinio juris of a state or a group of states. One special form of state practice is referred to in Article 31(3)(b) of the VCLT, namely ‘any subsequent practice in the application of the treaty which establishes the agreement of the parties regarding its interpretation’. A broad definition of ‘state practice’ will be used in what follows. However, acts of a respondent state of specific relevance to the case in question, indicating whether the state in that case has accepted or relied upon a specific interpretation of the term to be interpreted,226 will normally not constitute ‘state practice’ in the sense used here. Nevertheless, if it is argued that such an act confirms a general pattern of state practice in similar situations, it could constitute part of ‘state practice’.227

State practice must be distinguished from arguments based on subsequent agreement between the parties to the dispute: see section 11 above. Only when such agreements are invoked as part of a general practice of states will they be of relevance here. In addition to the practice of states, the practice of international institutions may be of relevance. While ‘state practice’ may include decisions made by organs of international institutions composed of their states parties, we may define ‘institutional practice’ as decisions or documents issued by organs of the institutions that enjoy a degree of independence from the states parties, in particular secretariats and compliance committees. As there was no significant use of institutional practice in the ICSID tribunals’ case law, we shall not pursue such practice further here.228 State practice must also be distinguished from acts adopted by non-governmental institutions. Of particular potential interest would be documents adopted and issued by business and industrial organizations or by international standardization organizations. However, the use of arguments based on acts by non-governmental institutions has been insignificant in ICSID case law, and such interpretive arguments will not be addressed further in this article.229

It was often difficult to distinguish the use of state practice in relation to the interpretation of treaties from the use of state practice in the context of customary international law or general principles of law, in particular where the interpretive problem related to both treaty law and customary international law or general principles of law.230 Where the use of state practice is clearly related to the application of rules of customary international law or general principles of law, it falls under sections 3, 8 and 9 above and will not be addressed here.

State practice was used as an interpretive argument in 52 of the 98 decisions. In general, tribunals neither explicitly nor implicitly classified state practice as a principal or supplementary means of interpretation: see Articles 31–32 of the VCLT.231 References of a general nature, i.e., references which did not specify the kind of state practice relied upon, were found in 14 decisions.232

The following sections will examine decisions that referred to specific kinds of state. State practice that occurred at an inter-state level in the form of treaties and decisions of international institutions will be addressed in section 13B. Thereafter, section 13C will address state practice in the form of unilateral acts, including acts relating to the implementation of treaties.

State Practice at the Inter-state Level

Bilateral investment treaties were the state practice at the inter-state level that was most frequently used as an interpretive argument. Such treaties were used in 28 decisions,233 while tribunals referred to other investment-related instruments in 14 decisions234 and to instruments that were not investment related in nine decisions.235 In general, the instruments were much more often used as arguments in relation to jurisdictional and procedural issues (38 decisions) than in relation to substantive issues (12 decisions). There was no significant difference between the categories of treaties in this respect. It is worth noting that only three decisions made reference to the WTO regime despite the close link between this and the investment regime, in particular between the GATS and the TRIMs Agreement and international investment law.236 There was no decision in which the tribunal made use of state practice in the form referred to in Article 31(3)(b) of the VCLT.237

State practice was more often used as a non-essential argument than as an essential one. BITs were more often used as essential arguments than the other treaties,238 and they were also more often used as general starting points for the interpretation than were the other treaties. Hence, it seems that tribunals in general attributed greater importance to arguments based on BITs than to arguments based on other treaties.239

Treaties were quite frequently used as a basis for a contrario arguments,240 in the sense that a rule found in treaty practice was used as an argument against interpreting a treaty that did not contain such a rule to include the rule.241 Hence, it may seem that tribunals in general were more supportive of a contrario arguments than of analogies.242 There were tribunals that warned against the use of a contrario arguments.243

Unilateral State Practice

We can distinguish between three ways in which unilateral acts of states are used by tribunals: as part of the facts of the case,244 as part of the determination of the rights and duties of the parties to the dispute,245 and as interpretive arguments in relation to treaty provisions, customary international law, or general principles of law. It is this third use of unilateral acts that is referred to as ‘unilateral state practice’ in the following.

Unilateral state practice was used for interpretive purposes in 30 of the 98 decisions. The following categories of practice were identified:

  • Model investment treaties (five decisions);246

  • Instruments or proceedings concerning the implementation of obligations under investment treaties (seven decisions);247

  • Arguments presented in proceedings for international tribunals (five decisions);248

  • Other unilateral practice – legislative (three decisions), executive (four decisions), judicial (13 decisions).249

The relatively high number of references to domestic court decisions as compared to the other categories of unilateral practice is remarkable. It may indicate a trend in ICSID tribunals to be more case law oriented than legislation oriented. Moreover, all the decisions using model investment treaties as an interpretive argument referred to the US Model Treaty.250

One may ask whether arguments presented in proceedings of international tribunals should be regarded as unilateral state practice. It has been argued that ‘[c]ounsel representing the State in arbitration proceedings have the duty to put forward all the arguments they deem appropriate to defend their position, but a tribunal could not presume that each of those arguments constitutes the expression of a unilateral act that obligates the State’.251 A contrary argument is that states can be expected to maintain a degree of consistency in how they interpret their obligations under treaties. Against this background, it can be argued that statements that are highly case specific, typically where a state is defending itself against a claim, should not be taken into account, while more general statements, typically those made by states as third parties to disputes,252 may be taken into account.

Unilateral state practice was often closely related to the case or dispute in question. In many instances, it may thus be better to characterize such acts as expressions of legal opinion than as expressions of practice. While inter-state practice was sometimes used as a generalized practice followed over a period of time, unilateral practice was rarely used in a generalized form.

The purposes for which ICSID tribunals used unilateral state practice did not differ significantly from their use of multilateral state practice. Both were used extensively in relation to jurisdictional and procedural issues. Moreover, the importance attributed to the argument did not differ significantly from that attributed to multilateral state practice. Hence, differences in the nature of inter-state and unilateral practice did not seem to have any major consequences for the way in which the tribunals used the argument.

Some Concluding Remarks

The category of state practice referred to in Article 31(3)(b) of the VCLT was absent in the decisions examined here. State practice as applied by the ICSID tribunals was probably their most heterogeneous category of interpretive argument, both in terms of the broad variety of acts that was regarded as relevant, and in terms of how such arguments were used by the tribunals. The use of generalized inter-state practice may contribute to the general development of international investment law and indicate that tribunals should be regarded as ‘legislator-oriented’. However, few ICSID tribunals made any extensive assessment of generalized state practice. Moreover, they made significant use of unilateral state practice. Hence, the ICSID tribunals’ use of state practice indicates that they can be regarded as more ‘dispute-oriented’ than ‘legislator-oriented’.

Preparatory Work to the Treaty and the Circumstances of Its Conclusion

Preparatory work covers a broad range of events prior to the adoption or entry into force of a treaty,253 including the general context in which the negotiations took place.254 It has sometimes been asked whether tribunals should regard as preparatory work statements by governments at the time of the conclusion of the agreement or in the subsequent domestic process of signing, ratifying, or implementing the treaty.255 While the former may possibly be regarded as part of the preparatory work,256 it seems appropriate to regard the latter as state practice rather than as part of the preparatory work, since the statements were made in a domestic setting and they were thus not subject to any exchange of opinions.

Preparatory work was used as an interpretive argument in 25 of the 98 decisions.257 There were few cases in which references to preparatory work were specific and discussed in detail.258 In relative terms, preparatory work was most often used in relation to NAFTA and less frequently used in relation to BITs.259 Comments in case law indicated that preparatory work to BITs would have been taken into account more often had it been available.260 Hence, the use of preparatory work seems to reflect its availability.261

Preparatory work was almost exclusively used in the context of jurisdictional and procedural issues. It was very rarely used for the interpretation of substantive provisions.262 Contrary to what was expected, there were few examples where preparatory work was used to clarify the object and purpose of the treaty or provision.263 It was more common for tribunals to use it to identify the states’ and drafters’ intentions.264 Only exceptionally was preparatory work used as an a contrario argument in the sense that the omission of an issue in the preparatory work indicated that a certain meaning had not been intended.265

In accordance with Article 32 of the VCLT, the preparatory work to a treaty may be used as a ‘supplementary’ means of interpretation ‘in order to confirm the meaning resulting from the application of article 31, or to determine the meaning when the interpretation according to article 31: (a) leaves the meaning ambiguous or obscure; or (b) leads to a result which is manifestly absurd or unreasonable’. There were few tribunals that emphasized the ‘supplementary’ nature of the preparatory work.266 On the contrary, tribunals frequently resorted to preparatory work as the starting point for their analysis or as an essential argument, although it was most often used as a non-essential argument.267 These findings support the conclusion that ‘resort to travaux préparatoires seems to be determined less by their position among the canons of interpretation than by their availability’.268 This conclusion seems valid, not only for the extent to which preparatory work is applied, but also for how it is applied. Such use of preparatory work may favour the views of certain countries over those of others due to differences in their ability to participate in negotiations and the ability of their administrations to document their positions and statements during negotiations.

Legal Doctrine

Article 38(d) of the Statute of the ICJ refers to ‘the teachings of the most highly qualified publicists’ together with ‘judicial decisions’ as a ‘subsidiary means for the determination of rules of law’. Legal doctrine is not explicitly referred to as a relevant interpretive argument in the VCLT, but it can be assumed to be covered by Article 32 as a supplementary means of interpretation.

Legal doctrine was used as a source of information about interpretive arguments in a number of decisions.269 In other instances, decisions referred to legal doctrine as a reference source for readers needing additional information.270 Such references are not relevant here, and must be distinguished from the use of the legal doctrine's analyses of, conclusions on, or points of view on interpretive questions.271 Some instances where legal experts were interviewed as witnesses and where their statements were taken into account as interpretive arguments were counted as ‘legal doctrine’.272

Legal doctrine was used as interpretive argument in 73 of the 98 decisions. This made it the second most frequently used interpretive argument, second only to ICSID case law. References to legal doctrine were almost as frequent in relation to substantive issues as in relation to jurisdictional and procedural issues.273 Hence, when compared to their use of other interpretive arguments, tribunals in general made more frequent use of legal doctrine in relation to substantive issues. The tribunals referred to legal doctrine in relation to interpretive methodology only in exceptional instances.274 Legal doctrine was used extensively in relation to specific questions of treaty interpretation and questions concerning rules of customary international law.

The tribunals referred to books or articles written by individual experts in almost all instances. Notable exceptions were a few references to the Restatement of the Law Third, The Foreign Relations Law of the United States,275 some references to publications from international institutions,276 and some instances of unspecified references.277 The clear dominance of individual authors over other types of publications was remarkable.

Legal doctrine was used as an essential interpretive argument in a majority of the cases in which the argument occurred.278 Moreover, it was used to establish a starting point for the legal analysis in more than half the decisions in which it was used.279 The tribunals frequently used doctrine to establish or define ‘tests’ to be used as conditions for applying rules.280

These findings indicate that legal doctrine in general was regarded by ICSID tribunals as one of the most important interpretive arguments. One reason could be that there is an extensive legal doctrine of a particularly high quality in the field of international investment law. Another reason could be that the interpretation of many of the provisions of the treaties is no longer a contentious issue, and that many tribunals find it sufficient to refer to the doctrine in order to justify their interpretation. However, even a cursory comparison of ICSID tribunals’ use of legal doctrine with that of the ICJ and the Appellate Body of the WTO, which make much less use of legal doctrine,281 indicates that these are not the reasons. This leaves us with the fact that there are relatively few experts on international investment law, and that these experts to a large extent are both main contributors to the literature in the field and main participants in arbitration proceedings. Moreover, many of the people serving on tribunals are professors of law, and they are thus raised in a tradition in which references to literature are highly internalized. These seem to be the main reasons why legal doctrine plays such an important role in the case law of ICSID tribunals.

ICSID tribunals’ active use of legal doctrine must be regarded as an essential element in their contribution to the development of international investment law, and is thus an argument in favour of regarding them as ‘legislator-oriented’. Their focus on literature specific to international investment law indicates that they have an attitude which may contribute to the fragmentation of international law.

Reasonable Results

Article 32(b) of the VCLT shows that reasonableness can be used as an argument in order to set aside an interpretation that would lead to ‘a result which is manifestly absurd or unreasonable’ on the basis of supplementary means of interpretation. There was no decision that explicitly applied reasonableness in such a context. The VCLT does not address the potential use of reasonableness as an interpretive argument in other contexts. Nevertheless, numerous authors have considered issues such as equity and public policy to be relevant to treaty interpretation.282

Many ICSID tribunals regarded it as important to arrive at conclusions based on interpretations that were reasonable. This was in most cases expressed as a need to avoid interpretations that would generally be regarded as unreasonable. A number of tribunals made statements that emphasized the need to arrive at a result that was reasonable in the case in question. Such specific reasonableness was in general related to the process of applying the rule to the facts in question rather than to the process of interpretation, and is thus not relevant for our purposes.283 A third group of statements is those where considerations of reasonableness were explicit elements of the rule to be applied. In these cases, reasonableness was related to the process of applying, and not interpreting, the rule.284 These cases are thus not relevant here.

Reasonableness is closely linked to two groups of interpretive arguments dealt with earlier in this article. First, there is a close link between reasonableness and teleological arguments, the main difference being that while the object and purpose are specified in the treaty or derived from its related interpretive materials, reasonableness is general considerations relates to the consequences of an interpretation.285 This distinction may, however, be difficult to apply in practice.286 Secondly, reasonableness may be closely related to contextual arguments. The main difference between these arguments is that contextual arguments take into account the somewhat narrow legal context of the rule and reasonableness takes into account a much broader legal and political context. Nevertheless, it may occasionally be difficult to distinguish clearly between the two.287

Keeping these distinctions in mind, it was remarkable that arguments based on reasonableness could be identified in 36 of the 98 decisions.288 Arguments based on reasonableness were mainly related to jurisdictional and procedural issues and less frequently related to substantive issues.289 Given the distribution of decisions, this difference was not significant.

Reasonableness was far more frequently used as an essential argument than as a non-essential one.290 The main explanation seems to be that reasonableness was invoked most frequently to reject interpretations other than those preferred by the tribunals and to illustrate potential negative consequences of not adopting the preferred interpretations. Only exceptionally was there any thorough assessment of the reasonableness of an interpretation, for example in the form of an explicit assessment of advantages and disadvantages.291

A focus on a reasonable solution in the individual case indicates a ‘dispute-oriented’ attitude, while a focus on finding an interpretation that would be reasonable should it be applied in future cases indicates a ‘legislator-oriented’ attitude. The way ICSID tribunals used reasonableness in their decisions pointed in the direction of the latter rather than the former. While the focus of the tribunals on reasonableness as a means to reject alternative interpretation may strengthen the argument that they should be regarded as ‘legislator-oriented’, the lack of thorough assessments of the reasonableness weakens this argument.

Concluding Remarks

Some Remarks on Potential Weaknesses of the Study

Before we conclude this study, some potential weaknesses of the empirical approach used should be discussed. One such weakness could be that not all decisions during the period have been available for assessment. There were 15 decisions that could not be included in the study.292 This is a fairly low number compared to the 98 decisions covered. Moreover, as these decisions are unavailable, their role as part of the developing case law of ICSID has been limited. Finally, one reason many of the decisions were unavailable was the delay in the publication of decisions. Against this background, the problem relating to the unavailability of decisions seems to be insignificant.

Another potential weakness could be the limited time span of the study. It can be argued that such a study should cover all ICSID cases, not just those from 1998 to 2006, and that 1998 seems to be a randomly selected year. When the study was designed, the primary objective was to give a description of how ICSID functions in its current legal environment. This environment includes recent multilateral instruments, such as the NAFTA and the Energy Charter Treaty, the high number of recent bilateral investment treaties, the increasing number of bilateral economic integration agreements containing investment chapters, and an increasing volume of foreign direct investment. Moreover, while the early decisions of ICSID tribunals reflected legal uncertainties in the regime's start-up phase, we may expect recent decisions to reflect a more settled legal tradition. The year 1998 was chosen in order to include a sufficiently high number of decisions to be a basis for the drawing of conclusions with some degree of certainty. In sum, there seems to be no significant problem relating to limiting the selection of decisions to those adopted after 1998.

