6 - The Interpretation of ‘Direction or Control’ in Investor–State Arbitration

Summary

This chapter explores how the interpretation of customary international law (CIL) can be shaped by the underlying premises and political values of a system. The argument it develops focuses on how investor–state arbitration has interpreted the CIL law rule establishing that the actions of state-owned enterprises will be attributed to the controlling state, as expressed in Article 8 of the Articles on the Responsibility of States for Internationally Wrongful Acts.

1 Introduction

In this chapter I explore how the interpretation of customary international law can be shaped by the underlying premises and political values of a system. My argument focuses on how investor–state arbitration has interpreted the customary international law rule establishing that the actions of state-owned enterprises will be attributed to the controlling state, as expressed in Article 8 of the Articles on the Responsibility of States for Internationally Wrongful Acts (ARSIWA).¹

International lawyers use the term ‘state-owned enterprise’ to refer to corporate entities that are not part of a state but are still owned or controlled by a state. State-owned enterprises pose challenges for international law because, although they are formally independent and autonomous entities, their corporate and operational ties with states can make it hard to distinguish their activities and goals from those of the states that own or control them. This sometimes nebulous relationship between states and state-owned enterprises has generated international concern that states might use state-owned enterprises to cover reprehensible and even illegal activities.²

Without a specific international legal regime for state-owned enterprises, discussions on state involvement in the international activities of these entities should be grounded on general international law, including the customary rules on attribution that are expressed in the ARSIWA. Investment arbitration tribunals have generally agreed on the applicability of customary international law on state responsibility to cases that deal with attribution to states of actions that violate treaty protections.³ However, they have approached these cases from several different perspectives and standards (Section 2). I argue that such variation arises from the specific characteristics of investment law – namely, a cautious approach towards state-owned enterprises, coupled with an individualized application of the ‘effective control’ test. These two elements have pushed tribunals towards interpreting Article 8 of the ARSIWA in an individualized way that calibrates general international law standards in the light of common investment law practices, the rules specifically agreed by the states involved, and the specific needs and circumstances of the parties (Section 3).

2 The Interpretation of Control in International Law and Investment Arbitration

The ARSIWA were developed by the International Law Commission (ILC) as a way of codifying and developing the international law on state responsibility. Although the ARSIWA themselves do not constitute a binding instrument, many of their provisions, including Article 8, are considered to reflect customary international law.⁴ The ARSIWA themselves are rather a subsidiary instrument⁵ that describes these provisions in a way that is considered authoritative. It is worth mentioning here that despite the subsidiary nature of the ARSIWA as a source of international law, the recognized authority of the rules expressed in the ARSIWA has led international lawyers (including myself in this chapter) to broadly refer to these rules as binding.

According to Article 8 of the ARSIWA, the conduct of a non-state actor, including a state-owned enterprise, is attributable to a state if such entity ‘is in fact acting on the instructions of, or under the direction or control of, that State in carrying out the conduct’.⁶ As I describe below, the interpretation of the term ‘control’ has been a subject of discussion in international law. The ILC, in its commentary to the ARSIWA, emphasized that there were different ways of interpreting the term ‘control’.⁷ It is therefore up for those applying Article 8 to determine how to interpret and apply it in light of the circumstances of each case. As noted by the ILC:

It is clear then that a State may, either by specific directions or by exercising control over a group, in effect assume responsibility for their conduct. Each case will depend on its own facts, in particular those concerning the relationship between the instructions given or the direction or control exercised and the specific conduct complained of.⁸

Although the ‘effective control’ test of the International Court of Justice’s (ICJ) is referred to as the standard test for attributing actions of state-controlled entities to states, it has been shaped to different circumstances. Although there are many possible approaches, tribunals have been generally consistent and applied similar tests when faced with similar sets of circumstances (Section 2.1). This effort for consistency contrasts with broader variations in approaches seen in the context of investment tribunals ruling on the attribution to states of the actions of state-owned enterprises (Section 3.2).

2.1 ‘Control’ in International Law

The best-known test for determining whether a state has enough control to be held responsible over the actions of other entities is the ‘effective control’ test set by the ICJ. In all cases where the ICJ has discussed the issue of attribution to states of actions committed by other entities, it has found that states could only be held responsible for the actions of actors that they effectively controlled. In the case of Military and Paramilitary Activities in and against Nicaragua (Nicaragua v. USA), the ICJ established what it considered to be the threshold of control for attribution of conduct to states based on the rule expressed in Article 8. Nicaragua had turned to the Court claiming that the United States was responsible for supporting armed actions by a non-state armed group known as the Contras in Nicaraguan territory. The Court concluded that the United States should be held responsible for supporting the Contras. However, it also found that it could not hold the United States responsible for illegal actions committed by the Contras in their armed operations because, although the United States played a relevant role in the Contras’ operations, it did not have effective control over them.⁹ The ICJ later reiterated its understanding that attribution under Article 8 of the ARSIWA would require evidence that a state has effective control over the actions of a non-state actor in the Armed Activities in the Congo case¹⁰ and in the Bosnia v. Serbia case.¹¹

Other tribunals read the ICJ’s interpretation of control as the standard customary law test for attribution of conduct under Article 8 of the ARSIWA.¹² However, the solution found by the ICJ was not the only possible one. For example, those operating in specialized fields of international law have found that it was sometimes necessary to adopt a lower threshold of control in light of the legal and practical circumstances of the cases they were handling.

