Dual Nationals in Investment Treaty Arbitration: An Emerging Field of Inconsistent Decisions

Introduction

Nationality is a crucial marker for protection in international investment law. Only investors that qualify as nationals of a contracting party can access investment treaties. Most investment treaties, however, are premised on broad provisions defining eligible nationals. With respect to individuals, these instruments typically define protected investors as physical persons who hold the nationality of the home state party in accordance with its domestic laws. Investment treaties, in other words, rarely regulate situations where investors seeking protection hold the nationality of both contracting parties. An exception to this broad nationality criterion can be found in the ICSID Convention. Article 25(2)(a) expressly excludes host state nationals from the Centre’s jurisdiction. But what if an investor with dual nationality resorts to other arbitral fora, such as ad hoc arbitration under the UNCITRAL Arbitration Rules?

The tribunal in Serafín García Armas and Karina García Gruber v Venezuela was the first to examine a claim by a dual national in a non-ICSID setting (García Armas I, UNCITRAL, 15 November 2014). The tribunal held that, since the applicable Venezuela-Spain BIT was silent on the standing of dual nationals, the investors qualified for treaty protection. As I predicted in a previous blogpost, this award has opened the floodgates for claims by dual nationals. At the time of writing, the number of claims has reached at least fifteen and tribunals have rendered awards in several of these cases. In doing so, arbitrators have grappled with a host of complex questions of treaty interpretation: are dual nationals subject to the customary international law rule of ‘dominant and effective’ nationality? Should an investor hold the nationality of the home state party at the time the investment was made? Should the definition of the term ‘investor’ be read as incorporating ICSID’s host state nationality restriction? Tribunals have answered these questions inconsistently and some awards have been challenged in different jurisdictions. This blogpost provides an overview of this case law, with a focus on the most recent decisions and the key interpretative issues raised therein.

1. The customary rule of ‘dominant and effective’ nationality

The point of departure for an overview on dual nationality cases must be the decision in García Armas I. The claimants, two Venezuelan-Spanish nationals, initiated proceedings under the Spain-Venezuela BIT, which provides for both ICSID and UNCITRAL arbitration. To overcome ICSID’s jurisdictional hurdle, the claimants resorted to UNCITRAL arbitration. The BIT and the UNCITRAL Rules impose no express restrictions on claims by dual nationals. Venezuela objected to jurisdiction on three main grounds, all of which have been invoked by states in subsequent cases.

As to the first ground, Venezuela argued that, under Article 31(3)(c) of VCLT, the tribunal should apply ‘any relevant rules of international law applicable in the relations between the parties’ to determine the standing of dual nationals under the BIT. For Venezuela, the relevant rule that filled the lacuna left in the treaty was the customary rule of ‘dominant and effective’ nationality. This rule, which has historically applied in the context of diplomatic protection claims, allows the espousal of claims only when it can be established that the dual national maintains more connections with the claimant state. In that way, the rule primarily aims to avoid the ‘internationalisation’ of purely domestic disputes (Aghahosseini and Draft Articles on Diplomatic Protection).

Venezuela averred that, since the claimants had stronger links with the respondent state (e.g., family, economic ties and habitual residence), they fell outside the BIT’s scope of protection. The tribunal rejected this objection, holding that ‘[t]he existence of a mechanism to resolve disputes directly between investors and the host state excludes diplomatic protection and its rules for being inconsistent with BITs’ (García Armas I, para 173). Accordingly, for the tribunal, insofar as the applicable investment treaty does not explicitly bars claims by dual nationals, an investor could rely on that treaty to sue their own state of nationality. This finding was recently upheld by the Paris Court of Appeal during annulment proceedings. Examining the award for a third time (after two remands from the French Cour de cassation), the Court of Appeal was unconvinced that general international law prevented the investors from accessing the BIT (Judgment, Paris Court of Appeal, 4 July 2023).