A third potential weakness relates to the fact that the study focuses on decisions and not on cases (which frequently consist of more than one decision). It can be argued that this will lead to the over-representation of the approaches of tribunals that split their cases into several decisions.293 However, ICSID dispute resolution is organized in a way that in general contributes to the likelihood that certain perspectives or views may be over-represented, since many arbitrators participate in numerous cases,294 and tribunals are supported by secretaries from the ICSID Secretariat, many of whom are responsible for multiple cases. Against this background, it can be argued that over-representation is not a challenge related to this study, but rather a potential challenge related to the organization of the ICSID dispute settlement system. This study may shed some light on the extent to which such over-representation of certain views or perspectives constitutes a problem for ICSID.

Finally, it can be argued that each interpretive issue must be approached on its own merits and cannot be generalized, for example because a tribunal may choose not to state all its interpretive arguments or only explicitly to address the arguments raised by the parties to the dispute. Hence, an approach based on registering the use of specific approaches to interpretive arguments does not give a full and correct picture of the interpretive processes. However, such an argument would constitute an argument against the usefulness of making any general statement concerning interpretive processes. The ‘empirical’ approach of this study does not give definitive answers to how individual tribunals have approached interpretive issues, but it may contribute to a general and system-oriented discussion of interpretive issues, and it may serve to falsify some general statements that have been made concerning investment tribunals’ use of interpretive arguments. Moreover, it may serve as a useful basis for comparative studies of international dispute settlement mechanisms.

Conclusions295

If we arrange the interpretive arguments found in the 98 decisions according to the number of decisions in which the arguments occur, we get the following list:

  • 1 ICSID case law – 90 decisions;

  • 2 Legal doctrine – 73 decisions;

  • 3 State practice – 52 decisions;

  • 4 Contextual arguments – 49 decisions;

  • 5 Object and purpose – 48 decisions;

  • 6 ICJ case law – 46 decisions;

  • 7 Reasonableness – 38 decisions;

  • 8 UNCITRAL case law – 30 decisions;

  • 9 Case law from other investment tribunals – 30 decisions;

  • 10 The Report of the Executive Directors – 26 decisions;

  • 11 Preparatory work – 25 decisions;

  • 12 Customary international law – 24 decisions;

  • 13 Case law from the Iran–US Claims Tribunal – 22 decisions;

  • 14 A contrario arguments – 21 decisions;

  • 15 Agreement between the parties to treaties – six decisions;296

  • 16 General principles of law – four decisions.

The most remarkable aspects of this list are the high number of decisions that used legal doctrine, various forms of case law, and state practice, and the relatively low number of cases that used the context, object and purpose, preparatory work, agreement between parties to treaties, and general principles of law. These aspects indicate that ICSID tribunals are ‘dispute-oriented’ rather than ‘legislator-oriented’.297 Moreover, they indicate that ICSID tribunals tend to follow an approach based in a common law, rather than a civil law, tradition when addressing interpretive issues.298

Figure 3 shows how ICSID tribunals used the interpretive arguments based on the distinction between essential arguments, non-essential arguments, and starting points for the legal reasoning.299 The figure shows the percentage of decisions300 that used the interpretive argument.

Figure 3.

Use of interpretive arguments by ICSID tribunals.

Figure 3.

Use of interpretive arguments by ICSID tribunals.

The figure does not include case law from ICSID due to the extensive use of such case law in the decisions.301 The main points of interest are the high proportion of essential arguments found in relation to reasonableness, contextual arguments, object and purpose, the ICJ, and doctrine (more than 60 per cent of the decisions), and the high proportion of non-essential arguments in case law from investment tribunals, state practice, and preparatory work. The extent to which tribunals used object and purpose, case law from the ICJ, and the legal doctrine as starting points for their legal analysis is also noteworthy. These findings to some (limited) extent confirm a description of ICSID tribunals as ‘dispute-oriented’ and relying on a common law approach. However, when taken together with the list above, Figure 3 indicates a complex situation in which the approaches of ICSID tribunals to interpretive issues vary significantly. Moreover, the more detailed assessments of ICSID tribunals’ use of the interpretive arguments above indicate that ICSID tribunals can be regarded as ‘legislator-oriented’ in many cases. In sum, it seems appropriate to indicate that whether ICSID tribunals should be regarded as ‘legislator-oriented’ or ‘dispute-oriented’ depends on the context, while the way in which they use case law as an interpretive argument is ‘legislator-oriented’, the relative importance that they attribute to different categories of interpretive arguments is ‘dispute-oriented’.

On the question concerning the extent to which ICSID tribunals contribute to the development of international investment law, it can be observed that ICSID tribunals make use of a broad range of interpretive arguments, and that the tribunals vary significantly as to how they use the arguments. Moreover, it has been indicated that ICSID tribunals in general have a tendency to be ‘dispute-oriented’. Hence, while it can be concluded that ICSID tribunals’ use of interpretive arguments indicates that they contribute significantly to the long-term development of international investment law, it can also be observed that the tribunals in general have significant potential for increasing their contribution.

On the question concerning the extent to which ICSID tribunals contribute to the fragmentation of international law, the tribunals’ willingness to use case law from other tribunals and to take into account state practice and customary international law indicate a general interest in aligning ICSID tribunals with other areas of international law. However, there is significant potential for ICSID tribunals to broaden their perspective by selecting arguments from materials that are related to other areas of international law. This is particularly relevant for their use of state practice and legal doctrine. In this context, one main challenge for the tribunals is to look beyond the arguments presented by the parties to the dispute, while maintaining the dispute within a manageable frame. Another main challenge is for the tribunals to align the way they use interpretive arguments with the ways in which other international tribunals use such arguments.

While it has been argued convincingly that ‘the dispute resolution system devised by [the international] society must make available both centralized and decentralized mechanisms for attending to the social needs of [the] evolving structure [of the international society]’,302 it is the opinion of this author that international tribunals’ approaches to interpretive issues should to a significant degree be ‘centralized’. The way in which ICSID tribunals use interpretive arguments in practice is often quite far removed from the structures set out in Articles 31–32 of the VCLT.303 On the other hand, it can be argued that other tribunals depart from these provisions as well, and that the differences between the approaches of ICSID tribunals and other tribunals are in general minor. Nevertheless, it seems appropriate to conclude that ICSID tribunals could do significantly more to align their approaches with those of other tribunals. The ICSID Secretariat is likely to play an essential role in this context, both through its selection of arbitrators and through its appointment of secretaries to the tribunals. The potential establishment of one centralized (or perhaps several decentralized) appeals procedure could contribute significantly to streamlining the decisions of ICSID tribunals and increasing the compatibility of their interpretive approaches with those of other international tribunals.

Finally, on the question of the extent to which ICSID tribunals create a predictable legal framework in which the interests of investors, states, and interested third parties are properly taken into account, the rather extensive use of reasonableness and legal doctrine as interpretive arguments is of particular interest. Such arguments are likely to contribute to the understanding of the general and long-term implications of decisions. On the other hand, the rather infrequent use of object and purpose, which in most cases was narrowly defined, the rather unsystematic use of state practice, as well as the tendency to be ‘dispute oriented’, indicates that ICSID tribunals have significant potential to increase their ability to take into account interests other than those represented to them by investors and host countries. The proliferation of instruments that ensure transparency and participation in investment proceedings is likely to contribute significantly in this context.304

Annex: ICSID Case Law Covered by the Study

This study covers 98 decisions from the period between 1 January 1998 and 31 December 2006 by tribunals established under ICSID. Not covered by the study are:

  • decisions during the period that were not public;305

  • decisions that concern only the application of a contract or domestic law (such decisions addressing jurisdictional or procedural issues under the ICSID Convention are included);306

  • decisions on rectification or supplementation;

  • decisions on stay of enforcement in annulment cases;

  • most decisions that contain only procedural orders (decisions that deal in some depth with interpretive questions are included).

The list below is organized alphabetically according to the short title of the case used in the article. Cases can be accessed electronically at www.worldbank.org/icsid/cases/cases.htm or http://ita.law.uvic.ca.

ADC v. Hungary (ADC Affiliate Limited and ADC & ADMC Management Limited v. Republic of Hungary), ICSID Case ARB/03/16, BIT Hungary–Cyprus Award 2006.

ADF v. US (ADF Group Inc. v. United States of America), ICSID Case ARB(AF)/00/1. NAFTA Award 2003, 18 ICSID Rev – FILJ (2003) 195.

AES v. Argentina (AES Corporation v. Argentine Republic), ICSID Case ARB/02/17. BIT Argentina–US Decision on jurisdiction 2005.

Aguas v. Bolivia (Aguas del Tunari SA v. Republic of Bolivia), ICSID Case ARB/02/3. BIT Bolivia–Netherlands Decision on jurisdiction 2005, 20 ICSID Rev – FILJ (2005) 450.

Autopista v. Venezuela (Autopista Concesionada de Venezuela, CA v. Bolivarian Republic of Venezuela), ICSID Case ARB/00/5. Contract and domestic law; Decision on jurisdiction 2001, 6 ICSID Rep (2004) 419.

Azinian v. Mexico (Robert Azinian and others v. United Mexican States), ICSID Case ARB(AF)/97/2. NAFTA Award 1999, 39 ILM (2000) 537.

Azurix v. Argentina (Azurix Corp. v. Argentine Republic), ICSID Case ARB/01/12. BIT Argentina–US Decision on jurisdiction 2003, 43 ILM (2004) 262, Award 2006.

Banro v. Congo (Banro American Resources, Inc and Société Aurifère du kivu et du Maniema Sàrl v. Democratic Republic of Congo), ICSID Case ARB/98/7. Contract, Decision on jurisdiction 2000, 17 ICSID Rev – FIJL (2003) 382.

Bayindir v. Pakistan (Bayindir Insaat Turizm Ticaret Ve Sanayi AS v. Islamic Republic of Pakistan), ICSID Case ARB/03/29. BIT Pakistan–Turkey Decision on jurisdiction 2005.

Biwater v. Tanzania (Biwater Gauff (Tanzania) Ltd v. United Republic of Tanzania), ICSID Case ARB/05/22. BIT Tanzania–UK Provisional measures 1, 2006 (Procedural Order No 1); Provisional measures 2, 2006, 46 ILM (2007) 15.

Camuzzi v. Argentina 1 (Camuzzi International SA v. Argentine Republic (1)), ICSID Case ARB/03/2. BIT Argentina–Belgium/Luxembourg Decision on jurisdiction 2005.

Camuzzi v. Argentina 2 (Camuzzi International SA v. Argentine Republic (2)), ICSID Case ARB/03/7. BIT Argentina–Belgium/Luxembourg Decision on jurisdiction 2005.

Casado v. Chile (Victor Pey Pey Casado and President Allende Foundation v. Republic of Chile), ICSID Case ARB/98/2. BIT Chile–Spain Provisional measures 2001, 16 ICSID Rev – FILJ (2001) 567.

CDC v. Seychelles (CDC Group plc v. Republic of Seychelles), ICSID Case ARB/02/14. Contract; Annulment proceedings 2005, 11 ICSID Rep (2007) 237.

Cement Shipping v. Egypt (Middle East Cement Shipping and Handling Co SA v. Arab Republic of Egypt), ICSID Case ARB/99/6. BIT Egypt–Greece Award 2002, 18 ICSID Rev – FILJ (2003) 602.

CGE v. Argentina (Compañía de Aguas del Aconquija SA and Vivendi Universal v. Argentine Republic), ICSID Case ARB/97/3. BIT Argentina–France Award 2000, 40 ILM (2001) 426; Preliminary issue 2001, 17 ICSID Rev – FILJ (2002) 168; Annulment proceedings 2002, 41 ILM (2002) 1135.

Champion v. Egypt (Champion Trading Company and Ameritrade International, Inc v. Arab Republic of Egypt), ICSID Case ARB/02/9. BIT Egypt–US Decision on jurisdiction 2003, 19 ICSID Rev – FILJ (2004) 275; Award 2006.

CMS v. Argentina (CMS Gas Transmission Company v. Argentine Republic), ICSID Case ARB/01/8. BIT Argentina–US Decision on jurisdiction 2003, 42 ILM (2003) 788; Award 2005, 44 ILM (2005) 1205.

Continental v. Argentina (Continental Casualty Company v. Argentine Republic), ICSID Case ARB/03/9. BIT Argentina–US Decision on jurisdiction 2006.

CSOB v. Slovakia (Ceskoslovenska obchodni banka, as v. Slovak Republic), ICISD Case ARB/97/4. BIT Slovak–Czech Republic Decision on jurisdiction 1, 1999, 14 ICSID Rev – FILJ 251 (1999); Decision on jurisdiction 2, 2000, 15 ICSID Rev – FILJ (2000) 530.

Duke v. Peru (Duke Energy International Peru Investments No. 1 Ltd v. Republic of Peru), ICSID Case ARB/03/28. Contract; Decision on jurisdiction 2006.

El Paso v. Argentina (El Paso Energy International Company v. Argentine Republic), ICSID Case ARB/03/15. BIT Argentina–US Decision on jurisdiction 2006.

Enron v. Argentina (Enron Corporation and Ponderosa Assets, LP v. Argentine Republic), ICSID Case ARB/01/3. BIT Argentina–US Decision on jurisdiction 1, 2004, 11 ICSID Rep (2007) 273; Decision on jurisdiction 2, 2004, 11 ICSID Rep (2007) 295.

Fedax v. Venezuela (Fedax NV v. Republic of Venezuela), ICSID Case ARB/96/3. BIT Venezuela–Netherlands Award, 1998, 37 ILM (1998) 1391.

Feldman v. Mexico (Marving Roy Feldman Karpa v. United Mexican States), ICSID Case ARB(AF)/99/1. NAFTA Decision on jurisdiction 2000, 18 ICSID Rev – FILJ (2003) 469; Award 2002, 42 ILM (2003) 625.

Fireman's Fund v. Mexico (Fireman's Fund Insurance Company v. United Mexican States), ICSID Case ARB(AF)/02/1. NAFTA Decision on jurisdiction 2003, 10 ICSID Rep (2006) 214.

Gas Natural v. Argentina (Gas Natural SDG, SA v. Argentine Republic), ICSID Case ARB/03/10. BIT Argentina–Spain Decision on jurisdiction 2005.

Generation v. Ukraine (Generation Ukraine, Inc v. Ukraine), ICSID Case ARB/00/9. BIT Ukraine–US Award 2003, 44 ILM (2005) 404.

Genin v. Estonia (Alex Genin, Eastern Credit Limited, Inc. and AS Baltoil v. The Republic of Estonia), ICSID Case ARB/99/2. BIT Estonia–US Award 2001, 17 ICSID Rev – FILJ (2002) 395.

Goetz v. Burundi (Antoine Goetz and others v. Republic of Burundi), ICSID Case ARB/95/3. BIT Burundi–Belgium/Luxembourg Award 1999, 15 ICSID Rev – FILJ (2000) 457.

Gruslin v. Malaysia (Philippe Gruslin v. Malaysia), ICSID Case ARB/99/3. BIT Malaysia–Belgium/Luxembourg Decision on jurisdiction 2000, 5 ICSID Rep (2002) 484.

IBM v. Ecuador (IBM World Trade Corporation v. Republic of Ecuador), ICSID Case ARB/02/10. BIT Ecuador–US Decision on jurisdiction 2003.

Impregilo v. Pakistan (Impregilo SpA v. Islamic Republic of Pakistan), ICSID Case ARB/03/3. BIT Pakistan–Italy Decision on jurisdiction 2005.

Inceysa v. El Salvador (Inceysa Vallisoletana SL v. Republic of El Salvador), ICSID Case ARB/03/26. BIT El Salvador–Spain Decision on jurisdiction 2006.

Interagua v. Argentina (Suez, Sociedad General de Aguas de Barcelona SA and Interagua Servicios Integrales de Agua SA v. Argentine Republic), ICSID Case ARB/03/17. BIT Argentina–Spain Preliminary issues 2006; Decision on jurisdiction 2006.

Joy Mining v. Egypt (Joy Mining Machinery Limited v. Arab Republic of Egypt), ICSID Case ARB/03/11. BIT Egypt–UK Decision on jurisdiction 2004, 44 ILM (2005) 73.

Lanco v. Argentina (Lanco International, Inc v. Argentine Republic), ICSID Case ARB/97/6. BIT Argentina–US Decision on jurisdiction 1998, 40 ILM (2001) 457.