This flexibility is illustrated by the International Criminal Tribunal for the Former Yugoslavia (ICTY), which attributed the conduct of armed forces to states by making use of what became known as the ‘overall control’ test. In the Tadić case, the ICTY acknowledged the ‘effective control’ test developed by the ICJ, but found that its high threshold was incompatible with the strict hierarchy under which military operations take form. According to the ICTY: ‘The requirement of international law for the attribution to States of acts performed by private individuals is that the State exercises control over the individuals. The degree of control may, however, vary according to the factual circumstances of each case.’¹³

Following this logic, the ICTY concluded that the ‘effective control’ test set by the ICJ was suited for situations where it was necessary to determine whether the actions of a private individual (or group of individuals) are attributable to a state. In these cases, the ICTY stressed, there must be evidence that such an individual received from a state specific instructions to act as its de facto agent.¹⁴ However, such detailed examination would not be necessary when the state already has authority over an individual or group that operates under an organized hierarchy. Organized and hierarchical structures, such as military units, operate under more predictable, preexisting logics and strategies and therefore do not require a threshold as high as the one expressed by the ICJ in Nicaragua v. USA. The ICTY clarified that:

Normally a member of the group does not act on his own but conforms to the standards prevailing in the group and is subject to the authority of the head of the group. Consequently, for the attribution to a State of acts of these groups it is sufficient to require that the group as a whole be under the overall control of the State.¹⁵

The ICJ responded specifically to the Tadić test in its Bosnia v. Serbia judgment, suggesting that the approach taken by the ICTY in this case was ‘unpersuasive’ and dismissing the possibility of applying it in the context of state responsibility.¹⁶ The ICTY, however, kept following its own test as set in Tadić,¹⁷ with defenders of this approach arguing that both tests can coexist because they are designed to deal with different circumstances.¹⁸

The peculiarities of state-owned enterprises may also justify a specific approach to establish that a state controls them. In its commentary to Article 8 of the ARSIWA, the ILC discussed the attribution of conduct of enterprises that are owned and controlled by the state as an example of a situation that could warrant a threshold different from that used in ICJ case law. According to the ILC, the conduct of a state-owned enterprise should be attributed to the state pursuant to Article 8 of the ARSIWA when there is ‘evidence … that the State was using its ownership interest in or control over a corporation specifically in order to achieve a particular result’.¹⁹

The ILC’s commentary relied on an example from the case law of the Iran–US Claims Tribunal (IUSCT) on the attribution of the conduct of state-owned enterprises. The IUSCT, a tribunal created as part of a set of measures to manage the political crisis between the United States and Iran arising from the 1979 hostage crisis and the subsequent freezing of Iranian assets by the United States, decided several cases that involved the attribution to the state of the actions of state-owned enterprises. These decisions were grounded on interpretation of Article VII(3) and (4) of the Claims Settlement Agreement, which provided for the attribution to Iran and the United States respectively of the acts of the entities they controlled.

In most cases, the IUSCT attributed the actions of state-owned enterprises to the state only if there was proof that the latter was using its controlling stake to direct the actions of the enterprise.²⁰ However, in cases such as Foremost Tehran v. Iran the tribunal’s decision on attribution rested on the corporate characteristics of an enterprise – namely, ‘the identity of its shareholders and the composition and behaviour of its board of directors, which must be examined together’,²¹ rather than on the specific interactions between the state and the corporations.²² The reason for such variation was the position taken by the minority of judges, as stated in Judge Khalilian’s opinion in the case of SEDCO v. Iran:

Until now, the majority has taken the position, in its past Awards, that the term ‘control’ in Article VII (paras. 2, 3, 4) involves a qualitative predominance and, in the words of Chamber Three, ‘control over management’. … Nonetheless, it is worth noting, as a digression, that this same Tribunal has elsewhere correctly given the meaning of ‘control’ as set forth in paragraph 3, since it has held that two conditions must be met, before there can be a finding of control.²³