I have analysed elsewhere the repercussions and merits of the reasoning deployed by the tribunal in García Armas I (see here, here and here). My position remains that the legal architecture of investment treaties allows for a reassessment of the conventional wisdom on the relationship between investment treaty law and the law of diplomatic protection. The law of diplomatic protection complements international investment law, as opposed to both regimes being mutually exclusive. As such, derogation from the rule of ‘dominant and effective’ nationality, which applies as customary international law, should only be accepted to the extent that the treaty parties have clearly and expressly stated such an intention. This conclusion is supported by the principle of ‘systemic integration’ embedded in Article 31(3)(c) VCLT, according to which treaty parties ‘are to have taken to refer to general principles of international law for all questions which [the treaty] does not itself resolve in express terms’ (McLachlan, 373).

Respondent states have invoked the rule of ‘dominant and effective’ nationality in subsequent cases where the applicable treaty also contained a formalistic nationality definition. While some tribunals have replicated the position adopted in García Armas I, others have taken the opposite view. In Bahgat v Egypt, for instance, the tribunal rejected Egypt’s arguments that the investor, a dual Finish-Egyptian national, did not qualify for protection under the Egypt-Finland BIT by virtue of the customary rule. The tribunal held that ‘general international law principles concerning the consequences of dual nationality in respect of jurisdiction ratione personae do not trump the BIT’s explicit language’ (Bahgat v Egypt, UNCITRAL, 30 November 2017, para 232). Egypt challenged the award before the Hague Court of Appeal, which agreed with the tribunal that the BIT’s silence on the standing of dual nationals did not justify the application of customary international law (Judgment, Hague Court of Appeal, 20 October 2021). In Rawat v Mauritius, an arbitration initiated by a dual French-Mauritius national under the France-Mauritius BIT, the tribunal was similarly not willing ‘to add conditions to the BIT’ given that ‘[t]here is no express exclusion of dual nationals’ therein (Rawat v Mauritius, UNCITRAL, 6 April 2018, paras 170-172). For other decisions in support of García Armas I see Uzan v Turkey, SCC, 20 April 2016 and Khalil v Senegal, UNCITRAL, 24 October 2019.

On the other side of the spectrum, tribunals have held that where the underlying BIT does not clarify whether dual nationals can access the treaty, the rule of ‘dominant and effective’ nationality should be considered to ascertain personal jurisdiction. Interestingly, one of these cases includes a claim brought by other members of the García Armas family also under the Spain-Venezuela BIT. In Manuel García Armas and others v Venezuela (García Armas II), the tribunal agreed with Venezuela that, in conformity with Article 31(3)(c) VCLT, ‘rules and principles of general international law apply unless an express derogation is agreed by the treaty parties’ (García Armas II, UNCITRAL, 13 December 2019, para 704). The tribunal found that, since the claimants’ dominant nationality was that of Venezuela, based on the links they had with that state as opposed to Spain, they were not protected under the BIT.

More recently, in Valle Ruiz v Spain, the tribunal similarly remarked that investment treaties ‘must be interpreted against the background of broader international law rules’ (Valle Ruiz v Spain, UNCITRAL, 13 March 2023, para 448). In this connection, the tribunal added that ‘while BITs have advanced the protection of aliens in many respects […] this does not mean that they intended to do away with the basic underlying principles, save where States expressly said so’ (at para 457). The tribunal decided that the claimants holding dual Mexican-Spanish nationality were entitled to access the Spain-Mexico BIT because their dominant nationality was that of the claimant state. For other decisions disagreeing with García Armas I see Heemsen v Venezuela, PCA, 29 October 2019 and Trapote v Venezuela, PCA, 30 January 2022.

Dual nationality cases do not only raise questions regarding the application of the rule of ‘dominant and effective’ nationality. Respondent states have also argued that the investor must show that they held the nationality of the home state party at the time the investment was made.