LESI v. Algeria 1 (Consortium Groupement LESI - DIPENTA v. People's Democratic Republic of Algeria), ICSID Case ARB/03/8. BIT Algeria–Italy Decision on jurisdiction 2005, 19 ICSID Rev – FILJ (2004) 426.

LESI v. Algeria 2 (LESI SpA and Astaldi SpA v. People's Democratic Republic of Algeria), ICSID Case ARB/05/3. BIT Algeria–Italy Decision on jurisdiction 2006.

LGE v. Argentina (LG&E Energy Corp, LG&E Capital Corp and LG&E International Inc v. Argentine Republic), ICSID Case ARB/02/1. BIT Argentina–US Decision on jurisdiction 2004, 11 ICSID Rep (2007) 414; Award 2006, 46 ILM (2007) 40.

Loewen v. US (The Loewen Group, Inc and Raymond L. Loewen v. United States of America), ICSID Case ARB(AF)/98/3. NAFTA Decision on jurisdiction 2001, 7 ICSID Rep (2005) 421; Award 2003, 42 ILM (2003) 811.

Lucchetti v. Peru (Lucchetti SA and Lucchetti Peru SA v. Republic of Peru), ICSID Case ARB/03/4. BIT Peru–Chile Decision on jurisdiction 2005, 19 ICSID Rev – FILJ (2004) 359.

Maffezini v. Spain (Emilio Agustin Maffezini v. Kingdom of Spain), ICSID Case ARB/97/7. BIT Spain–Argentina Provisional measures 1999, 16 ICSID Rev – FILJ (2001) 207; Decision on jurisdiction 2000, 16 ICSID Rev – FILJ (2001) 212; Award 2000, 16 ICSID Rev – FILJ (2001) 248.

Metalclad v. Mexico (Metalclad Corp v. United Mexican States), ICSID Case ARB(AF)/97/1. NAFTA Award 2000, 40 ILM (2001) 36.

Metalpar v. Argentina (Metalpar SA and Buen Aire SA v. Argentine Republic), ICSID Case ARB/03/5. BIT Argentina–Chile Decision on jurisdiction 2006.

Mihaly v. Sri Lanka (Mihaly International Corp v. Democratic Socialist Republic of Sri Lanka), ICSID Case ARB/00/2. BIT Sri Lanka–US Decision on jurisdiction 2002, 41 ILM (2002) 867.

Mitchell v. Congo (Patrick Mitchell v. Democratic Republic of the Congo), ICSID Case ARB/99/7. BIT Congo–US Annulment proceedings 2006.

Mondev v. US (Mondev International Ltd v. United States of America), ICSID Case ARB(AF)/99/2. NAFTA Award 2002, 42 ILM (2003) 85.

MTD v. Chile (MTD Equity Sdn Bhd and MTD Chile SA v. Republic of Chile), ICSID Case ARB/01/7. BIT Chile–Malaysia Award 2004, 44 ILM (2005) 91.

Noble v. Romania (Noble Ventures, Inc v. Romania), ICSID Case ARB/01/11. BIT Romania–US Award 2005.

Nul v. Egypt (Jan de Nul NV and Dredging International NV v. Arab Republic of Egypt), ICSID Case ARB/04/13. BIT Egypt–Belgium Decision on jurisdiction 2006.

Olguín v. Paraguay (Eudoro A. Olguín v. Republic of Paraguay), ICSID Case ARB/98/5. BIT Paraguay–Peru Decision on jurisdiction 2000, 6 ICSID Rep (2004) 156; Award 2001, 6 ICSID Rep (2004) 164.

Pan American v. Argentina (Pan American Energy LLC and BP Argentina Exploration Company v. Argentine Republic, and BP America Production Company and others v. Argentine Republic), ICSID Cases ARB/03/13 and ARB/04/8. BIT Argentina–US Decision on jurisdiction 2006.

Plama v. Bulgaria (Plama Consortium Ltd v. Republic of Bulgaria), ICSID Case ARB/03/24. Energy Charter Treaty and BIT Bulgaria–Cyprus Decision on jurisdiction 2005, 44 ILM (2005) 721; Provisional measures 2005.

PSEG v. Turkey (PSEG Global Inc and Konya Ilgin Elektrik Üretim ve Ticaret Limited Sirketi v. Republic of Turkey), ICSID Case ARB/02/5. BIT Turkey–US Decision on jurisdiction 2004, 44 ILM (2005) 465.

RFCC v. Morocco (Consortium RFCC v. Kingdom of Morocco), ICSID Case ARB/00/6. BIT Morocco–Italy Decision on jurisdiction 2001; Award 2003. 20 ICSID Rev – FILJ (2005) 391.

Salini v. Morocco (Salini Costruttori SpA and Italstrade SpA v. Kingdom of Morocco), ICSID Case ARB/00/4. BIT Morocco–Italy Decision on Jurisdiction 2001, 42 ILM (2003) 609.

Salini v. Jordan (Salini Costruttori SpA and Italstrade SpA v. The Hashemite Kingdom of Jordan), ICSID Case ARB/02/13. BIT Jordan–Italy Decision on jurisdiction 2004, 44 ILM (2005) 573; Award 2006.

Sempra v. Argentina (Sempra Energy International v. Argentine Republic), ICSID Case ARB/02/16. BIT Argentina–US Decision on jurisdiction 2005.

SGS v. Pakistan (SGS Société Générale de Surveillance SA v. Islamic Republic of Pakistan), ICSID Case ARB/01/13. BIT Pakistan–Switzerland Decision on jurisdiction 2003, 42 ILM (2003) 1290.

SGS v. Philippines (SGS Société Générale de Surveillance SA v. Republic of the Philippines), ICSID Case ARB/02/6. BIT Philippines–Switzerland Decision on jurisdiction 2004, 8 ICSID Rep (2005) 518.

Siemens v. Argentina (Siemens AG v. Argentine Republic), ICSID Case ARB/02/8. BIT Argentina–Germany Decision on jurisdiction 2004, 44 ILM (2005) 138.

Soufraki v. UAE (Hussein Nuaman Soufraki v. United Arab Emirates), ICSID Case ARB/02/7. BIT United Arab Emirates–Italy Decision on jurisdiction 2004.

Tanzania Electric v. Tanzania (Tanzania Electric Supply Company Ltd v. Independent Power Tanzania Ltd), ICSID Case ARB/98/8. Contract; Provisional measures 1999, 8 ICSID Rep (2005) 226.

Técnicas v. Mexico (Técnicas Medioambientales Tecmed SA v. United Mexican States), ICSID Case ARB(AF)/00/2. BIT Mexico–Spain Award 2003, 43 ILM (2004) 133.

Telenor v. Hungary (Telenor Mobile Communications AS v. Republic of Hungary), ICSID Case ARB/04/15. BIT Hungary–Norway Decision on jurisdiction 2006.

Tokelės v. Ukraine (Tokios Tokelės v. Ukraine), ICSID Case ARB/02/18. BIT Ukraine–Lithuania Decision on jurisdiction 2004, 20 ICSID Rev – FILJ (2005) 205; Provisional measures 2005, 11 ICSID Rep (2007) 352.

Tradex v. Albania (Tradex Hellas SA v. Republic of Albania), ICSID Case ARB/94/2. National legislation; Award 1999, 14 ICSID Rev – FILJ (1999) 197.

Vivendi v. Argentina (Suez, Sociedad General de Aguas de Barcelona SA and Vivendi Universal SA v. Argentine Republic), ICSID Case ARB/03/19. BIT Argentina–Spain Preliminary issues 2005; Decision on jurisdiction 2006.

Waste Management v. Mexico 1 (Waste Management, Inc v. United Mexican States (1)), ICSID Case ARB(AF)/98/2. NAFTA Decision on jurisdiction 2000, 40 ILM (2001) 56.

Waste Management v. Mexico 2 (Waste Management, Inc v. United Mexican States (2)), ICSID Case ARB(AF)/00/3. NAFTA Preliminary issues 2001, 6 ICSID Rep (2004) 541; Decision on jurisdiction 2002, 41 ILM (2002) 1315; Award 2004, 43 ILM (2004) 967.

WDFC v. Kenya (World Duty Free Company Ltd v. Republic of Kenya), ICSID Case ARB/00/7. Contract. Award 2006, 46 ILM (2007) 339.

Wena v. Egypt (Wena Hotels Ltd v. Arab Republic of Egypt), ICSID Case ARB/98/4. BIT Egypt–UK Decision on jurisdiction 1999, 41 ILM (2002) 881; Award 2000, 41 ILM (2002) 896; Annulment proceedings 2002, 41 ILM (2002) 933; Interpretation proceedings 2005.

Hanne Rignes has contributed extensively to the initial phases of my work on this article. The work was undertaken while I stayed as visiting researcher at Peking University, invited by Professors Edward J. Shao and Bai Guimei. This article is part of a research project at the Faculty of Law, University of Oslo, on international tribunals: legalization and constitutionalization – implications for domestic constitutional structures. I thank researcher Jo Stigen and Professors Geir Ulfstein and Giuditta Cordero Moss, as well as an anonymous referee for comments on drafts of the article