The ILC, when examining the IUSCT case law, focused on the majority cases where the tribunal looked for evidence that the state had used its corporate power to push the state-owned enterprise towards adopting specific actions.²⁴ The ILC found these majority cases representative, not only of the IUSCT practice for dealing with state-owned enterprises but of the position of international law in general.²⁵

These examples show how there are multiple ways of interpreting and applying Article 8 of the ARSIWA. The term ‘control’, in the way that it is used in the ARSIWA, or even the ‘effective control’ terminology used by the ICJ, are polysemic terms – they do not encompass a single type of relationship between states and non-state entities. It is up to those applying Article 8 to examine whether each situation they face meets the conditions for attribution under that provision. However, this case-by-case approach does not mean that Article 8 should be interpreted indiscriminately or inconsistently. In fact, all bodies examined in this section have sought a consistent application of Article 8 when dealing with similar sets of facts. Consistency in these cases is achieved not by applying the exact same text, but rather through constant judicial dialogue. This logic can be seen in the intense dialogue between ICJ and the ICTY about whether there should be a single control test, and if so, which should be the correct one. Both concluded that there were different possible interpretations for the term ‘control’ and that the test that each of them had been applying so far was the correct one for the specific circumstances they were facing.²⁶ Similarly, in the IUSCT the judges that were in the minority in cases such as SEDCO engaged in internal judicial dialogue when dissenting from its own case law.²⁷

2.1.1 Control Tests in Investment Arbitration

Investment tribunals have followed a tendency similar to that of their counterparts in other fields of international law. However, there is a key difference between the bodies examined in Section 2.1 and investment arbitration tribunals. Unlike the former which are permanent institutions with a stable composition of judges, investment arbitration is a fragmented dispute settlement system. Although investment arbitration is developed on the basis of common principles and goals, cases are grounded on agreements that are negotiated and drafted individually, each with its own particularities. Neither these agreements nor the tribunals that settle the disputes arising from them have formal ties with each other. There is also no formal obligation to follow precedent, meaning that tribunals may rely on previous case law when appropriate, but are not forced to follow it.²⁸ As a result, the findings made by a tribunal in one case do not bind other tribunals, making it hard to make definitive affirmations about the field.

With that caveat made, investment tribunals have generally acknowledged the ICJ’s ‘effective control’ test as the standard approach for holding states responsible for the actions of non-state actors. In the case of Marfin v. Cyprus, for example, the tribunal stressed that investment arbitration jurisprudence had consistently upheld the ICJ test and that there was no reason to depart from it.²⁹ As a consequence, the tribunal found that Cyprus could not be held responsible for treaty violations allegedly committed by a private bank, because although Cyprus had sometimes interfered in the bank’s administration and guided some of its decisions, there was no evidence that Cyprus had effective control over the bank’s alleged violations.³⁰

Arbitral tribunals’ attitudes towards the ICJ test vary, however, when the cases concern state responsibility for the actions of a state-owned enterprise. While many cases have followed the ICJ’s threshold of effective control, others have adopted much lower standards. Cases can be divided into three groups:

1. 1. those applying the ‘effective control’ standard set by the ICJ;

2. 2. those that consider the ‘effective control’ standard too high for cases involving state-owned enterprises and instead determine that states that cleared or authorized an action by a state-owned enterprise are responsible for those actions; and

3. 3. those that presume that states are responsible for the actions of any enterprise in which they hold more than 50 percent of the shares.

Cases applying the ICJ’s ‘effective control’ standard include Jan de Nul v. Egypt, where the tribunal examined whether actions of an Egyptian agency with distinct legal personality were attributable to the state of Egypt. The tribunal focused strictly on international rules on attribution and followed general standards for the interpretation of ARSIWA Articles 4, 5, and 8,³¹ making specific reference to the ICJ’s effective control test.³² Such an approach was also followed by the tribunals in Hamester v. Ghana, White Industries v. India, and Tulip Real Estate v. Turkey.³³

In several other cases the tribunals adopted much lower thresholds of control. In the Bayindir v. Pakistan case, the tribunal had to determine whether Pakistan should be held responsible for clearing actions of the National Highway Authority (NHA) that violated investment protections owed to the claimant. The tribunal was aware that state clearance met neither of the thresholds set in the ICJ’s effective control test and the ICTY’s overall control test, but chose nevertheless to attribute the NHA’s conduct to Pakistan under Article 8 of the ARSIWA.³⁴ The tribunal summarized its logic in the following way:

[T]he Tribunal is aware that the levels of control required for a finding of attribution under Article 8 in other factual contexts, such as foreign armed intervention or international criminal responsibility, may be different. It believes, however, that the approach developed in such areas of international law is not always adapted to the realities of international economic law and that they should not prevent a finding of attribution if the specific facts of an investment dispute so warrant.³⁵