2. The nationality of the investment

In García Armas I, the investors exclusively held Venezuelan nationality at the time they began their business operations in Venezuela that led to the establishment and development of the investment. Venezuela argued that the investor must be a home state national – and not a host state national – at the time the investment is made. Venezuela grounded its argument on Article I(2) of the BIT, which defines the term ‘investment’ as ‘any type of asset invested by nationals of one Contracting Party in the territory of the other Contracting Party’ (García Armas I, para 50). A majority rejected the objection, holding that ‘the relevant dates of nationality to invoke the protection of the BIT are (a) the date on which the alleged violation occurred (in this case, the government measures); and (b) the date on which the arbitral proceedings were initiated’ (ibid). In the setting aside proceedings, the Paris Court of Appeal sided with the findings of the majority, ruling that the BIT did not contain any condition that its nationality criterion must be met at the time when the investment was made (Judgment, Paris Court of Appeal, 27 June 2023).

The tribunals in Okuashvili v Georgia and Pugachev v Russia also dealt with a similar objection. The former case involved a claim under the Georgia-UK BIT brought by an investor who holds Georgian nationality by birth and who later naturalised as a British national. His investments in Georgia were made well before he acquired UK nationality. Georgia argued that, since the investor ‘was exclusively a Georgian national’ at the time of the investment, he ‘must be properly categorised as a domestic investor who is not entitled to invoke the protections afforded by the Treaty’ (Okuashvili v Georgia, SCC, 31 August 2022, para 269). Despite acknowledging the ‘unusual feature of the case’ (i.e., the investment was ‘purely domestic’), the tribunal rejected Georgia’s objection. It held that ‘the plain, unqualified formulation of the express provisions of the Treaty’ do not specifically refer to investments made by nationals of the home state party (at paras 275 and 283).  

In Pugachev v Russia, the tribunal took a different approach. The claimant, a Russian national from birth who acquired French nationality through naturalisation, commenced UNCTIRAL arbitration under the France-Russia BIT. The claimant himself confirmed that he made his investments before he obtained French nationality. The tribunal considered that ‘in order to be protected, an investment must be […] transnational (cross-border) from inception’ and thus ‘the Claimant [must have] French nationality at the time he made his alleged investments’ (Pugachev v Russia, UNCITRAL, 18 June 2020, paras 397 and 417). As a result, the majority declined jurisdiction. The investor tried unsuccessfully to annul the award before the Madrid Hight Court of Justice, which held that an award can only be set aside on jurisdictional grounds where an arbitration agreement does not exist or is invalid (Judgment, Madrid High Court of Justice, 10 November 2021).

A final argument typically raised by respondent states concerns the relationship between the nationality criterion in investment treaties and Article 25(2)(a) of the ICSID Convention.

3. The ‘importation’ of ICSID’s nationality restriction

The number of fora that investment treaties offer investors to choose from has increased over time, ICSID and UNICTRAL arbitration being by far the most frequently proposed fora (OECD Working Papers on International Investment 2012/02). As previously mentioned, Article 25(2)(a) of the ICSID Convention contains a negative nationality requirement. This provision provides that the term:

‘National of another Contracting State […] does not include any person who [on the date on which the parties consented to submit such dispute to arbitration as well as on the date on which the request was registered] also had the nationality of the Contracting State party to the dispute’.

To bypass this jurisdictional obstacle, investors have resorted to UNCITRAL arbitration or other mechanisms.

Perhaps to minimise opportunism by investors, respondent states have argued that the inclusion of ICSID as a mechanism to resolve investment disputes should be read as excluding dual nationals from the protection of the treaty, even if they resorted to another arbitral forum. The tribunal in Garcia Armas I was the first to examine this argument. Article XI(2) of the applicable Spain-Venezuela BIT gives the investor the choice to submit the dispute to either national courts or ICSID. Articles XI(3) allows the investor to access UNCITRAL Arbitration Rules if the first two mechanisms are not available. Venezuela asserted that the explicit reference to the ICSID Convention in Article XI(2) confirms that the intention of the Contracting Parties when concluding the BIT was to avoid being sued by their own nationals before an international tribunal; especially considering that ICSID is the first mechanism offered by the BIT. The tribunal was unconvinced. It held that ‘the provisions of the ICSID Convention are only applicable to arbitrations submitted under the Convention’ (García Armas I, para 194).