1
See S. Rosenne, The Law and Practice of the International Court 1920–2005, Volume III Procedure (4th edn, 2006), at 1550–1551. Commission, ‘Precedent in Investment Treaty Arbitration. A Citation Analysis of a Developing Jurisprudence’, 24(2) J Int’l Arb (2007) 129 examines ‘the sources of law cited’ in decisions of investment tribunals, but does not make a clear distinction between sources and interpretive materials: see in particular at 151–152.
2
The article will not address arguments used in dissenting opinions.
3
Hence, the approach of this article differs from that of a ‘citation analysis’: see Drahozal, ‘The Iran–U.S. Claims Tribunal and Investment Arbitration: A Citation Analysis’, 3(2) Transnat’l Dispute Management (2006) 1, at 3–5.
4
See the Report of the International Law Commission from its 58th session, in Official Records of the General Assembly, Sixty-first Session, Supplement No. 10 (A/61/10), at 405, para. 247: ‘[t]he rationale for the Commission's treatment of fragmentation is that the emergence of new and special types of law, so-called “self-contained regimes” and geographically or functionally limited treaty-systems, creates problems of coherence in international law’. See also paras 11–16 of the Conclusions of the ILC's Study Group on Fragmentation of International Law, at 410–412.
5
J. G. Merrills, International Dispute Settlement (4th edn, 2005) may serve to illustrate the approaches generally taken.
6
Examples are E. McWhinney, Judicial Settlement of International Disputes. Jurisdiction, Justiciability and Judicial Law-making on the Contemporary International Court (1991) and O. Spiermann, International Argument in the Permanent Court of International Justice. The Rise of the International Judiciary (2005).
7
These studies emphasize materials that may indicate reasons for the judgments other than those that appear in the decision itself: see Spiermann, supra note 6, at 129–132.
8
Other studies that have used somewhat similar approaches include M.O. Hudson, The Permanent Court of International Justice 1920–1942. A Treatise (1943), at 640–661, and H. Lauterpacht, The Development of International Law by the International Court (1958). Relevant studies of investment arbitration include in particular C.H. Schreuer, ‘Diversity and Harmonization of Treaty Interpretation in Investment Arbitration’, 3(2) Transnat’l Dispute Settlement (2006) 1. See also Commission, supra note 1, at 142–143 with further references. Commission's study is essentially limited to an assessment of the role of case law.
9
The Convention on the Settlement of Investment Disputes between States and Nationals of Other States (1965): see International Center for the Settlement of Investment Disputes, ICSID Convention, Regulations and Rules, ICSID/15/Rev.1 (2003), at 7, available at: www.worldbank.org/icsid/.
10
See the Agreement among the Government of Brunei Darussalam, the Republic of Indonesia, Malaysia, the Republic of the Philippines, the Republic of Singapore, and the Kingdom of Thailand for the Promotion and Protection of Investments (The ASEAN Agreement for the Promotion and Protection of Investments, 1987), the Energy Charter Treaty (ECT, 1990), the North American Free Trade Agreement (NAFTA, 1994) and the Protocol of Colonia for the Promotion and Reciprocal Protection of Investments within MERCOSUR (1994).
11
For an examination of the use of investment clauses in economic integration agreements see UNCTAD, Investment Provisions in Economic Integration Agreements (2006).
12
The main examples are contracts between investors and states or entities representing states, and permits or concessions issued by public authorities. See, e.g., Duke v. Peru, at para. 121 (full references to the cases covered by the study are set out in the Annex to this article). The complete references of ICSID cases are set out in the Annex to the article.
13
See Inceysa v. El Salvador, at 100 and para. 332.
14
The commentary to Clause 10 of the ICSID Model Clauses states that ‘[t]he parties are free to agree on rules of law defined as they choose. They may refer to a national law, international law, a combination of national and international law, or a law frozen in time or subject to certain modifications.’
15
One question is whether the tribunal is to take national law as the starting point and supplement such law by international law, or whether the tribunal is to regard international law as the basis for its legal reasoning. In Tradex v. Albania, at para. 70 the tribunal concluded that it ‘will make use of sources of international law insofar as that seems appropriate for the interpretation of terms used in the 1993 Law, such as “expropriation”’. In Compañía del Desarrollo de Santa Elena SA v. Republic of Costa Rica, ICSID Case ARB/96/1, at para. 65 the tribunal held that ‘[t]he parties’ apparently divergent positions lead, in substance, to the … conclusion … that, in the end, international law is controlling. The Tribunal is satisfied that, under the second sentence of Article 42(1), the arbitration is governed by international law.’ The latter approach has prevailed in subsequent cases.
16
Difficulties in drawing a clear distinction between rules and facts have not had significant consequences for the analysis of interpretive arguments.
17
In light of the increasing availability of relevant case law it is not surprising that one would find ‘increasing citation of ICSID precedents by ICSID tribunals’:, see Commission, supra note 1, at 149–150. However, such an increase can be assumed to level out after a period of time unless there is a subsequent development in the use of case law as an interpretive argument.
18
Of these, 62 cases including 84 decisions were made under the Arbitration Rules and 10 cases including 14 decisions were made under the Additional Facility Rules.
19
An overview of decisions and cases included in and those left out of the study can be found in the Annex. Conciliation and fact-finding cases fall outside the study since in general they do not lead to decisions based on the interpretation of substantive or procedural rules: see Rules 30 and 32 of the Conciliation Rules, see International Center for the Settlement of Investment Disputes, supra note 9, and Art. 15(4) of the Fact-Finding (Additional Facility) Rules, see International Center for the Settlement of Investment Disputes, ICSID Additional Facility Rules, ICSID/11/Rev.1 (2003), at 13, available at: www.worldbank.org/icsid/. As of November 2005, only five of 198 registered cases were conciliation cases and there were no fact-finding cases: see Onwuamaegbu, ‘The Role of ADR in Investor–State Dispute Settlement: the ICSID Experience’, 22(2) News from ICSID (2005) 12, at 12.
20
See Art. 41(2) of the ICSID Convention, supra note 9. The distinction between jurisdictional issues and the merits of the case is unclear, in particular where the respondent argues that the tribunal lacks jurisdiction because the claim is not sufficiently substantiated: see Bayindir v. Pakistan, at paras 185 ff. Hence, jurisdictional decisions do sometimes address the interpretation of substantive provisions.
21
See Art. 47 of the ICSID Convention, supra note 9, and Arbitration Rule 39.
22
Preliminary issues may, e.g., concern challenges to members of tribunals, third party participation in proceedings, and whether the case is manifestly without legal merit: see Arbitration Rule 41(5) and (6).
23
See Art. 50 of the ICSID Convention, supra note 9, and Arbitration Rule 51.
24
See Art. 52 of the ICSID Convention, supra note 9, and Arbitration Rules 50(1)(c)(ii) and 54.
25
See also Arts 52(1)(i) and 54 of Schedule C to the Arbitration (Additional Facility) Rules. supra note 19.
26
See Art. 42(3) of the ICSID Convention, supra note 9. None of the issues taken up in the cases examined were resolved ex aequo et bono (see, however, Corn Products v. Mexico (preliminary issue), which in reality was decided on an ex aequo et bono basis, and Técnicas v. Mexico, at para. 190, where the issue was raised). Ex aequo et bono decisions fall outside the scope of this article since they do not address the interpretation of rules.
27
‘The Parties shall interpret and apply the provisions of this Agreement in the light of its objectives set out in paragraph 1 and in accordance with applicable rules of international law.’
28
Annex D sets out how Arts 5 and 29 ECT shall be interpreted in the light of relevant rules under the WTO Agreement. There are some examples of similar rules under bilateral economic integration agreements, but these rules generally apply to the interpretation of provisions related to trade in goods and services: see, e.g., the Record of Understanding relating to the Agreement between the EFTA States and Singapore (2002), available at: http://secretariat.efta.int/Web/legaldocuments/.
29
See Art. 41 of the Agreement between Canada and the Republic of Peru for the Promotion and Protection of Investments (2006), available at: www.international.gc.ca/assets/trade-agreements-accords-commerciaux/pdfs/Canada-Peru10nov06-en.pdf, which sets out a separate procedure for the interpretation of Annexes to the treaty and instructs the tribunal to respect the interpretation established through this procedure.
30
See, e.g., Mitchell v. Congo (annulment), at paras 40–41.
31
One recent example of a decision that was annulled can be found in CGE v. Argentina (annulment proceedings).
32
See UNCTAD, Investor-State Disputes Arising from Investment Treaties. A Review (2005), at 57: ‘this is contemplated in several recent treaties involving the United States, and is the subject of increasing attention from Governments and academics. According to ICSID, by mid-2005 as many as 20 countries may have signed treaties with provisions on an appeals mechanism in investor-State disputes.’ For an example see Annex 10-F to the Dominican Republic–Central America–United States Free Trade Agreement (2004), available at: www.ustr.gov/Trade_Agreements/Bilateral/CAFTA/CAFTA-DR_Final_Texts/Section_Index.html. See also Gantz, ‘An Appellate Mechanism for Review of Arbitral Decisions in Investor – State Disputes: Prospects and Challenges’, 39 Vanderbilt J Transnat’l L (2006) 39 and Franck, ‘The Legitimacy Crisis in Investment Treaty Arbitration: Privatizing Public International Law through Inconsistent Decisions’, 73 Fordham L Rev (2005) 1587.
33
Supra note 7.
34
See the difference in approaches between ADC v. Hungary, where the tribunal focused narrowly on the disagreements between the parties to the dispute, and WDFC v. Kenya (in particular paras 180–181) and Biwater v. Tanzania (provisional measures 2) (in particular paras 121 ff.), where the tribunals took into account a broad range of interests.
35
This list can be compared to those of other authors. Hudson, supra note 8, at 640–661 distinguished between: intention of the Parties, ‘natural’ meaning, context, nature and purpose of an instrument, travaux préparatoires, legal background, political and social background, analogous provisions, action by the Parties, liberal or restrictive interpretation, and special rules of interpretation. Lauterpacht, supra note 8, established a shorter list: case law (including from other tribunals), teachings of publicists, preparatory works, general principles of law, customary international law, and object and purpose. Rosenne, supra note 1, at 1552–1562 distinguished between: judicial decisions, teachings of publicists, practice of states, and resolutions of the UN General Assembly. As we see, the categories used by these authors differ significantly. These examples indicate that there are currently no authoritative categories of interpretive arguments beyond those that can be derived from the Statute of the ICJ and the VCLT.
36
Where lists are omitted, they can be obtained from the author at: o.k.fauchald@jus.uio.no.
37
A treaty may contain an implicit reference to customary international law or general principles of law. In such cases, it may be unclear whether the use of customary law or general principles of law is to be regarded as a direct application of such rules or whether they are applied as interpretive arguments. An example is the provisions on ‘fair and equitable treatment’. Cases where there is no explicit reference to a rule outside the treaty will be classified as cases concerning treaty interpretation for the purpose of this study.
38
In cases where a treaty makes use of general concepts it can possibly be argued that there is a legal lacuna, such as where the ICSID Convention uses the term ‘investment’ without further definition: see, e.g., Mihaly v. Sri Lanka, paras. 33 and 58. Such cases are classified as cases concerning treaty interpretation for the purpose of this study.
39
Customary international law was not found to replace existing treaty obligations in any of the cases examined.
40
As explained in para. 40 of the Report of the Executive Directors of the International Bank for Reconstruction and Development on the Convention on the Settlement of Investment Disputes between States and Nationals of Other States, see International Center for the Settlement of Investment Disputes, supra note 9, (2003), at 35 (hereinafter ‘Report of the Executive Directors’), the use of the phrase ‘international law’ is to be understood as this phrase has been defined in Art. 38 of the Statute of the International Court of Justice, i.e., as referring to both customary international law and general principles of law.
41
For an example see Art. 1131(1) of NAFTA.
42
In Inceysa v. El Salvador, at para. 227, the tribunal stated: ‘[w]ithout attempting to define what the general principles of law are, the Tribunal notes that, in general, they have been understood as general rules on which there is international consensus to consider them as universal standards and rules of conduct that must always be applied and which, in the opinion of important commentators, are rules of law on which the legal systems of the States are based’. The unclear distinction between customary law and general principles can be illustrated by the principles enumerated in Asian Agricultural Products Ltd. v. Sri Lanka, ICSID Case ARB/87/3, at paras 40 and 56. Many principles of customary international law originate in domestic law, and it is often unclear when such a principle qualifies as a rule of customary international law. See also WDFC v. Kenya, which uses the concepts ‘international public policy’ and ‘transnational public policy’ at para. 157.
43
See Wena v. Egypt (award), at para. 111, WDFC v. Kenya, at paras 138–157, CSOB v. Slovakia (jurisdiction 1), at para. 47, and ADC v. Hungary, at para. 475.
44
The following jurisdictional issues were discussed: The date for assessing whether the conditions for jurisdiction was fulfilled (CSOB v. Slovakia (jurisdiction 1), at para. 31), exhaustion of domestic remedies (Maffezini v. Spain (jurisdiction), at para. 29), the relationship between jurisdiction according to contract, legislation, and treaty (SGS v. Philippines, at paras 138, 141–142, 145, and 154, and Continental v. Argentina, at paras 88–90), piercing the corporate veil as a means to reject jurisdiction (Tokelės v. Ukraine (jurisdiction), at paras 53–56), the requirement that an agreement to confer jurisdiction must be clear and unambiguous (Plama v. Bulgaria (jurisdiction), at para. 198) and whether claims can be made on grounds which were not invoked before domestic courts where the dispute concerns the lawfulness of the domestic court decision (Loewen v. US (award), at para. 87).
45
The following procedural issues were discussed: Questions concerning estoppel (CSOB v. Slovakia (jurisdiction 1), at para. 47, Gruslin v. Malaysia, at para. 20, and Camuzzi v. Argentina 2, at para. 64), and ‘abuse of process’ as a justification for dismissing a claim (Waste Management v. Mexico 2 (jurisdiction), at paras 48–49).
46
Substantive issues were mostly discussed on the basis of explicit references to customary international law in the treaties, but not always. The following substantive issues were discussed: expropriation (Feldman v. Mexico (award), at paras 115–116 and Técnicas v. Mexico, at paras 116 and 139), and fair and equitable treatment (Técnicas v. Mexico, at paras 153–155, ADF v. US, at paras 179 and 183, ADC v. Hungary, at para. 423, Mondev v. US, at paras 111–125, Loewen v. US (award), at paras 129–130, and Genin v. Estonia, at para. 367).
47
In its Commentaries to the Draft Articles on Responsibility of States for Internationally Wrongful Acts, UN Doc A/56/10, at 59, para. 1, the International Law Commission defined ‘secondary rules of State responsibility’ as ‘the general conditions under international law for the State to be considered responsible for wrongful actions or commissions, and the legal consequences which flow therefrom’. The following ‘secondary rules’ issues were discussed: retroactive application of treaty or unilateral act (Mondev v. US, at paras 68 and 70, Salini v. Jordan (jurisdiction), at para. 177, Impregilo v. Pakistan, at para. 310, Técnicas v. Mexico, at paras 63, 70–71, and 172, and Bayindir v. Pakistan, at para. 178), when does an act have legal effects (CSOB v. Slovakia (jurisdiction 1), at para. 46 and PSEG v. Turkey, at para. 142), consequences to be drawn from silence (Salini v. Jordan (award), at para. 95), internal law as excuse for non-compliance with international obligations (CGE v. Argentina (award), at para. 49, Metalclad v. Mexico, at para. 100, and IBM v. Ecuador, at paras 71–72), attribution of acts of different actors to the state (Azinian v. Mexico, at paras 98–99, CGE v. Argentina (award), at para. 49, Metalclad v. Mexico, at para. 73, Generation v. Ukraine, at para. 10.3, Enron v. Argentina (jurisdiction 1), at para. 32, Noble v. Romania, at paras 69–70, 81–82, and 84, Nul v. Egypt, at para. 84, Maffezini v. Spain (jurisdiction), at paras 76–78, and Salini v. Morocco, at para. 31), prescription of claim (Wena v. Egypt (award), at para. 106), public order (corruption) (Wena v. Egypt (award), at para. 111), the relationship between breach of contract and breach of treaty (CGE v. Argentina (annulment), at paras 95–96 and 110, n. 78), effect of treaty for non-parties (Mihaly v. Sri Lanka, at para. 23), circumstances precluding wrongfulness (state of necessity) (CMS v. Argentina (award), at paras 315–331, 379–382, and 383–394), obligation to avoid acts that would aggravate the conflict in question (Casado v. Chile, at para. 67 and Biwater v. Tanzania (provisional measure 2), at para. 135) and remedies (CMS v. Argentina (award), at paras 399–404 and ADC v. Hungary, at paras 481–495).