The tribunal unfortunately did not go further into clarifying what were the ‘specific facts of the investment dispute’ that justified its departure from tests that were already established in international case law. Nevertheless, such interpretation was later followed by the tribunal in Deutsche Bank v. Sri Lanka. In this case, the tribunal found that the actions of a state-owned entity should be attributed to Sri Lanka on the basis of evidence that the entity was following a governmental directive when it started the program that led to the dispute.³⁶

Most tribunals are not, however, so explicit on the reasons that led them to depart from the general ‘effective control’ test. Take, for example, the case of Maffezini v. Spain. The question that was raised before the tribunal concerned the attribution to the state of the actions of SODIGA, a state-owned company engaged in a joint venture with the claimant. In its analysis, the tribunal stressed that the applicable law to determine attribution was that of state responsibility, in particular the provisions contained in the (then) draft ARSIWA. However, its analysis followed very different lines from those in the ARSIWA or those used by other international law courts. The tribunal ruled that the actions of state-owned entities are presumed to be attributable to the state if that entity (1) was owned by the state, (2) was controlled by the state, or (3) performed state functions.³⁷ The tribunal then focused on what it described as a ‘structural test’, through which the tribunal allowed itself to presume that any entity in which the government owned more than half of the shares would be operating under its control. Based on this test, the tribunal concluded that the Spanish state was responsible for SODIGA’s actions because it owned ‘no less than’ 51 percent of its shares.³⁸

The tribunals in RFCC v. Morocco and Salini v. Morocco approached ‘control’ in Article 8 from a similar perspective. Both cases arose from contracts between foreign investors and the Société Nationale des Autoroutes du Maroc (ADM), a state-owned enterprise responsible for many of Morocco’s highways. In RFCC v. Morocco, the tribunal concluded that Morocco had control over ADM after examining the information on its corporate ownership and its administration contained in its bylaws.³⁹ In Salini v. Morocco, the tribunal went further with its suggestion of attributing to states the actions of all state-owned enterprises, which the tribunal defined as ‘any commercial company dominated or predominantly controlled by the State or by State institutions, whether it has a legal personality or not’.⁴⁰ Years later, this approach was reiterated in the case of Helnan v. Egypt, where the tribunal attributed the actions of a corporation to Egypt based solely on the fact that it was owned by Egypt.⁴¹

In some cases, tribunals have also combined elements from different interpretations, tailoring their analysis of control to the circumstances of the case. In the case of Karkey Karadeniz v. Pakistan, the tribunal applied a test that included both corporate and effective control when applying Article 8 of the ARSIWA. The case concerned contracts signed by the claimant with a Pakistani corporation as part of a broader governmental plan to tackle the Pakistani energy crisis. The tribunal started its examination by pointing out that the company itself had acknowledged that it was ‘owned and controlled by the government of Pakistan’, suggesting that it would apply the criterion of corporate control to the ‘control’ threshold in Article 8 of the ARSIWA.⁴² However, the tribunal then stressed that the evidence on record indicated that it was the government of Pakistan that had negotiated the contract and selected the corporation that was going to enter into it as a buyer.⁴³ Ultimately the tribunal’s conclusion was that the company’s ‘acts related to the conclusion and execution of the Contract were directed, instructed or controlled by Pakistan, and are accordingly attributable to Pakistan, irrespective of the disclaimer contained in the RFP [Request for Proposal]’.⁴⁴

The interpretation of the term ‘control’ in Article 8 of the ARSIWA will therefore vary depending on the specific circumstances of a case, the broader legal framework, and the rules which the tribunal is applying to the case. In this sense, investment arbitration tribunals dealing with questions of attribution of conduct of state-owned enterprises to states have followed similar patterns to the ones seen in other international law contexts. The use of the ‘effective control’ test by the ICJ in preference to a lower standard is a way to prevent it from overstepping its functions and ensuring that it does not meddle in states’ domestic affairs. On the other hand, the ICTY employed a lower threshold justified by the fact that armed troops operate under the oversight of the state and follow a rigid hierarchy.

However, as shown by the IUSCT, the particularities of state-owned enterprises may warrant for a specific interpretation of the term ‘control’. The control test applied by the majority of the IUSCT determined, first, whether the state was able to use its shares or other corporate or institutional mechanisms to control the actions of the enterprise. Secondly, the tribunal would look at whether the state had indeed used those powers to control the enterprise and have it perform the specific actions discussed in the case at hand. In investment law, tribunals applying Article 8 of the ARSIWA have employed three different tests, holding states responsible (1) only for the actions of state-owned enterprises they effectively controlled and specifically directed, adopting a test similar to that of the IUSCT; (ii) for any action they authorized; or (iii) for any action of the state-owned enterprises over which they held corporate control.