The tribunal in García Armas II, also constituted under the Spain-Venezuela BIT, ruled differently. It reasoned that ‘through the explicit reference to ICSID arbitration’ in Article XI(2), the Contracting Parties have tacitly excluded Spanish-Venezuelan nationals from the BIT’s ambit of protection (García Armas II, para 721). According to the tribunal, this conclusion is in line with the rules of the VCLT, which requires that the definition of ‘investor’ be interpreted in its ‘context’, that is, by considering other provisions of the BIT. The tribunal added that ‘[t]he practical consequence of this conclusion is that the definition of ‘‘investor’’ in the BIT does not have different meanings depending on the forum the investor relies upon’ (at para 722).

The tribunal in Rawat v Mauritius took a similar approach. Article 9 of the applicable France-Mauritius BIT contains an unusual dispute resolution clause. It specifies that investment contracts between an investor and the host state must include a dispute resolution clause providing for ICSID arbitration. The investor requested that the tribunal apply the most-favoured-nation (MFN) clause in the BIT to import the UNCITRAL arbitration mechanism contained in the Finland-Mauritius BIT. Mauritius argued that the investor could not ‘escape the effects of the ICSID Convention, since the ICSID Convention is the only treaty that is referred to in th[e] BIT’ (Rawat v Mauritius, para 118). The tribunal sided with the state, holding that the explicit reference to ICSID created ‘a strict and conventional alignment between the notion of “ressortissant” under the ICSID Convention and under the France-Mauritius BIT’ (at para 178). The Belgian Court de cassation upheld the tribunal’s interpretation in setting aside proceedings, dismissing the investor’s arguments that the arbitrators had erred in their approach to the interpretation of the BIT (Judgment, Belgian Cour de cassation, 6 April 2023).

In Valle Ruiz v Spain, the respondent state also invited the tribunal to ‘import’ the ICSID requirements on nationality within the applicable Spain-Mexico BIT. ICSID arbitration is one of the four options listed in the dispute resolution clause of treaty, together with UNCITRAL arbitration. In stark contrast to the reasoning adopted in García Armas II, the tribunal held that the ‘Claimants have opted for the legal framework [of] the UNCITRAL Arbitration Rules, which do not include any rule analogous to Article 25(2)(a) of the ICSID Convention’ (Valle Ruiz v Spain, para 437). Further, the tribunal did not considered that the BIT’s ‘notion of investor must always be the same whether a claimant opts for arbitration under the ICSID Convention or under any of the other available fora’ (at para 442).

Conclusion

The foregoing discussion allows us to draw three conclusions. The first is that claims by dual nationals is an emerging field within international investment law, a field plagued with inconsistent decisions. The case law shows a divergence in approaches to the application of the customary rule of ‘dominant and effective’ nationality, the relevant dates to ascertain the nationality of the investor and the role of Article 25(2)(a) of the ICSID Convention. This case law no doubt adds a further layer of uncertainty in international investment law, making it difficult for states and investors to ascertain the personal scope of investment treaties.

Second, case law on dual nationality illustrates that, while corporations have long been strategists in matters of nationality, individuals are increasingly adopting this role, seizing opportunities that states have inadvertently created for them. Cases like García Armas I and Okuashvili v Georgia show how investors now enjoy the benefit of having more than one passport that can be invoked to make and operate the investment in the host state and when it becomes convenient, to seek compensation under an investment treaty. Stated differently, the nationality of natural persons has also become a manipulable category, something that can be instrumentalized.

Lastly, it is probable that claims by dual nationals will continue to increase, and more awards will soon be rendered in pending cases. States that find these claims objectionable should regulate the issue in future and present treaties. Newly concluded treaties expressly incorporate the rule of ‘dominant and effective’ nationality or the stricter requirement found in the ICSID Convention as part of the definition of investor (for an analysis on reform options see here). States play a fundamental role in curbing the phenomenon and we can be cautiously optimistic about their ability to do so.

The author acted as counsel for Venezuela during the jurisdictional phase of the UNCITRAL arbitration in García Armas I.