48
One example is Mondev v. US, at paras 110–113.
49
See in particular CMS v. Argentina (award), at paras 315–331, Mondev v. US, at paras 110–127, and ADC v. Hungary, at paras 481–495. One tribunal cautioned against the use of what it labelled ‘underdeveloped and peripheral principles’ of customary law: see Feldman v. Mexico (award), at para. 62.
50
At para. 116. See also ADF v. US, at para. 179 and Mondev v. US, at para. 125.
51
For an example see Autopista v. Venezuela, at para. 316.
52
See the discussion of res judicata in Waste Management v. Mexico 2 (jurisdiction), at paras 38–47.
53
The following procedural issues were discussed: burden of proof (Cement Shipping v. Egypt, at paras 89–90 and Salini v. Jordan (award), at para. 70), and three procedural principles selected from a list presented by the tribunal in Asian Agricultural Products Ltd. v. Sri Lanka, ICSID Case ARB/87/3 (Tradex v. Albania, at para. 84).
54
The following ‘secondary rules’ issues were discussed: The effect of subsequent conduct (ADC v. Hungary, at para. 475), the duty to mitigate damages (Cement Shipping v. Egypt, at para. 167), no one can transfer a better title than he or she has (Mihaly v. Sri Lanka, at para. 24), ‘public policy’ as a general exception (WDFC v. Kenya, at paras 138–157 and Inceysa v. El Salvador, at paras 245–250), good faith (Inceysa v. El Salvador, at paras 230–233), a right based on fraud does not enjoy protection (i.e., the principle nemo auditur propriam turpitudinem allegans) (Inceysa v. El Salvador, at paras 240–241), and unlawful enrichment (Inceysa v. El Salvador, at paras 253–254).
55
See Waste Management v. Mexico 2 (jurisdiction), at paras 38–47, Inceysa v. El Salvador, at paras 222–254, and WDFC v. Kenya, at paras 138–157.
56
There is no generally accepted definition of ‘self-contained regimes’ in international law: see Report of the International Law Commission, supra note 4, at 410–412. For the purpose of this article, the concept is used to illustrate a continuum between a regime that is fully integrated under the general rules, institutions, and procedures of international law, and a regime that functions in full isolation from such rules, institutions, and procedures (i.e., an extreme version of a self-contained regime).
57
Para. 13 of the Report of the Executive Directors, supra note 40, emphasizes that the tribunals should be equally available to states wanting to bring cases against investors, as they would be available to investors bringing cases against states. The actual development shows that this expectation has not been met. Only exceptionally have states brought cases against investors to ICSID arbitration (see Gabon v. Société Serete SA, ICSID Case ARB/76/1 and Tanzania Electric v. Tanzania), and there are few rules that may serve as a legal basis for legal claims in such cases. In Genin v. Estonia the state raised a counterclaim against the investor: see paras 201, 235, 309, 314, and 376–378. See also SGS v. Philippines, at para. 40.
58
See also paras. 40 and 45 of the Report of the Executive Directors, supra note 40.
59
This finding is in contrast to Schreuers’ statement that ‘[t]ribunals almost invariably start by invoking Article 31 of the Vienna Convention on the Law of Treaties (VCLT) when interpreting treaties’: see Schreuer, supra note 7, at 1.
60
For the years 1999–2002 there were references in 21% of the decisions, and for the years 2003–2006 there were references in 47% of the decisions.
61
These include the following: CSOB v. Slovakia (jurisdiction 1), at para. 57, Champion v. Argentina (jurisdiction), at 16–17, Fireman's Fund v. Mexico, at paras 94–95, Técnicas v. Mexico, at paras 121 and 155, ADF v. US, at para. 149, Enron v. Argentina (jurisdiction 1), at para. 47, Enron v. Argentina (jurisdiction 2), at para. 32, Tokelės v. Ukraine (jurisdiction), at paras 27–38, 75, and 85, MTD v. Chile, at paras 112–113, Siemens v. Argentina, at paras 80–81, Plama v. Bulgaria (jurisdiction), at paras 117, 147, 158, 160, 188, 194, and 196, Sempra v. Argentina, at paras 141–143, Camuzzi v. Argentina 1, at paras 133–134, CMS v. Argentina (award), at para. 360, CDC v. Seychelles, at paras 33–34, Noble v. Romania, at paras 50 and 55, Aguas v. Bolivia, at paras 88, 90–93, 226, 230, 250, and 291, Azurix v. Argentina (award), at paras 307, 359, 361, and 391, LGE v. Argentina (award), at paras 122 ff., Metalpar v. Argentina, at para. 91, and Mitchell v. Congo, at para. 43.
62
Mihaly v. Sri Lanka, at para. 58, Inceysa v. El Salvador, at paras 225–229, and Waste Management v. Mexico 2 (jurisdiction), at para. 39.
63
The most significant are Plama v. Bulgaria (jurisdiction), at paras 117, 147, 158, 160, 188, 194, and 196, Tokelės v. Ukraine (jurisdiction), at paras 27–38, and Aguas v. Bolivia, at paras 88, 90–93, 226, 230, 250, and 291.
64
See Mitchell v. Congo, at para. 43, Mondev v. US, at para. 43, and Noble v. Romania, at para. 55. See also Gruslin v. Malaysia, at para. 21.6 where the tribunal stated that: ‘[i]f its meaning is found to be clear, the Tribunal will not reduce its reach by reference to general consideration of assumptions derived from extrinsic sources of the sort relied upon by the Respondent in its materials and arguments’. Subsequently, it found it unnecessary to address the ‘extrinsic sources’ invoked. On the use of case law from the ICJ as interpretive arguments more generally see sect. 12.4 below.
65
See, e.g., Azinian v. Mexico, at paras 100–120, Wena v. Egypt (award), at paras 85–94, Goetz v. Burundi, at paras 63–70, Maffezini v. Spain (jurisdiction), at paras 62–63, Metalclad v. Mexico, at paras 109–111, Mihaly v. Sri Lanka, at para. 24, ADF v. US, at paras 196–197, Tokelės v. Ukraine, at paras 57–70, Aguas v. Bolivia, at para. 118, Inceysa v. El Salvador, at paras 208–264, Loewen v. US (award), at paras 54–70, Joy Mining v. Egypt, at para. 63, Plama v. Bulgaria (jurisdiction), at para. 225, Sempra v. Argentina, at para. 46, Noble v. Romania, at para. 184, and El Paso v. Argentina, at paras 70–85 and 138.
66
See, e.g., Mondev v. US, at para. 153, Técnicas v. Mexico, at para. 56, Bayindir v. Pakistan, at para. 103, Champion v. Egypt (award), at para. 134, and Mitchell v. Congo, at para. 48.
67
Examples include Plama v. Bulgaria (jurisdiction), at para. 193, Sempra v. Argentina, at para. 141, Camuzzi v. Argentina 1, at para. 133, CDC v. Seychelles, at para. 33, and Inceysa v. El Salvador, at paras 227–228.
68
See the International Law Commission's commentaries on Art. 31 (Art. 27 in ILC's draft) of the VCLT in the Yearbook of the International Law Commission (1966), ii, at 218 and 220–221, paras 2 and 11–12.
69
See Aguas v. Bolivia, at para. 91.
70
Examples that illustrate different ways in which an objective approach to treaty interpretation was expressed include Gruslin v. Malaysia, at para. 21.6, Mondev v. US, at para. 79, Aguas v. Bolivia, at paras 225–238, and Continental v. Argentina, at paras 58 ff.
71
See Lanco v. Argentina, at para. 29, Maffezini v. Spain (provisional measure), at para. 9, Maffezini v. Spain (jurisdiction), at paras 34 and 74, CGE v. Argentina (annulment), at para. 54, Waste Management v. Mexico 1, at para. 27, Técnicas v. Mexico, at para. 70, ADF v. US, at para. 161, Siemens v. Argentina, at para. 85, Plama v. Bulgaria (jurisdiction), at para. 147, Impregilo v. Pakistan, at para. 299, Aguas v. Bolivia, at para. 89, LGE v. Argentina (award), at para. 88, and Mitchell v. Congo, at para. 20.
72
Tribunals made use of dictionaries in 12 of the 98 decisions: Feldman v. Mexico (award), at para. 96, ADF v. US, at para. 161, SGS v. Philippines, at para. 131, Tokelės v. Ukraine (jurisdiction), at paras 26 and 75, MTD v. Chile, at para. 113, CMS v. Argentina (award), at para. 291, Aguas v. Bolivia, at paras 227, 229, and 231–234, Tokelės v. Ukraine (provisional measures), at para. 16, Nul v. Egypt, at paras 99–100, Azurix v. Argentina (award), at paras 360 and 392, LGE v. Argentina (award), at para. 157, and Mitchell v. Congo, at para. 20. An illustrative example of a tribunal that made extensive use of such approaches is Aguas v. Bolivia, at paras 227–233.
73
Examples can be found in CSOB v. Slovakia (jurisdiction 1), at para. 57, Wena v. Egypt (jurisdiction), at sections IV.C and IV.D, Metalclad v. Mexico, at para. 70, and Mondev v. US, at para. 119. The tribunal noted the following warning in Plama v. Bulgaria (jurisdiction), at para. 193: ‘[h]ere, the Tribunal is mindful of Sir Ian Sinclair's warning of the “risk that the placing of undue emphasis on the ‘object and purpose’ of a treaty will encourage teleological methods of interpretation [which], in some of its more extreme forms, will even deny the relevance of the intentions of the parties”’.
74
The above findings differ from Schreuer's observation (supra note 7, at 3): ‘[a]mong the principles contained in Article 31 VCLT an interpretation that looks at the treaty's object and purpose is particularly popular’.
75
See, e.g., M.S. McDougal, H.D. Lasswell, and J.C. Miller, The Interpretation of International Agreements and World Public Order. Principles of Content and Procedure (1994), at p. xvi: ‘[t]he primary aim of a process of interpretation by an authorized and controlling community decision-maker can be formulated in the following proposition: discover the shared expectations that the parties to the relevant communication succeeded in creating in each other’.
76
The main examples include Salini v. Jordan (award), at paras 117–118, Plama v. Bulgaria (jurisdiction), at paras 197, 207–209, 212, and 219, SGS v. Pakistan, at paras 167 and 173, Impregilo v. Pakistan, at para. 135, Sempra v. Argentina, at paras 52, 69, 111, 142, and 145, Camuzzi v. Argentina 1, at paras 100 and 134, Aguas v. Bolivia, at para. 153, Telenor v. Hungary, at para. 95, Inceysa v. El Salvador, at paras 177–178, 198, and 200, ADF v. US, at para. 195, Técnicas v. Mexico, at para. 156, and Wena v. Egypt (jurisdiction), at sections IV.C and IV.D.
77
See, e.g., Salini v. Jordan (award), at paras 117–118 and Wena v. Egypt (jurisdiction), at section IV.D.
78
See Inceysa v. El Salvador, at para. 200: ‘[s]o, after analysing the intent of Spain and El Salvador obvious in the travaux preparatoires of the Agreement, we must look at its own terms’.
79
For a discussion of these principles in relation to ICSID case law see Schreuer, supra note 7, at 4–6. He restricts his analysis to cases concerning the interpretation of ‘umbrella clauses’ and refers to case law that supports both principles.
80
Some tribunals use the Latin phrase ut magis valeat quam pereat: see, e.g., Pan American v. Argentina, at para. 132.
81
Possible links can be found in SGS v. Philippines, at para. 116, Feldman v. Mexico (award), at para. 183, Mondev v. US, at para. 91, and Noble v. Romania, at para. 52.
82
In addition to cases in which tribunals rejected the principles of both effectiveness and restrictiveness (see infra note 87), the principle of effectiveness was explicitly rejected in Banro v. Congo, at para. 6.
83
Arguments related to the effet utile of the provision to be interpreted were found in Salini v. Jordan (jurisdiction), at para. 95, Técnicas v. Mexico, at para. 156, Maffezini v. Spain (jurisdiction), at para. 36, CSOB v. Slovakia (jurisdiction 1), at paras 39 and 67, SGS v. Pakistan, at para. 172, Pan American v. Argentina, at para. 132, El Paso v. Argentina, at para. 110, Noble v. Romania, at paras 50–54, and Inceysa v. El Salvador, at para. 202. Arguments relating to the effet utile of other specific provisions were found in Siemens v. Argentina, at paras 87–90, El Paso v. Argentina, at para. 109, and SGS v. Pakistan, at para. 150. Arguments relating to the effet utile of the treaty in question were found in PSEG v. Turkey, at para. 162, Continental v. Argentina, at para. 80, Pan American v. Argentina, at paras 93 and 105, El Paso v. Argentina, at para. 76, Camuzzi v. Argentina 1, at para. 56, and Sempra v. Argentina, at paras 69 and 94.
84
Effet utile arguments are closely related to contextual arguments: see infra section 6.
85
El Paso v. Argentina, at para. 77, SGS v. Pakistan, at para. 171, and Noble v. Romania, at para. 55.
86
Metalpar v. Argentina, at para. 92, Interagua v. Argentina (jurisdiction), at paras 59 and 64, Loewen v. US (jurisdiction), at para. 51, Vivendi v. Argentina, at paras 61 and 66, Aguas v. Bolivia, at para. 91, and Continental v. Argentina, at para. 80. In this context, Schreuer, supra note 7, at 4, states that: ‘[t]he idea that treaty provisions that constitute derogations from sovereignty call for a restrictive interpretation is usually rejected nowadays’.
87
Mondev v. US, at para. 43, Wena v. Egypt (annulment), at para. 18, Siemens v. Argentina, at para. 81, El Paso v. Argentina, at paras 68–70, Pan American v. Argentina, at paras 97–99, Mitchell v. Congo (annulment), at para. 19, Interagua v. Argentina (jurisdiction), at para. 64, and Vivendi v. Argentina (jurisdiction), at para. 66.
88
See, concerning inconsistencies, Maffezini v. Spain (provisional measures), at para. 9 and Feldman v. Mexico (jurisdiction), at para. 44, and, concerning mutually supportive rules, Inceysa v. El Salvador, at para. 206 (‘harmonious interpretation of the Agreement’).
89
An example of an a contrario argument can be found in Cement Shipping v. Egypt, at para. 87: ‘[w]hile that provision [Art. 11 of the BIT] requires the application of additional provisions of the national law if more favorable for the investor … by argumentum a contrario it does not permit application of provisions of national law limiting any claims found by the Tribunal to exist under the BIT’.
90
See RFCC v. Morocco (award), at para. 40, Fireman's Fund v. Mexico, at para. 95, ADF v. Hungary, at paras 131 and 133, Tokelės v. Ukraine (jurisdiction), at para. 77, and Aguas v. Bolivia, at paras 145–148, 240, and 242–243.
91
See CGE v. Argentina (annulment), at para. 69, Waste Management v. Mexico 2 (award), at para. 144, Siemens v. Argentina, at paras 85–86, Salini v. Jordan (jurisdiction), at paras 77–79, and CMS v. Argentina (award), at para. 290.
92
For an example see Continental v. Argentina, at paras 78–79.
93
For examples see SGS v. Pakistan, at paras 169–170 and Plama v. Bulgaria (jurisdiction), at para. 192.
94
For an example see ADF v. US, at paras 164–165.
95
For examples see Lanco v. Argentina, at para. 26 and Banro v. Congo, at para. 21.
96
See Biwater v. Tanzania (provisional measures 2), at para. 123.
97
For the purpose of this article, ICSID rules and regulations were classified as contextual arguments. However, it remains unclear how they should classified: see Vivendi v. Argentina (preliminary issues), at para. 13, where the tribunal stated: ‘the unanimous adoption of Arbitration Rule 53 can be seen, if not as an actual agreement by the States parties to the Convention as to its interpretation [see Art. 31(3) of the VCLT], at least as amounting to subsequent practice relevant to its interpretation’. In almost all cases where rules and regulations were mentioned they were applied directly and not used as interpretive arguments.
98
The numbers were as follows: Preamble – six decisions (CSOB v. Slovakia (jurisdiction 1), at para. 64, Salini v. Morocco, at para. 52, RFCC v. Morocco (jurisdiction), at para. 60, Bayindir v. Pakistan, at para. 137, Mitchell v. Congo, at para. 32, and Metalclad v. Mexico, at para. 71), annexes etc. – four decisions (CMS v. Argentina (jurisdiction), at para. 119, CMS v. Argentina (award), at para. 362, Siemens v. Argentina, at paras 85–86, and Fireman's Fund v. Mexico at, paras 86–90), and rules and regulations – two decisions (ADF v. US, at para. 144 and Biwater v. Tanzania (provisional measures 2), at paras 123 and 128).
99
See Aguas v. Bolivia, at paras 276–279, Gas Natural v. Argentina, at para. 30, and Pan American v. Argentina, at para. 50.
100
See ADC v. Hungary, at para. 291 and CGE v. Argentina (award), at n. 16.
101
Contextual arguments between BITs will be addressed infra in section 13 (state practice). Contextual arguments between a BIT or the ICSID Convention on the one hand and customary international law or general principles of law on the other will be studied infra in sections 8 and 9 respectively. Even broader contextual arguments based on the general function of international investment law are addressed infra in section 7 (object and purpose) and section 16 (reasonableness).
102
For an example see Fireman's Fund v. Mexico, at para. 95.
103
For an example see Fireman's Fund v. Mexico, at paras 79 and 86–90.
104
For an example see SGS v. Philippines, at para. 132b.
105
Schreuer, supra note 7, at 3.
106
While tribunals referred to the preamble as a source in 13 decisions, they omitted to make any reference in 31 decisions (including two in which the tribunals made only general references to what they inferred from the treaty or the general regime for investment protection).
107
The latter use of object and purpose is closely related to the use of reasonableness as an interpretive argument: see infra section 16.
108
Such references were found in 12 decisions.
109
Such references were found in 6, 3, and 4 decisions, respectively.
110
Such references were found in 37 decisions.
111
Such references were found in 2, 3, and 16 decisions, respectively.
112
Such references were found in six decisions. For an example see CSOB v. Slovakia (jurisdiction 1), at para. 57.
113
An outstanding example is ADF v. US, at para. 147 where the tribunal made a general assessment of how to use the object and purpose under NAFTA in response to the claimant's request to address the issue.