3 Investment Law Shaping the Concept of Control

In the absence of specific provisions on state responsibility for the actions of state-owned enterprises, the applicable rules are those contained in the ARSIWA, which is deemed to mostly reflect customary international law. However, the interpretation of such rules is shaped by the specific factual and legal context of each case. The factual context relates to the particularities of the relationship between the state and the entity, including the powers that the state holds as a shareholder and the specific interactions between the state and the enterprise in the circumstances of the case. The legal context comprises the characteristics of the field of international law and the instruments applicable to the relationship between the parties, including the legal principles that guide them, and how the stakeholders see and apply the rules contained in these instruments.⁴⁵

In this section, I discuss two particularities of international investment law that may help explain why arbitrators adopt lower thresholds of control when applying Article 8 of the ARSIWA to the actions of state-owned enterprises. International investment law is a field that is largely inspired by the goal of depoliticizing investment relations, which translates into discouraging states from participating directly in investments.⁴⁶ This underlying goal may push arbitrators to approach state-led investment initiatives cautiously (Section 3.1). On top of that, arbitrators have in other situations equated ‘effective control’ with corporate control, thus suggesting that investment tribunals interpret Article 8 of the ARSIWA much more flexibly than courts that operate in other fields of international law (Section 3.2). Ultimately, a tribunal will calibrate the threshold of the control by looking at the characteristics of the field, general international law standards, and the particularities of each case (Section 3.3).

3.1 Caution towards State-Led Initiatives in the Development of International Investment Law and Investment Arbitration

Investment arbitration is often guided by a general cautiousness towards state-led initiatives. International investment law was developed around the principle of investment neutrality, which assumes that state involvement in investment relations may generate political risks that harm foreign investment.⁴⁷ The term ‘political risk’ is not defined legally but is used in investment and business circles to refer to risks associated with political fluctuations such as changes of power and policy. Many of the protections contained in investment agreements are aimed at reducing political risks that foreign investors may face in the host state – for example, most-favored-nation and fair-and-equitable-treatment clauses. Most-favored-nation clauses prevent the host states’ relationships with other states from interfering with investment relations by allowing foreign investors to claim most-favored-nation treatment.⁴⁸ Fair-and-equitable-treatment provisions require foreign investors to be treated ‘fairly and equitably’⁴⁹ – terms that have been interpreted as a blanket protection for foreign investors against host state actions that seem to be politically motivated or ‘irrational’.⁵⁰

This logic of limiting state power also applies to investment arbitration. Investors often feel that the host states’ domestic courts – the default forum for them to settle these disputes – may be biased against foreign investors or lack the expertise to deal with investment treaty claims. Even before clauses providing for arbitration as a means to settle disputes started to be introduced in investment treaties, states and investors were already reaching out to organizations such as the World Bank to request informal assistance to settle their disputes.⁵¹ At the same time, the international dispute settlement options that existed then were limited and politically sensitive. To start a case against the host state, investors had to convince their home states to exercise diplomatic protection. This forced investors to spend significant time and political resources without any guarantee that the case would be picked up by their home states or lead to a desirable outcome.⁵² Investment arbitration as a mechanism of dispute settlement that investors can use directly, without having to appeal to their state of origin, was thus considered a more politically neutral option than both domestic dispute settlement and diplomatic protection.⁵³

In addition to depoliticizing the access of investors to international dispute settlement mechanisms, there is also the goal of depoliticizing the resolution of disputes themselves. Investment law often deals with subjects that are highly sensitive to states,⁵⁴ and one of the central goals of investment arbitration is to reduce state influence over litigation outcomes.⁵⁵ This includes, for example, prohibiting investors’ home states from directly engaging in investment disputes, as provided in Article 27 of the ICSID Convention.⁵⁶

The ICSID Convention also prevents states from making use of investment arbitration as investors themselves. The idea of having investor-states as parties to investment arbitration was raised and rejected during the negotiations of the ICSID Convention. Many of the states involved in negotiations feared that allowing states to access investment arbitration as investors would facilitate political capture of the system, which would defeat the purpose of creating a depoliticized system.⁵⁷ Article 25(1) of the ICSID Convention, which establishes ICSID jurisdiction, therefore excludes the possibility of interstate disputes:

The jurisdiction of the Centre shall extend to any legal dispute arising directly out of an investment, between a Contracting State (or any constituent subdivision or agency of a Contracting State designated to the Centre by that State) and a national of another Contracting State …

Similarly, the early development of investment arbitration seems to have embedded in its discussions the idea that, although formally separate from states, state-owned enterprises would probably not be eligible claimants in investment arbitration because they operated under state control. Early ICSID drafts determined that, for jurisdictional purposes, state-owned enterprises should be equated to states.⁵⁸ The final draft of the ICSID Convention also allowed states to consent to foreign investors filing claims directly against their state-owned enterprises, therefore presuming that those enterprises were part of host states themselves.⁵⁹ Even today, there are investment treaties with provisions limiting the protections afforded to state-owned investors and, as a consequence, their capacity to resort to investment arbitration.⁶⁰