114
This was done in 18 decisions.
115
The numbers were: essential argument – 31 decisions and non-essential argument – 13 decisions.
116
See Maffezini v. Spain (award), at para. 64, Autopista v. Venezuela, at para. 138, Waste Management v. Mexico 2 (award), at para. 116, and LESI v. Algeria, at para. 83(iii).
117
See CGE v. Argentina (preliminary issue), at para. 11 and CMS v. Argentina (award), at paras 353–356.
118
See Siemens v. Argentina, at para. 92: ‘[w]hile these considerations may follow a strict logical reasoning based on the terms of the Treaty, their result does not seem to accord with its purpose. More consistent with it is to consider that, in Article 3, treatment of the investments includes treatment of the investors and hence the need to provide for exceptions that refer to them. In the same vein, the reference to investors and investments in Article 4 is a matter of emphasis, not of exclusion’.
119
Cases where treaty provisions make explicit references to international law, areas in which the treaty does not address the issue (legal lacunae), or cases where it may be argued that subsequent customary international law has replaced treaty provisions are addressed supra in section 3B. Mondev v. US, at paras 111–125, is a borderline case, and may serve to illustrate the distinction between the use of customary international law as an interpretive argument and as a separate legal basis (source of law). Here, this decision has been classified as ‘separate legal basis’ due to the reference to customary international law in Art. 1105 of NAFTA and the interpretation issued by the NAFTA Free Trade Commission. See also Loewen v. US (award), at paras 129–130.
120
See in particular infra section 13 on state practice. An example is the interpretation of the term ‘investment’ in Art. 25(1) of the ICSID Convention. A decision which is classified as ‘practice’ is Mihaly v. Sri Lanka: see para. 33: ‘the definition [of “investment” in the ICSID Convention] was left to be worked out in the subsequent practice of states, thereby preserving its integrity and flexibility and allowing for future progressive development of international law on the topic of investment’. See also para. 58.
121
Loewen v. US (jurisdiction), at para. 73 is an example of the latter: ‘[w]e accept that an important principle of international law should not be held to have been tacitly dispensed with by an international agreement, in the absence of words making clear an intention to do so … Such an intention may, however, be exhibited by express provisions which are at variance with the continued operation of the relevant principle of international law’.
122
See also Art. 1131(1) of NAFTA: ‘[a] Tribunal established under this Section shall decide the issues in dispute in accordance with this Agreement and applicable rules of international law’.
123
In addition, there were two decisions in which such arguments were rejected: see Aguas v. Bolivia. at para. 235 and SGS v. Pakistan, at para. 177.
124
Such issues include: the nationality of the claimant (Feldman v. Mexico (jurisdiction), at paras 30–36, Olguín v. Paraguay (award), at para. 62, and Tokelės v. Ukraine (jurisdiction), at para. 70), jurisdiction in relation to domestic court decisions (Loewen v. US (jurisdiction), at paras 45–47 and 61–74), submission of claim (Waste Management 2 (jurisdiction), at paras 36–37), the definition of investor (CMS v. Argentina (jurisdiction), at paras 43–45 and 48), agreement to jurisdiction (Inceysa v El Salvador, at paras 179–181, Sempra v. Argentina, at para. 156, and Camuzzi v. Argentina 1, at para. 144), jurisdiction of the annulment committee (CDC v. Seychelles (annulment), at para. 36), retroactivity (Mondev v. US, at para. 98), general jurisdictional issues (AES v. Argentina, at para. 23), and procedural issues (Goetz v. Burundi, at paras 54–57).
125
Such issues include: fair and equitable treatment (Feldman v. Mexico (award), at paras 139–141, Noble v. Romania, at paras 164–165, RFCC v. Morocco (award), at para. 51, and Azurix v. Argentina (award), at paras 361–373), discrimination (Feldman v. Mexico (award), at para. 137 and Genin v. Estonia, at para. 368), arbitrariness (LGE v. Argentina (award), at para. 157), expropriation (Feldman v. Mexico (award), at paras 103–108, LGE v. Argentina (award), at paras 185–197, and Técnicas v. Mexico, at para. 116), MFN clauses (Maffezini v. Spain (jurisdiction), at paras 46–50), umbrella clauses (Noble v. Romania, at paras 53–55), compensation (Impregilo v. Pakistan (jurisdiction), at para. 154), state of necessity (LGE v. Argentina (award), at paras 245–258 and CMS v. Argentina (award), at para. 374), and attributability to the state (ADF v. US, at para. 166 and Metalclad v. Mexico, at para. 73).
126
See in particular Azurix v. Argentina (award), at paras 361–373.
127
See Lanco v. Argentina, at para. 27, CSOB v. Slovakia (jurisdiction 1), at para. 39, Técnicas v. Mexico, at para. 63, SGS v. Philippines, at para. 166, Plama v. Bulgaria (jurisdiction), at para. 140, Impregilo v. Pakistan, at para. 148, and Metalclad v. Mexico, at para. 100.
128
See Wena v. Egypt (interpretation), at paras 77–79, Casado v. Chile, at para. 2, and Goetz v. Burundi, at paras 54–56.
129
See section 3B above.
130
The concept ‘general principles of law’ is defined infra in section 3A, and the use of such principles as a separate legal basis is addressed in section 3C.
131
See Lanco v. Argentina, at para. 46, Técnicas v. Mexico, at para. 119, and Olguín v. Paraguay (award), at paras 83–84. In addition, we may mention Fedax v. Venezuela, at para. 30, in which it remains unclear which role general principles of law had.
132
The only explicit references to analogies found in the decisions examined were in ADF v. US, at para. 144 (see also para. 135), Maffezini v. Spain (jurisdiction), at para. 79, Mondev v. US, at paras 142–143 and SGS v. Philippines, at n. 95.
133
Such arguments are often referred to as expressio unius est exclusio alterius or as the ‘expressio unius principle’.
134
Supra note 7, at 6–7.
135
See para. 203: ‘[t]his shows that in NAFTA and probably in the FTAA the incorporation by reference of the dispute settlement provisions set forth in other BITs is explicitly excluded. Yet, if such language is lacking in an MFN provision, one cannot reason a contrario that the dispute resolution provisions must be deemed to be incorporated. The specific exclusion in the draft FTAA is the result of a reaction by States to the expansive interpretation made in the Maffezini case. That interpretation went beyond what State Parties to BITs generally intended to achieve by an MFN provision in a bilateral or multilateral investment treaty.’
136
See also Siemens v. Argentina, at para. 140 and Enron v. Argentina (jurisdiction 1), at paras 46–47.
137
Mitchell v. Congo, at para. 43.
138
See, e.g., Waste Management v. Mexico 2 (award), at para. 85.
139
See, e.g., Casado v. Chile, at para. 85.
140
See infra, section 13B.
141
Waste Management v. Mexico 2 (preliminary issues), at para. 21, CMS v. Argentina (award), at paras 368–370, Waste Management v. Mexico 2 (award), at para. 85, Casado v. Chile,at paras 85–86, Cement Shipping v. Egypt, at para. 87, SGS v. Pakistan, at paras 151–152 and 176–177, SGS v. Philippines, at para. 132(b), Fireman's Fund v. Mexico, at para. 84, Tokelės v. Ukraine (jurisdiction), at paras 30 and 33–36, ADC v. Hungary, at paras 292 and 359, and Mitchell v. Congo, at para. 43.
142
MTD v. Chile, at para. 104, LESI v. Algeria 1, at section II, para. 25.ii, LESI v. Algeria 2, at para. 84(ii), Plama v. Bulgaria (award), at para. 191, Interagua v. Argentina (jurisdiction), at para. 56, Vivendi v. Argentina (jurisdiction), at para. 58, Nul v. Egypt, at para. 116, Azurix v. Argentina (award), at para. 363, CGE v. Argentina (annulment), at paras 55 and 98, and Aguas v. Bolivia, at para. 234.
143
This category thus overlaps with the interpretive arguments addressed supra in section 6.
144
See Yearbook of the International Law Commission, supra note 68, at 221, para. 14.
145
See A. Aust, Modern Treaty Law and Practice (2000), at 189–194, who states that the agreement ‘must be a clear expression of the intention of the parties’, and T.O. Elias, The Modern Law of Treaties (1974), at 75–76.
146
See Yearbook of the International Law Commission, supra note 68, at 221, para. 13.
147
For an illustration of these two issues see Yaung Chi Oo Trading PTE Ltd v. Government of the Union of Myanmar, ASEAN I.D. Case ARB/01/1, at paras 23–25.
148
See, in this direction, Aguas v. Bolivia, at para. 251: ‘[t]he position taken by Bolivia in this proceeding and the statements made by Ministries of the Government of the Netherlands to the Parliament of the Netherlands, despite the fact that they both relate to the present dispute, are not a “subsequent agreement between the parties.” The coincidence of several statements does not make them a joint statement. And, it is clear that in the present case, there was no intent that these statements be regarded as an agreement.’
149
For an example see Biwater v. Tanzania (provisional measures 2), at paras 117–120, where the tribunal discussed an agreement that was recorded by the ICSID in Minutes of the First Session.
150
ICSID Rules and Regulations are binding documents that tribunals have an obligation to apply directly. Hence, these documents are normally not interpretive arguments. Moreover, they can hardly be regarded as agreements ‘regarding the interpretation’ of the Convention. See, in the same direction, CGE v. Argentina (preliminary issues), at para. 12: ‘the unanimous adoption of Arbitration Rule 53 can be seen, if not as an actual agreement by the States parties to the Convention as to its interpretation, at least as amounting to subsequent practice relevant to its interpretation’. Hence, the tribunal indicates that such a decision could be classified as state practice according to Art. 31(3)(b) of the VCLT.
151
Supra note 40.
152
A parallel can be drawn to the commentaries adopted by the International Law Commission on its draft treaties, available at: www.un.org/law/ilc/index.htm.
153
See on these issues the Yearbook of the International Law Commission, supra note 68, at 220, para. 10.
154
See in particular paras 22, 26, 27, 29, and 40.
155
See ICSID, ICSID Convention, Regulations and Rules, ICSID doc. ICSID/15, 2006.
156
See Lanco v. Argentina, at paras 42–43 and 47, CSOB v. Slovakia (jurisdiction 1), at paras 63–64, Wena v. Egypt (jurisdiction), at sections IV.A and V, Goetz v. Burundi, at paras 80–83, Banro v. Congo, at para. 20, Salini v. Morocco, at para. 51, RFCC v. Morocco (jurisdiction), at para. 59, Wena v. Egypt (annulment), at para. 38, Azurix v. Argentina (jurisdiction), at para. 58, Tokelės v. Ukraine (jurisdiction), at paras 19 and 99, LGE v. Argentina (jurisdiction), at para. 71, PSEG v. Turkey, at para. 137, Joy Mining v. Egypt, at para. 42, LESI v. Algeria 1, at part II, para. 13.i, Plama v. Bulgaria (jurisdiction), at para. 193, Impregilo v. Pakistan, at para. 146, Gas Natural v. Argentina, at para. 20, Aguas v. Bolivia, ar para. 278, Bayindir v. Pakistan, at para. 125, Pan American v. Argentina, at para. 54, Interagua v. Argentina (jurisdiction), at para. 35, Vivendi v. Argentina (jurisdiction), at para. 35, Nul v. Egypt, at para. 74, LESI v. Algeria 2, at para. 72, Continental v. Argentina, at para. 69, and Inceysa v. El Salvador, at para. 167.
157
See infra section 14. The Report has in particular been used as an interpretive argument in recent cases, i.e., cases from 2003–2006.
159
See in particular sections II.A, III.C, III.D, V.A and VII.
160
The only cases in which they were mentioned are Casado v. Chile, at para. 68 and CDC v. Seychelles, at para. 53.
161
See Art. 15.26 of the United States–Singapore Free Trade Agreement (2003), available at: www.ustr.gov/Trade_Agreements/Bilateral/Singapore_FTA/Section_Index.html.
162
See supra section 6.
163
See, e.g., Arts 30.3 and 31.2 of the 2004 Model BIT of the USA, available at: www.ustr.gov/assets/Trade_Sectors/Investment/Model_BIT/asset_upload_file847_6897.pdf.
164
Schreuer, supra note 7, at 19 states that ‘[p]lans to create institutionalized mechanisms to achieve uniform interpretations have yielded limited results so far’. However, one example can be found in Art. 41 of the Agreement between Canada and the Republic of Peru for the Promotion and Protection of Investments (2006), supra note 29.
165
The distinction between amendments and interpretations may be difficult to draw. Methanex v. US (award), at pt IV, ch. C, paras 20–23, indicates that states have a broad discretion when choosing the form in which to adopt an instrument.
166
See Gruslin v. Malaysia, at para. 23, which concerned subsequent agreement through exchange of letters concerning the concept ‘approved investment’. See para. 23.4: ‘at this peak level of intercourse between states it must be regarded as an enduring and authoritative engagement expressing to the Belgo-Luxemburg Union the manner in which the Respondent regards and applies the terms of the IGA with regard to investments made in its territory by nationals of the Belgo-Luxemburg Union’. See also Plama v. Bulgaria (jurisdiction), at para. 195 where the tribunal took into account the failed negotiation of amendments to a BIT.
167
See Annex 2 to the Final Act of the European Energy Charter Conference, Energy Charter Secretariat, The Energy Charter Treaty and Related Documents (2004), at 133.
168
NAFTA Free Trade Commission, Notes of Interpretation of Certain Chapter 11 Provisions, adopted 31 July 2001. This interpretation shall, in accordance with Art. 1131(2) of NAFTA, ‘be binding on a Tribunal established under this Section’.
169
See the UNCITRAL cases of Pope & Talbot Inc. v. Government of Canada, award in respect of damages, at paras 48–65 and Methanex Corporation v. United States of America, final award of the tribunal on jurisdiction and merits, at pt II, paras 12–14 and 19–21, and at pt IV, ch. C, paras 10 ff. These cases are available at: http://ita.law.uvic.ca/. See the ICSID cases of Mondev v. US, at paras 119–125, ADF v. US, at para. 177, Loewen v. US (award), at paras 125–126, and Waste Management v. Mexico 2 (award), at paras 90–97.
170
Among the few ICSID tribunals that mention the Statute of the ICJ is that in Mihaly v. Sri Lanka, at para. 58: ‘[o]nly subject to Article 59 of the Statute of the International Court of Justice are judicial decisions to be considered as such subsidiary sources of law’. See also Sempra v. Argentina, at para. 147 and Camuzzi v. Argentina 1, at para. 135 (identical): ‘[i]t is also the duty of tribunals called upon to settle a dispute, particularly when the question is to interpret the meaning of the terms used in a treaty. This is precisely the role of judicial decisions as a source of international law in Article 38(1) of the Statute of the International Court of Justice …’.
171
Schreuer, supra note 7, at 10. See also Commission, supra note 1, at 130–131.
172
At para. 293. See also Camuzzi v. Argentina 2, at para. 19, El Paso v. Argentina, at para. 39, SGS v. Philippines, at para. 97, Feldman v. Mexico (award), at para. 107, and Pan American v. Argentina, at para. 42.
173
For examples see AES. v. Argentina, at paras 23–33, Bayindir v. Pakistan, at para. 76, Nul v. Egypt, at paras 63–64, and Metalpar v. Argentina, at para. 50. See also Schreuer, supra note 7, at 11–15 with further references. Only exceptionally do ICSID tribunals make use of the concept ‘precedent’. For examples see CMS v. Argentina (jurisdiction), at para. 76 and IBM v. Ecuador, at para. 48.
174
SGS v. Philippines, at para. 97.
175
For a more detailed overview of ICSID case law on this issue see Commission, supra note 1, at 144–148.
176
There were only 15 instances in which tribunals made general references to case law of the kind found in Loewen v. US (award), at para. 132: ‘[n]either State practice, the decisions of international tribunals nor the opinion of commentators support the view that bad faith or malicious intention is an essential element of unfair and inequitable treatment or denial of justice amounting to a breach of international justice’.
177
See reg. 22 of the Administrative and Financial Regulations, see International Center for the Settlement of Investment Disputes, supra note 9, at 51. See also Aguas v. Bolivia, at para. 288: ‘[w]ithout access to the full records of these cases, the Tribunal does not believe it possible to assess their significance for the present arbitration’.
178
See, e.g., Waste Management 2 (award), at paras 163–174 and Feldman v. Mexico (award), at paras 143–152.
179
See Feldman v. Mexico (award), at paras 151–152 and Loewen v. US (jurisdiction), at para. 49. Only one decision seems to have placed some emphasis on the fact that a statement was an obiter dictum: see Waste Management v. Mexico 2 (award), at para. 170.
180
The six decisions in which case law was not used as interpretive arguments were CSOB v. Slovakia (jurisdiction 2), Waste Management v. Mexico 1, Gruslin v. Malaysia, Maffezini v. Spain (award), Fireman's Fund v. Mexico, and Soufraki v. UAE.
181
See, e.g., Wena v. Egypt (jurisdiction), at section VI, Metalclad v. Mexico, at para. 108, Autopista v. Venezuela, at para. 108, Nul v. Egypt, at paras 69 and 91, and Bayindir v. Pakistan, at paras 130–138.
182
See, e.g., Salini v. Morocco, at para. 31, RFCC v. Morocco (jurisdiction), at para. 35, Loewen v. US (jurisdiction), at paras 47–49 and 53, Generation v. Ukraine, at paras. 20.25, SGS v. Philippines, at paras 120–126, and Nul v. Egypt, at para. 64.
183
See, e.g., Cement Shipping v. Egypt, at para. 94, where the tribunal found it unnecessary to repeat the reasoning of a previous tribunal in order to arrive at the same conclusions concerning applicable customary international law.
184
See, e.g., CSOB v. Slovakia (jurisdiction 1), at para. 34, Azurix v. Argentina (jurisdiction), at paras 88–89, Banro v. Congo, at paras 11–12, and LGE v. Argentina (award), at paras 159–161.
185