However, the creators of this dispute settlement system were aware that these preconceptions were subjective and could not reasonably justify a blanket prohibition against state-owned claimants making use of investment arbitration. When faced with the question of whether state-owned enterprises should be excluded from filing investment arbitration cases, Aaron Broches, former general counsel of the World Bank and one of the main sponsors of the ICSID Convention, acknowledged that they could play distinct roles in an economy, but seemed to believe that truly independent and market-oriented state-owned enterprises were an exception. In Broches’s words, ‘[t]here would appear to be no reason why a state-owned enterprise, if it elected to assimilate itself to a private enterprise rather than a government agency’⁶¹ should be treated differently from a privately owned corporation.

Most of the discussions on state-owned enterprises in investment law were, however, informal and were not translated into explicit treaty provisions. Broches’s writings on the ICSID negotiations indicate that the drafters of these treaties may have thought that it was not necessary to add such a level of detail on state-owned enterprises, since they shared a ‘common understanding’ on the subject. This understanding was that although state-owned enterprises could not be automatically equated to states, most of them presented a risk for a free market economy because they operated under strict state supervision. This notion, however, covered only the political behaviour and characteristics of state-owned enterprises.

It is worth noting that this tendency to overgeneralize treaty terms on state-owned enterprises also prevails in modern investment treaties. With the exception of the EU–China Comprehensive Agreement on Investment, which specifically lists the entities covered by its provisions,⁶² the few treaties that contain provisions on how to distinguish state-owned from other corporations or from states still do so in very broad terms, such as ‘when it exercises any regulatory, administrative or other governmental authority’.⁶³

While the legal instruments shaping international investment law do not provide much guidance on how to deal with legal questions involving state-owned enterprises, they show that stakeholders engaged in developing international investment law and investment arbitration were cautious towards recognizing rights of state-owned enterprises as investors. As stressed by the tribunal in Maffezini v. Spain, the overall framework of international investment law presumes that state-owned enterprises will, as a general rule, operate alongside states and even as part of their structure.⁶⁴ In the extreme, relying on such notions may lead to decisions such as the one taken by the tribunal in Bayindir v. Pakistan, which concluded that it could attribute to a state the actions of any enterprise that it owns, regardless of there being specific evidence showing that the state indeed exercised control over that entity.⁶⁵

3.2 Using ‘Effective Control’ as a Flexible Standard

A second element shaping how investment arbitration approaches the attribution to states of state-owned enterprise actions is the flexibility with which investment tribunals have approached the term ‘effective control’. Tribunals have relied on the terminology of effective control in various contexts and referring to different thresholds of control, including corporate control. In the context of investment arbitration, the term ‘effective control’ has been used to indicate that a tribunal must look at the facts of each case, rather than relying on general formulas that presume control.

The legal issue with regard to which the tribunals show the greatest flexibility in the interpretation of the standards of effective control is the determination of the nationality of foreign investors. As stated in Article 25(1) of the ICSID Convention, modern investment arbitration is supposed to be between a state and a national of another state. Therefore, a corporate foreign investor will be allowed to file an investment claim on its own only if it proves that it is a national of a state party to a relevant treaty. However, it is not easy to determine the nationality of an investor, since many host states require investments to be made through local subsidiaries that would not normally qualify as foreign investors. In general international law, the nationality of a corporate investor is determined on the basis of its place of incorporation,⁶⁶ which makes it impossible for foreign shareholders to file claims seeking relief for damages suffered through a local subsidiary. Aware that this would hamper the access of transnational companies to investment arbitration,⁶⁷ the drafters of the ICSID Convention included in Article 25(2)(b) the possibility for states to agree to treating local subsidiaries controlled by foreign corporations as foreign nationals:

(2) ‘National of another Contracting State’ means: …

(b) any juridical person which had the nationality of a Contracting State other than the State party to the dispute on the date on which the parties consented to submit such dispute to conciliation or arbitration and any juridical person which had the nationality of the Contracting State party to the dispute on that date and which, because of foreign control, the parties have agreed should be treated as a national of another Contracting State for the purposes of this Convention.⁶⁸

Many investment treaties thus include in their dispute settlement chapters provisions that also treat as foreign investors subsidiaries that are controlled by a national of a foreign state.⁶⁹ However, such provisions require proof that the claimant parent company has corporate control over the local subsidiary.⁷⁰ Although the most obvious form of corporate control would be to hold more than half of the subsidiary’s shares, there are many ways of controlling a corporation – for example, through special voting rights, golden shares, or joint ownership, to name but a few. Perhaps to avoid being blindsided by looking only at majority shareholding, tribunals have treated as nationals of a state claimants that are under the ‘effective’ corporate control of an entity established in that state.⁷¹