See, e.g., Fedax v. Venezuela, at para. 30 and Banro v. Congo, at para. 1.
186
See, e.g., Autopista v. Venezuela (jurisdiction), at para. 111.
187
See, e.g., CDC v. Seychelles, at para. 90.
188
See, e.g., Casado v. Chile, at paras 53–57 and Telenor v. Hungary, at para. 79. For a general discussion see AES v. Argentina, at paras 18–32. This also touches upon questions regarding cases where similar issues arise; should the cases be consolidated (see Pan American v. Argentina), should they be dealt with separately by the same tribunal (see Interagua v. Argentina and Vivendi v. Argentina), or should they be dealt with separately by different tribunals (see Corn Products v. Mexico)? The approach differs from case to case according to the opinion of the parties and the tribunals. Cases in which essential parts of the decisions are identical include: Salini v. Morocco and RFCC v. Morocco (jurisdiction), El Paso v. Argentina and Pan American v. Argentina, Interagua v. Argentina and Vivendi v. Argentina, and Camuzzi v Argentina 1 and Sempra v. Argentina.
189
See, e.g., SGS v. Philippines, at para. 132, AES v. Argentina, at para. 30, and Gas Natural v. Argentina, at para. 36.
190
For an example see CMS v. Argentina (jurisdiction), at paras 70–76. There may be an ‘upper limit’ for ICSID tribunals’ freedom to rely on case law. As indicated by Schreuer, supra note 7, at 15, ‘an application for annulment that alleges an excess of powers or a failure to state reasons because the tribunal has simply relied on earlier decisions without making an independent decision or developing its own reasons is entirely possible’.
191
For an example see Sempra v. Argentina, at para. 99.
192
As indicated by E. McWhinney, Judicial Settlement of International Disputes. Jurisdition, Justiciability and Judicial Law-Making on the Contemporary International Court (1991), at 16, both the common law and the civil law traditions reflect a willingness to recognize certain cases as ‘leading’. On leading cases under ICSID see Commission, supra note 1, at 154–156, who indicates that more analysis is needed before leading cases will crystallize, and T. Weiler (ed.), International Investment Law and Arbitration: Leading Cases from the ICSID, NAFTA, Bilateral Treaties and Customary International Law (2005).
193
See SGS v. Pakistan, at para. 164 where the tribunal initially emphasized that no case law was available and that it had a ‘case of first impression’.
194
For an example see Casado v. Chile, at paras 53–54.
195
For an example see ADF v. US, at para. 197.
196
Hence, the case law of ICSID tribunals does not seem to support the conclusion that there is a strong esprit de corps among ICSID and other arbitrators as suggested by Commission, supra note 1, at 136–141.
197
One recent example is the protection of shareholders under BITs: see Sempra v. Argentina, at paras 93–94. Another recent example is the conditions for being recognized as an investment under Art. 25 of the ICSID Convention: see Bayindir v. Pakistan, at paras 130–138. Some tribunals have used the civil law concept jurisprudence constante: see SGS v. Philippines, at para. 97 and AES v. Argentina, at para. 32.
198
See El Paso v. Argentina, at paras 69–79, where the tribunal first pointed out the divergence in case law, and subsequently made its own general findings without seeming to consider itself bound to follow a particular line of argument followed in previous case law (see at paras 79–80).
199
A multilateral appeals procedure in the context of ICSID was suggested, but was subsequently dropped: see Schreuer, supra note 7, at 21–22. See also supra note 32.
200
Schreuer, supra note 7, at 11, observes that ‘[e]ach tribunal is constituted ad hoc for the particular case. Therefore, ICSID cannot be expected to act like an international court such as the ICJ or ECHR.’ In a different direction, Commission, supra note 1, at 136–141, concludes that ICSID practice concerning qualifications, nationalities, and frequency of selection of ICSID arbitrators has created ‘an esprit de corps amongst arbitrators in investment treaty cases’.
201
See, e.g., Gas Natural v. Argentina, at para. 36 and AES v. Argentina at para. 30.
202
See Schreuer, supra note 7, at 16–17, with references to case law. Schreuer goes on to observe that ‘[i]n some cases tribunals did not follow earlier decisions but adopt different solutions. At times they simply adopted a different solution without distancing themselves from the earlier decision. At other times they referred to the earlier decision and pointed out that they were unconvinced by what another tribunal had said and that, therefore, their decision departed from the one adopted earlier. … Most tribunals carefully examine earlier decisions and accept these as authority most of the time. But sometimes they disagree with them and make their disagreement known.’ See also Aguas v. Bolivia, at n. 99, CGE v. Argentina (preliminary issue), at para. 22, Mondev v. US, at para. 69, and Franck, ‘The Legitimacy Crisis in Investment Treaty Arbitration: Privatizing Public International Law through Inconsistent Decisions’, 73 Fordham L Rev (2005) 1558.
203
See, e.g., Interagua v. Argentina (jurisdiction), at paras 30, 50, and 62–65, Autopista v. Venezuela, at para. 143, Continental v. Argentina, at para. 75, Nul v. Egypt, at paras 126–131, and Enron v. Argentina (jurisdiction 2), at paras 33–37 and 40–46. It can be noted that there are different grounds for distinguishing cases, including differences with regard to facts, legal basis, and available interpretive materials.
204
See UNCTAD, Dispute Settlement: Investor–State, in UNCTAD Series on issues in international investment agreements, 2003, doc UNCTAD/ITE/IIT/30, available at: www.unctad.org/en/docs/iteiit30_en.pdf, at 47, where it is observed that the provisions of ICSID ‘have been developed through the interpretive jurisprudence of successive ICSID Tribunals into a complex and technical body of procedural law, though it must be stressed that each Tribunal is free to interpret the Convention as it sees fit, there being no doctrine of precedent under the ICSID Convention. However, earlier decisions on admissibility undoubtedly form persuasive precedents upon which the parties and subsequent Tribunals may rely.’
205
No conclusion could be drawn in relation to the ECT as there was only one relevant decision.
206
These are tribunals established according to the 1976 Arbitration Rules adopted by the UN Commission on International Trade Law (UNCITRAL) and recommended by UN GA Res 31/98.
207
Of particular interest here is Drahozal, supra note 3, who carries out a quantitative citation analysis of ICSID tribunals’ references to decisions by the Iran–US Claims Tribunal.
208
See supra, the text accompanying notes 102–104.
209
For each group of decisions the percentage adds up to more than 100% due to the fact that decisions often contained more than one reference to the argument in question.
210
The only instances identified were Mondev v. US, at para. 43 and Loewen v. US (jurisdiction), at para. 51.
211
One possible example is Cement Shipping v. Egypt, at paras 174–175.
212
See, e.g., Metalclad v. Mexico, at paras 108, 120, and 122, RFCC v. Morocco (award), at paras 64–65, Generation v. Ukraine, at paras 20.23–20.25 and Feldman v. Mexico (award).
213
In the following, the case law from these two courts is addressed as one entity under the label ‘ICJ case law’.
214
The numbers are as follows: procedural and jurisdictional: UNCITRAL 63%–ICJ 65%, substantive: UNCITRAL 43%–ICJ 41%, and interpretive methodology: UNCITRAL 7%–ICJ 13%.
215
The argument can be characterized as ‘decisive’, in the sense that it was the only argument or the decisive argument for determining the content of the rule, in 12 of these decisions.
216
In exceptional cases tribunals took into account opinions expressed by individual judges: see in particular Plama v. Bulgaria (jurisdiction), at para. 118 concerning the separate opinion of Judge Higgins.
217
See Técnicas v. Mexico, at paras 146–147, Interagua v. Argentina (jurisdiction), at para. 50, and Vivendi v. Argentina (jurisdiction), at para. 50.
218
See, e.g., Impregilo v. Pakistan, at paras 238–258.
219
See Casado v. Chile, at paras. 2 and 15 and Goetz v. Burundi, at paras 54–56.
220
See Feldman v. Mexico (award), at para. 177, SGS v. Pakistan, at para. 171, ADF v. US, at para. 147, Vivendi v. Argentina (preliminary issues), at para. 15, and Interagua v. Argentina (preliminary issues), at para. 15.
221
There were five decisions referring to the European Court of Human Rights and two referring to the Inter-American Court of Human Rights: see Mondev v. US, at paras 69, 138, and 143–144, Loewen v. US (award), at para. 165, Técnicas v. Mexico, at paras 116 and 122, Azurix v. Argentina (award), at paras 311–312 and 391–392, ADC v. Hungary, at para. 497, and IBM v. Ecuador, at para. 72.
222
See Loewen v. US (jurisdiction), at para. 45, SGS v. Philippines, at para. 171, and IBM v. Ecuador, at para. 72.
223
It has even been argued that tribunals should show some restraint in making use of case law that the parties to the dispute have not commented upon: see Continental v. Argentina, at para. 19: ‘[s]ubsequent to the hearing the Tribunal received a communication from the Claimant pointing to recent ICSID decisions on jurisdiction issued in cases involving Argentina, and an answer from Argentina raising objections as to the relevance of those decisions. The Tribunal informed the parties, through the Secretariat, on July 20, 2005 that “it believes it is empowered to take judicial notice of such published decisions. However, in accordance with due process principles, the Tribunal is of the opinion that should it consider necessary for its decision on jurisdiction to specifically rely on points raised and discussed in those decisions, it should give an opportunity first to the parties to comment on those possibly relevant points. The Tribunal would accordingly do so should the situation envisaged occur”.’
224
Examples include Salini v. Jordan (jurisdiction), at paras 137–156 and Waste Management v. Mexico 2 (jurisdiction), paras 39–45.
225
Examples can be found in Aguas v. Bolivia, at paras 291–293, PSEG v. Turkey, at para. 146, and CMS v. Argentina (award), at para. 317.
226
For a discussion of such acts with further references to case law see Schreuer, supra note 7, at 18.
227
An example is the reference to the practice of Argentina and Germany in their BITs, invoked in Siemens v. Argentina, at paras 105–106. Wena v. Egypt (annulment), paras 42–45 is a borderline example which has been classified as state practice.
228
The only examples of interest were Salini v. Morocco, at para. 52, RFCC v. Morocco (jurisdiction), at para. 60, and Joy Mining v. Egypt, at para. 52, where the tribunals referred to decisions by the Secretary General of ICSID, and CDC v. Seychelles, at para. 53, Casado v. Chile, at para. 25, and ADF v. US, at n. 151, where the tribunals took into account explanatory notes provided by the ICSID Secretariat.
229
The only examples were CGE v. Argentina (preliminary issue), at paras 18 and 20, CMS v. Argentina (jurisdiction), at para. 48, and CMS v. Argentina (award), at para. 402.
230
Mihaly v. Sri Lanka may serve as an illustration. In para. 33, the tribunal observed that there was no definition of ‘investment’ in the ICSID convention and that ‘the definition was left to be worked out in the subsequent practice of states, thereby preserving its integrity and flexibility and allowing for future progressive development of international law on the topic of investment’. The tribunal went on to state in para. 58 that ‘[i]n the absence of a generally accepted definition of investment for the purpose of the ICSID convention, the Tribunal must examine the current and past practice of ICSID and the practice of States as evidenced in multilateral and bilateral treaties and agreements binding on states, notably the US-Sri Lanka BIT’.
231
See, however, Aguas v. Bolivia, at paras 291–293 where state practice is implicitly dealt with as a supplementary means of interpretation, and Gruslin v. Malaysia, at paras 21.2 and 21.6, where, inter alia, state practice is classified as ‘extrinsic sources’.
232
See Lanco v. Argentina, at para. 46, Casado v. Chile, at para. 87, CMS v. Argentina (jurisdiction), at para. 47, CMS v. Argentina (award), at para. 355, Loewen v. US (award), at para. 132, SGS v. Philippines, at para. 167, LESI v. Algeria 1, at para. 32(i), LESI v. Algeria 2, at para. 93(ii), Vivendi v. Argentina (preliminary issues), at para. 17, Interagua v. Argentina (preliminary issues), at para. 17, Bayindir v. Pakistan, at para. 190, Pan American v. Argentina, at para. 214, Salini v. Jordan (award), at para. 94, and Mitchell v. Congo, at para. 52. These references were used as essential arguments in six decisions, as non-essential arguments in five decisions, and as general starting points for the interpretation in four decisions. The references occurred mainly in relation to procedural and jurisdictional issues (nine decisions) and less frequently in relation to substantive issues (five decisions).
233
Noteworthy examples include Maffezini v. Spain (jurisdiction), at paras 52–53 and 57–61, Waste Management v. Mexico 2 (jurisdiction), at paras 29–30, Enron v. Argentina (jurisdiction 1), at paras 46–47, Plama v. Bulgaria (jurisdiction), at paras 201–206, Aguas v. Bolivia, at paras 216–217 and 289–314, Inceysa v. El Salvador, at paras 184–189, and Telenor v. Hungary, at paras 95–96.
234
The instruments referred to were NAFTA and the ECT and instruments from UNCITRAL, the Iran–US Claims Tribunal, and ASEAN. See SGS v. Pakistan, at paras 176–177, Loewen v. US (award), at paras 229–230, Tokelės v. Ukraine (jurisdiction), at paras 33–36, Salini v. Jordan (award), at para. 73, LESI v. Algeria 1, at para. 25(ii), LESI v. Algeria 2, at para. 84(ii), Plama v. Bulgaria (jurisdiction), at paras 155–157, 190, 195, and 198–206, CMS v. Argentina (award), at paras. 276, 283, and 368–370, Azurix v. Argentina (award), at paras 362–363, CGE v. Argentina (annulment), at paras 55 and 98, Biwater v. Tanzania (provisional measures 2), at paras 117–125, 130–131, 132, and 135, Vivendi v. Argentina (preliminary issue), at para. 25, Interagua v. Argentina (preliminary issue), at para. 24, and Waste Management v. Mexico 2 (preliminary issue), at para. 21.
235
These included the WTO Agreement, the GATT, the OECD Draft Convention on the Protection of Foreign Property (1967), the Locarno Arbitration Convention (1925), UNCLOS (1982), treaties on humanitarian law, dispute settlement arrangements under international law, a World Bank directive, and practice of the UN Security Council. See Waste Management v. Mexico 2 (jurisdiction), at para. 49, Loewen v. US (award), at para. 150, CMS v. Argentina (award), at paras 353, 363–364, and 370, Vivendi v. Argentina (preliminary issue), at para. 22, Interagua v. Argentina (preliminary issue), at para. 21, PSEG v. Turkey, at para. 160, Mitchell v. Congo, at para. 53, El Paso v. Argentina, at para. 61, and Pan American v. Argentina, at para. 80. All references to VCLT and the ICJ Statute were omitted here due to their close link to customary international law.
236
Among these decisions, only one reference related to a substantive issue: see CMS v. Argentina (award), at 370, which referred to the GATT when interpreting a safeguards clause. The other decisions were Vivendi v. Argentina (preliminary issue), at para. 22, and Interagua v. Argentina (preliminary issue), at para. 21.
237
The only statement of interest was found in Gas Natural v. Argentina, at n. 12: ‘[t]he Tribunal notes Argentina's argument that Spain's position in the Maffezini case reflects understanding of the Spain-Argentina BIT consistent with that of Argentina in this case. We do not believe, however, that an argument made by a party in the context of an arbitration reflects practice establishing agreement between the parties to a treaty within the meaning of Article 31(3)(b) of the Vienna Convention on the Law of Treaties.’
238
See, however, LGE v. Argentina (award), at para. 213, which indicates that one must avoid using a development in practice as an argument in favour of rights or duties not explicitly set out in the treaties.
239
This seems to support the observations of Schreuer, supra note 7, at 7, that ‘[t]he large number of BITs, often containing similar or identical provisions, lends itself to a comparative approach. Especially the BITs of the host State but also of the investor's home State with third countries often lead to extensive comparisons and inferences’.
240
See section 10 above.
241
See SGS v. Pakistan, at paras 176–177, Enron v. Argentina (jurisdiction 1), at paras 46–47, Tokelės v. Ukraine, at paras 33–36, LESI v. Algeria 1, at para. 25(ii), LESI v. Algeria 2, at para. 84(ii), CMS v. Argentina (award), at paras 368–370, ADC v. Hungary, at para. 359, Azurix v. Argentina (award), at para. 363, and CGE v. Argentina (annulment), at paras 55 and 98.
242
For an example see Azurix v. Argentina (award), at para. 363: ‘[t]he interpretation of the FTC or the examples of FTAs adduced by the Respondent may be evidence of a significant practice by one of the parties to the BIT, but the Tribunal has difficulty in reading it in the text of the BIT which governs these proceedings. The fact that the FTC interpreted Article 1105 in reaction to a tribunal's different understanding of this article and that, in recent agreements, the correlative clause has been drafted to reflect the FTC's interpretation show that the meaning of that article and similar clauses in other agreements could reasonably be understood to have a different meaning.’
243
See Enron v. Argentina (jurisdiction 1), at para. 46: ‘[t]he fact that a treaty may have provided expressly for certain rights of shareholders does not mean that a treaty not so providing has meant to exclude such rights if this can be reasonably inferred from the provisions of such treaty. Each instrument must be interpreted autonomously in the light of its own context and in the light of its interconnections with international law.’