Despite the use of terminology identical that used by the ICJ in the Nicaragua v. United States case, the test used by arbitral tribunals to determine the nationality of investors could not be more different from the one employed by the ICJ. Whereas the ICJ test focused on whether a state had de facto control over another entity’s specific actions, the test used by arbitral tribunals to determine the nationality of investors involves an inquiry into the company’s corporate control mechanisms. In the case of Rompetrol v. Romania, the tribunal examined the successive changes in the ownership of Rompetrol since 1993 to determine who had ‘effective control of the corporate entities in question’.⁷² The Vivendi I annulment committee adopted a similar approach to examine whether a transfer of control over one of the claimants could affect its nationality, stressing that a company could be considered a foreign national only if it was ‘effectively controlled, directly or indirectly, by nationals of one of the contracting parties’.⁷³

Neither of the tribunals in the cases of Rompetrol or Vivendi I referred or drew connections to Nicaragua v. United States or to Article 8 of the ARSIWA. This raises the question of whether these two investment tribunals had the intention to apply and interpret the ICJ’s effective control test through the lens of investment law when adopting such terminology. Nevertheless, the use of the term ‘effective control’ has led other tribunals to treat such case law as part of a common analytical framework. In Pan American Energy and BP Argentina v. Argentina, the claimants and the tribunal referred to case law on the determination of nationality and on attribution of conduct to justify the use of corporate control as a criterion to determine nationality.⁷⁴ The resulting award thus combined the case law on ‘effective control’, thereby creating another new test for attribution of conduct under these criteria.⁷⁵

Investment tribunals therefore approach the interpretation of the term ‘control’ under Article 8 of the ARSIWA from a rather unique perspective. In the absence of a distinction between ‘corporate nationality control’ and ‘control over non-state actor’s conduct’, the concept of ‘effective control’ ceases to be a specific threshold for control in the context of investment arbitration. Rather than using it as a reference to a threshold of control, investment tribunals seem to treat the ‘effective control’ standard as translating an obligation to examine each case individually and to evaluate different legal approaches to reach a solution that is appropriate to the specific circumstances of the case.

3.3 Judicial Reasoning as a Compass for Determining a Threshold for Control

The flexibility with which tribunals have approached the ‘effective control test’ is not exceptional in investment arbitration. As argued elsewhere, investment arbitration tribunals approach situations on a case-by-case basis, tailoring their legal solutions to the characteristics of the agreements at hand and the specific needs of the parties to the dispute.⁷⁶ Faced with the question of attribution of state-owned enterprise conduct to states based on Article 8 of the ARSIWA, investment tribunals will therefore determine what threshold of control to apply in each case by calibrating⁷⁷ standards used in general international law and the specific characteristics of the international investment law applicable to the case at hand.

An examination of the way in which investment arbitration tribunals applied Article 8 of the ARSIWA raises the question not only of the threshold of control that they apply but even more so the elements that they take into consideration to determine the threshold of control applicable to each case. A tribunal that rules on whether to attribute the actions of a state-owned enterprise to a state pursuant to Article 8 of the ARSIWA will have to consider whether the (very high) threshold of control set by the ICJ is appropriate to the circumstances of the case at hand,⁷⁸ including the rules agreed upon by the treaty parties; the common practices in the field;⁷⁹ the treaty parties’ laws on state-owned enterprises;⁸⁰ and evidence of the involvement of the treaty party in the state-owned enterprise or in events related to the case.⁸¹

The highest threshold of control that a tribunal could adopt would be the one applied by the ICJ in the Nicaragua v. United States case, which holds states responsible only for actions that they have specifically controlled by, for example, ordering or directing the state-owned enterprise to act in such a way. The lowest threshold would be to attribute to the state the actions of any entity in respect of which it has participatory decision-making power, regardless of whether it is indeed involved in the decisions taken concerning that entity. Between these two extremes there is a range of other possible thresholds that may be adopted depending on the context of a case. A tribunal may, for example, hold states responsible for the actions of non-state entities when they have evidence that the state formally cleared or authorized those actions. All of these approaches are in principle allowed by international investment law and investment arbitration, as arbitral tribunals have the power to interpret and apply international law to solve the conflict they have at hand.

Tribunals’ powers to interpret and apply open-ended rules such as the one expressed in Article 8 of the ARSIWA is tied to an obligation to provide reasoning. Article 48(3) of the ICSID Convention requires tribunals to state the reasons upon which they have made their decisions. Article 34(3) of the UNCITRAL Arbitration Rules contains a similar obligation, which applies unless the parties have agreed otherwise. In ICSID proceedings, an arbitral award that lacks an account of the tribunal’s reasoning may even be annulled pursuant to Article 52(1)(e) of the ICSID Convention. The amount of detail that a tribunal must give to fulfill its obligation to provide its reasoning will vary from case to case. As noted by the annulment committee in Wena, there is no single compulsory format in which the tribunal must present its reasons, and it can even leave its reasons implicit, ‘provided that they can be reasonably inferred from the terms used in the decision’.⁸² For example, we can assume that a tribunal dealing with Article 8 of the ARSIWA that applies the exact same threshold for control as the ICJ did in Nicaragua did so because it saw no specific reason to recalibrate the general international law standard. However, if a tribunal issues a decision that seems to depart from previous understandings, it should consider explaining why it decided to do so.⁸³

Although tribunals enjoy considerable freedom when interpreting open-ended rules such as that expressed in Article 8, references to different approaches in a single award suggest that, in investment arbitration, the term ‘effective control’ refers more to a standard of evaluation than a specific threshold of control.⁸⁴ Take the tribunal in Karkey Karadeniz, for instance, which justified its decision on attribution on the basis of elements that refer to corporate control, de facto control, and specific instructions – namely (1) the state-owned had acknowledged that it was owned and controlled by the government of Pakistan; (2) the government of Pakistan had approved the contract related to the dispute; (3) the government of Pakistan was the guarantor of that contract; and (4) the buyer was not allowed to interfere in licenses granted in the contract, which was interpreted as a sign of bad faith.⁸⁵ The diverse interpretations, each unique to a particular award, demonstrate the elasticity the term ‘effective control’ has within this system.

In the absence of a specific threshold of control to be applied uniformly across cases, a correct application of Article 8 of the ARSIWA relies on judicial reasoning. Rather than simply looking at the case law for possible tests and picking one, investment arbitration tribunals must explain their reasons when ruling on what would be the correct threshold of control to be applied in the case before them. As tribunals are not limited in the reasoning that they need to present to justify their decision, they have the power and freedom to decide what elements they will consider relevant in each case and how they weigh each of those elements. However, those reasons must be clear and rational from the stance of international investment law.

4 Conclusion

Although the ICJ’s effective control test is commonly described as a standard solution for applying Article 8 of the ARSIWA, the ICJ has itself recognized that this test cannot be applied universally.⁸⁶ Being a polysemic term, ‘control’ in Article 8 can only be interpreted in the light of the circumstances of a particular case. As a consequence, tribunals applying Article 8 have each had to develop their own control tests according to the legal and factual particularities of the cases before them. The process of interpreting these tests required tribunals to dialogue with each other in order to determine what elements it was appropriate for them to rely on when interpreting control in the cases before them.

Investment tribunals have often interpreted Article 8 of the ARSIWA differently in each case. Rather than establishing a single threshold that would be applied in every case involving state-owned enterprises, arbitral tribunals gauge the ideal threshold of control on the basis of the characteristics and needs of each case. Each tribunal operates independently and therefore has the power to decide alone what factors it will take into consideration when it determines the level of control, such as (1) the importance of following general international law precedent; (2) the specific characteristics of the case at hand, including the parties’ needs and the particularities of the treaty that they agreed upon; and (3) the legal principles that guide international investment law and common practices in investment arbitration. In this chapter, I have focused on the last of these elements and identified two characteristics of international investment law that play a relevant role in tribunals’ interpretations of Article 8 of the ARSIWA in cases involving state-owned enterprises.

The first of these characteristics is the principle of investment neutrality that inspires international investment law and guides investment agreements. The principle of investment neutrality assumes that state participation or interference in investment relations may create political risks prejudicial to investment relations, and acts as a warning to stakeholders to approach any state-led initiative with caution. In practice, this caution guides the drafting of treaty provisions focusing on the protection of foreign investment and may lead tribunals to interpret provisions in ways that extend the scope of state liability for participating directly in foreign investment. The second characteristic is the fact that investment tribunals have used the term ‘effective control’ to refer to other levels of control besides the very high and stable threshold set by the ICJ in Nicaragua v. United States. This flexible approach towards effective control suggests that investment tribunals approach cases from a more individualized perspective and one that takes into consideration the principles that guide the field.

However, this individualized approach towards the application of Article 8 of the ARSIWA increases the burden laid upon tribunals to justify their decisions. Tribunals have the obligation to provide in their awards the reasons that led them to decide in a certain way. While there is no specific format for presenting these reasons, they must be clear to the reader and follow a rational line of thought. When calibrating the threshold of control that they are going to apply in each case, tribunals must therefore clarify what elements they considered relevant for such calibration, how they have weighted them, and why they concluded that the threshold applied is best suited for that specific case.