244
See, e.g., Azinian v. Mexico, at paras 84 and 96–97, where the tribunal concludes that ‘[a]s the Mexican courts found that the Ayuntamiento's decision to nullify the Concession Contract was consistent with the Mexican law governing the validity of public service concessions, the question is whether the Mexican court decisions themselves breached Mexico's obligations under Chapter Eleven’.
245
See, e.g., Feldman v. Mexico, in particular at paras 79–84, where the tribunal discussed Mexican court decisions of direct relevance to the case.
246
Enron v. Argentina (jurisdiction 1), at para. 46, CMS v. Argentina (award), at para. 368, El Paso v. Argentina, at para. 80, Pan American v. Argentina, at para. 108, Mitchell v. Congo, at para. 32.
247
CSOB v. Slovakia (jurisdiction 1), at para. 44, Feldman v. Mexico (award), at para. 181, Generation v. Ukraine, at paras 15.4–15.7, CMS v. Argentina (jurisdiction), at para. 82, CMS v. Argentina (award), at paras 362 and 369, ADF v. US, at para. 195, and Metalpar v. Argentina, at paras 58–65.
248
Waste Management v. Mexico 2 (jurisdiction), at para. 28, Enron v. Argentina (jurisdiction 2), at paras 34, 36, and 39, Sempra v. Argentina, at para. 123, ADF v. US, at para. 179, and Camuzzi v. Argentina 1, at para. 112.
249
Legislative: Feldman v. Mexico (award), at para. 58, Aguas v. Bolivia, at para. 245, and Wena v. Egypt (annulment), at 42–45. Executive: ADF v. US, at para. 188, PSEG v. Turkey, at paras 144 and 146, Aguas v. Bolivia, at paras 248–263, and Telenor v. Hungary, at para. 78. Judiciary: Feldman v. Mexico (award), at para. 133, CSOB v. Slovakia (jurisdiction 2), at para. 29, Maffezini v. Spain (jurisdiction), at para. 54, CGE v. Argentina (preliminary issues), at para. 24, Casado v. Chile, at para. 87, Mondev v. US, at paras 133 and 143, Waste Management v. Mexico 2 (jurisdiction), at paras 44–45, Waste Management v. Mexico 2 (award), at paras 154 and 159, Loewen v. US (award), at para. 123, Enron v. Argentina (jurisdiction 2), at paras 38–39, SGS v. Philippines, at para. 170, Vivendi v. Argentina (preliminary issues), at paras 19–20, and Interagua v. Argentina (preliminary issues), at paras 18–19.
250
The tribunal rejected the argument in the only case in which a different model treaty was invoked: see Siemens v. Argentina, at para. 106 concerning the German Model Treaty.
251
Sempra v. Argentina, at para. 146. See also Enron v. Argentina (jurisdiction 1), at para. 48 and Enron v. Argentina (jurisdiction 2), at para. 39.
252
See rule 37(2) of the Arbitration Rules (‘non-disputing party’) and Art. 1128 of NAFTA (‘third party submission’). The latter has been frequently used in practice.
253
For an example relating to the distinction between preparatory work and state practice or subsequent agreement between states see Inceysa v. El Salvador, at paras 192–196, referring to ‘communications exchanged between El Salvador and Spain days before the entry into force of the Agreement’, which was regarded as preparatory work.
254
The tribunals used different concepts when referring to preparatory work, including ‘legislative history’, ‘negotiations’, a reference to the ‘drafters’, and more specific references to drafts.
255
See Mondev v. US, at para. 111: ‘[w]hether or not explanations given by a signatory government to its own legislature in the course of ratification or implementation of a treaty can constitute part of the travaux préparatoires of the treaty for the purposes of its interpretation, they can certainly shed light on the purposes and approaches taken to the treaty, and thus can evidence opinio juris’. See also Generation Ukraine v. Ukraine, at paras 15.4–15.7.
256
References to the Report of the Executive Directors are not included here: see supra section 12B.
257
CSOB v. Slovakia (jurisdiction 1), at para. 16, CGE v. Argentina (preliminary issues), at paras 9 and 12, CGE v. Argentina (award), at paras 52, 55, 57, and 79, CGE v. Argentina (annulment), at para. 69, Banro v. Congo, at paras 16 and 20, Autopista v. Venezuela, at paras 97, 106, and 112–113, Casado v. Chile, at paras 2 and 15, Waste Management v. Mexico 2 (preliminary issues), at para. 16, Mondev v. US, at para. 123, Feldman v. Mexico (award), at para. 165, Wena v. Egypt (annulment), at para. 40, CMS v. Argentina (jurisdiction), at para. 50, Fireman's Fund v. Mexico, at paras 70, 74, and 83, SGS v. Philippines, at paras 131 and 146, PSEG v. Turkey, at paras 137–138, Plama v. Bulgaria (jurisdiction), at paras 196–197, AES v. Argentina, at para. 79, Sempra v. Argentina, at para. 142, Camuzzi v. Argentina 1, at para. 134, Gas Natural v. Argentina, at paras 20 and 29, CDC v. Seychelles, at para. 36, Aguas v. Bolivia, at paras 235, 268–274, 280, 283–284, and 332, LGE v. Argentina (award), at paras 95 and 212, and Inceysa v. El Salvador, at paras 192–196.
258
An example of detailed discussion can be found in Inceysa v. El Salvador, at paras 192–196.
259
For NAFTA: three decisions constituting 23% of the total number of decisions in which NAFTA was invoked, for ICSID 17 decisions constituting 17% of the decisions, and for BITs eight decisions constituting 10% of the decisions.
260
See, e.g., Wena Hotels v. Egypt (jurisdiction), at section IV.C: ‘[n]o documents, such as the travaux préparatoires, that might assist in interpreting Article 8(1) are available. Accordingly, the Tribunal can only rely upon third party commentary and its own interpretation of the provision to determine the intent of the United Kingdom and Egypt in consenting to bring disputes under ICSID jurisdiction.’ See also Generation v. Ukraine, at para. 15.3 and Aguas v. Bolivia, at para. 268.
261
As to the use of preparatory work under other agreements see Plama v. Bulgaria, at para. 160 concerning the ECT and Yaung Chi Oo Trading PTE Ltd. v. Government of the Union of Myanmar, ASEAN I.D., Case ARB/01/1, at para. 80 concerning the ASEAN Agreement for the Promotion and Protection of Investments.
262
The numbers were: jurisdictional and procedural issues – 23 decisions and substantive issues – three decisions.
263
There were only three such decisions.
264
For an example see ADF v. US, at para. 195. However, it was often difficult to distinguish clearly between intent on the one hand and object and purpose on the other see CGE v. Argentina (award), at para. 52.
265
See Aguas v. Bolivia, at para. 235.
266
Examples are Sempra v. Argentina, at para. 142, Camuzzi v. Argentina 1, at para. 134, and Aguas v. El Salvador, at paras 268–274. See also Enron v. Argentina (jurisdiction 2), at para. 32.
267
The numbers were: essential argument – 12 decisions, non-essential argument – 14 decisions, and starting point – seven decisions.
268
Schreuer, supra note 7, at 9. The ICSID's preparatory work was made available in ICSID, History of the ICSID Convention. Documents Concerning the Origin and the Formulation of the Convention on the Settlement of Investment Disputes between States and Nationals of Other States (1968, reprinted 2001). Calls for the publication of the preparatory work to NAFTA finally resulted in the release in 2004 of a number of documents, available at: www.naftaclaims.com/commission.htm.
269
See, e.ge., Tokelės v. Ukraine (jurisdiction), at paras 91 and 92, and SGS v. Philippines, at para. 146.
270
See, e.g., Casado v. Chile, at para. 82, and CGC v. Argentina (annulment), at para. 98.
271
As applied here, this distinction has been based on how tribunals referred to the doctrine, and not on consulting the texts referred to. This approach may arguably detract from the accuracy of the findings. However, the way in which tribunals referred to the doctrine was generally a sufficiently clear indication of the role legal doctrine played.
272
See Loewen v. US (award), at paras 139, 147, and 151, MTD v. Chile, at para. 109, PSEG v. Turkey, at para. 138, CMS v. Argentina (award), at paras 364–365, 367–368, and 392, PSEG v. Turkey, at para. 138, and Nul v. Egypt, at paras 91, 121, and 135.
273
The numbers were: jurisdictional and procedural issues – 49 decisions and substantive issues – 35 decisions.
274
See Casado v. Chile, at para. 2, Plama v. Bulgaria (jurisdiction), at para. 193, and CDC v. Seychelles, at para. 33.
275
See Feldman v. Mexico (award), at paras 99, 101, 104–106, 137, and 170, Loewen v. US (jurisdiction), at para. 57, Técnicas v. Mexico, at para. 115, and LGE v. Argentina (award), at para. 196.
276
See Continental v. Argentina, at para. 81 (UNCTAD publications), and WDFC v. Kenya, at para. 139 (International Law Association report).
277
See, e.g., Salini v. Jordan (jurisdiction), at para. 115, and CMS v. Argentina (jurisdiction), at para. 56.
278
For noteworthy examples see Tokelės v. Ukraine (jurisdiction), at paras 67–69, and CDC v. Seychelles, at paras 33–37.
279
The numbers were as follows: essential argument – 48 decisions, non-essential argument – 39 decisions, and starting point – 37 decisions.
280
The following may serve as examples: CSOB v. Slovakia (jurisdiction 1), at paras 17, 46, and 47, Maffezini v. Spain (jurisdiction), at paras 79 and 94, Wena v. Egypt (award), at paras 99–100, CGE v. Argentina (preliminary issue), at para. 24, Tokelės v. Ukraine (jurisdiction), at para. 106, Joy Mining v. Egypt, at para. 53, CDC v. Seychelles, at para. 41, and Telenor v. Hungary, at para. 70.
281
On the practice of the ICJ see Rosenne, supra note 1, at 1558: ‘[w]ith regard to the “teachings of the most highly qualified publicists” … of the various nations, both Courts are very reticent in direct citation of named publicists in support of any proposition of law’. As regards the Appellate Body of the WTO, it does sometimes refer to legal doctrine in the context of general issues of international law. References to legal doctrine when interpreting specific provisions of the treaties covered by the Agreement establishing the World Trade Organization are very rare.
282
See, e.g., C.W. Jenks, The Prospects of International Adjudication (1964), at chs 7 and 8 and Weil, ‘L’équité dans la jurisprudence de la Cour Internationale de Justice. Un mystère en voie de dissipation?’ in V. Lowe and M. Fitzmaurice (eds), Fifty Years of the International Court of Justice (1996), at 121–144.
283
For the main examples see Azinian v. Mexico, at paras 122–124, Genin v. Estonia, at paras 345, 347, 348, and 370, Feldman v. Mexico (award), at paras 129 and 136, Loewen v. US (award), at paras 241–242, Técnicas v. Mexico, at paras 148–150, ADF v. US, at para. 136, SGS v. Philippines, at para. 162, LESI v. Algeria 1, at para. 40, and Impregilo v. Pakistan, at paras 289–290. A borderline example which has been regarded as relevant here is Mondev v. US, at para. 136.
284
Main examples are the rules on determination of damages or the division of costs: see, e.g., Generation v. Ukraine, at paras 24.1–24.8, ADC v. Hungary, at para. 534, CMS v. Argentina (award), at paras 406, 410–416, and 471, Maffezini v. Spain (award), at para. 96, Metalclad v. Mexico, at para. 130, Gruslin v. Malaysia, at paras 27.1 ff, and Olguín v. Paraguay (award), at para. 85. Examples relating to other rules can be found in Waste Management 2 (preliminary issues), at paras. 13 and 24, LGE v. Argentina (award), at para. 194, and Generation v. Ukraine, at para. 20.37.
285
CDC v. Seychelles, at para. 37, indicates that the tribunal followed such a distinction: ‘[k]eeping the object and purpose of the Convention as well as these underlying policy considerations in mind’.
286
Examples include Mondev v. US, at para. 91, and Plama v. Bulgaria (jurisdiction), at paras 161–164, which have been classified as object and purpose, and Plama v. Bulgaria (jurisdiction), at paras 207–209 and 219, and Maffezini v. Spain (jurisdiction), at paras 54–55, which have been classified as reasonableness.
287
For a borderline example that has been classified as reasonableness see Técnicas v. Mexico, at para. 69.
288
See ADC v. Hungary, at para. 435, Azinian v. Mexico, at paras 86 and 87, Maffezini v. Spain (jurisdiction), at paras 54–55, 56, and 62–63, Técnicas v. Mexico, at para. 69, Wena v. Egypt (award), at para. 129, Metalclad v. Mexico, at para. 67, Banro v. Congo, at paras 21 and 24–25, Mondev v. US, at paras 44 and 136, Feldman v. Mexico (award), at paras 103, 170, and 183, Generation v. Ukraine, at para. 14.5, SGS v. Pakistan, at paras 167–168 and 184, SGS v. Philippines, at paras 123, 132(c), 142, and 148, CMS v. Argentina (jurisdiction), at para. 125, CMS v. Argentina (award), at paras 317, 360, and 406, Loewen v. US (award), at para. 162, Enron v. Argentina (jurisdiction 1), at paras 50, 52, and 85, Tokelės v. Ukraine (jurisdiction), at para. 40, Joy Mining v. Egypt, at para. 54, Salini v. Jordan (jurisdiction), at para. 115, LESI v. Algeria 1, at para. 32(ii), Plama v. Bulgaria (jurisdiction), at paras 157, 207–209, and 219, Plama v. Bulgaria (provisional measure), at para. 40, Impregilo v. Pakistan, at para. 135, Sempra v. Argentina, at paras 53 and 102, Camuzzi v. Argentina 1, at paras 40 and 91, Vivendi v. Argentina (preliminary issues), at paras 15 and 21–22, Vivendi v. Argentina (jurisdiction), at paras 51 and 65, Interagua v. Argentina (preliminary issues), at paras 15 and 21–22, Interagua v. Argentina (jurisdiction), at paras 51 and 65, CDC v. Seychelles, at paras 36–37, Bayindir v. Pakistan, at paras 100 and 273, El Paso v. Argentina, at para. 82, Pan American v. Argentina, at paras 110 and 179, LGE v. Argentina (award), at para. 238, Biwater v. Tanzania (provisional measures 2), at paras 133, 136–137, and 140, and Telenor v. Hungary, at paras 93–94.
289
The numbers were: jurisdictional and procedural issues – 29 decisions and substantive issue – 11 decisions.
290
The numbers were: essential argument – 31 decisions and non-essential argument – 16 decisions.
291
See Vivendi v. Argentina (preliminary issues), at paras 15 and 20–21, and Interagua v. Argentina (preliminary issues), at paras 15 and 20–21.
292
See the Annex for a list of these decisions.
293
There were two tribunals that made three decisions and two tribunals that made two decisions (e.g., by distinguishing between their decisions on jurisdictional issues and on the merits). Annulment decisions are not counted since an ad hoc committee with new members is appointed to hear such cases: see Art. 52(3) of the Convention.
294
See Commission, supra note 1, at 137–141.
295
The conclusions presented here do not repeat those reached elsewhere in this article. They present general conclusions on the issues raised in the introduction of this article.
296
Including four decisions concerning the Notes of Interpretation issued by the FTC under the NAFTA, and excluding decisions concerning the Report of the Executive Directors.
297
See section 2 supra.
298
In the same direction see Commission, supra note 1, at 158.
299
See the text accompanying notes 102–104 supra concerning these distinctions.
300
For each group of decisions the percentage adds up to more than 100% due to the fact that decisions often contained more than one reference to the argument in question.
301
ICSID case law was used as an interpretive argument on multiple occasions in almost all decisions.
302
F. Orrego Vicuña, International Dispute Settlement in an Evolving Global Society. Constitutionalization, Accessibility, Privitization (2004), at 124.
303
A main example is ICSID tribunals’ use of supplementary means of interpretation: see Art. 32 VCLT. According to the ILC's Commentaries to Art. 28 (Art. 32 VCLT) in Yearbook of the International Law Commission (1966), ii, at 223, para. 19, the use of such arguments is an ‘exception to the rule that the ordinary meaning of the terms must prevail. … The Commission considered that the exception must be strictly limited, if it is not to weaken unduly the authority of the ordinary meaning of the terms. Sub-paragraph (6) is accordingly confined to cases where interpretation under article 27 gives a result which is manifestly absurd or unreasonable.’
304
See OECD, International Investment Law. A Changing Landscape (2005), ch. 1.
305
These include decisions that have been registered on ICSID's website, including cases where the following is noted: ‘[a]ward embodying the parties’ settlement agreement’. The following 15 decisions were not available as of 31 December 2006: WRB Enterprises and Grenada Private Power Limited v. Grenada, ICSID Case ARB/97/5 (award and settlement); Compagnie Française pour le Développement des Fibres Textiles v. Côte d’Ivoire, ICSID Case ARB/97/8 (award and settlement); Houston Industries Energy, Inc. and others v. Argentine Republic, ICSID Case ARB/98/1 (award); Patrick Mitchell v. Democratic Republic of the Congo, ICSID Case ARB/99/7 (award); Astaldi SpA & Columbus Latinoamericana de Construcciones SA v. República de Honduras, ICSID Case ARB/99/8 (award and settlement); Zhinvali Development Ltd v. Republic of Georgia, ICSID Case ARB/00/1 (award); Consortium RFCC v. Kingdom of Morocco, ICSID Case ARB/00/6 (annulment proceedings); Société d’Exploitation des Mines d’Or de Sadiola SA v. Republic of Mali, ICSID Case ARB/01/5 (award); AIG Capital Partners, Inc. and CJSC Tema Real Estate Company v. Republic of Kazakhstan, ICSID Case ARB/01/6 (award); Repsol YPF Ecuador SA v. Empresa Estatal Petroleos del Ecuador (Petroecuador), ICSID Case ARB/01/10 (award); F-W Oil Interests, Inc v. Republic of Trinidad & Tobago, ICSID Case ARB/01/14 (award); Fireman's Fund Insurance Company v. United Mexican States, ICSID Case ARB(AF)/02/1 (award); Total SA v. Argentine Republic, ICSID Case ARB/04/1 (decision on jurisdiction); BP America Production Company and others v. Argentine Republic, ICSID Case ARB/04/8 (decision on jurisdiction); and Helnan International Hotels A/S v. Arab Republic of Egypt, ICSID Case ARB/05/19 (decision on jurisdiction).
306
As regards Autopista v. Venezuela, the decision on jurisdiction is included while the decision on the merits is left out. As regards CDC v. Seychelles, the decision on the merits is left out, while the decision on annulment is included.