1. Claims by dual nationals against their State of nationality

We have already determined that most IIAs protect individual investors who possess the nationality of the home State party under its domestic legislation. Where the home State nationality is the only relevant criterion, a tribunal may have to decide on the recognition of this nationality in cases where the investor seeking protection is also a national of the host State. If this factual scenario arises in ICSID proceedings, the tribunal should decline jurisdiction.Footnote ³⁶ However, what if, for economic reasons or otherwise, the investor holding dual home–host State nationality decides to institute proceedings under a different dispute settlement mechanism, such UNCITRAL Arbitration?

Unlike the ICSID Convention, the UNCITRAL Arbitration Rules are silent on the standing of dual nationals. Thus, in the event of an UNCITRAL-based IIA claim, the tribunal will have to determine whether being a national of the home State is sufficient for a dual home–host State national to sue his or her own State of nationality. Put differently, the question is whether, in the absence of express language to the contrary in the applicable IIA, an investor holding the nationality of both Contracting Parties should be entitled to treaty protection. Recent investment tribunals have answered this question inconsistently, adding a further layer of uncertainty in international investment law.

The point of departure was the decision in Serafín García Armas and Karina García Gruber v Venezuela (García Armas I), considered as the first publicly known ISDS case in which a dual home–host State national has brought an investment treaty claim outside the ICSID Convention Regime.Footnote ³⁷ In this case, two Venezuelan–Spanish nationals brought a claim against Venezuela for alleged breaches of the Spain–Venezuela BIT. By virtue of the ISDS clause in the BIT, which provides for both ICSID and ad hoc UNCITRAL arbitration, the claimants chose to resort to UNCITRAL arbitration. This was perhaps because Venezuela's withdrawal from the ICSID Convention in 2012. The BIT contains the typical broad definition of ‘investor’, only requiring that investors hold the nationality of the home State (Spain in this case).Footnote ³⁸ Despite this, Venezuela objected to jurisdiction, arguing that the Spanish nationality of the claimants was a mere formality used conveniently with the intention of making their own State a respondent before an international tribunal. In support of its objection, Venezuela relied on the customary international law rule of ‘dominant and effective’ nationality.Footnote ³⁹ This rule is codified in the 2006 International Law Commission (ILC) Draft Articles on Diplomatic Protection, Draft Article 7 of which provides that:

According to this provision, a State may exercise diplomatic protection in respect of a dual home–host State national only if he or she does not have stronger connections (personal, economic, political, etc) with the respondent State. It was undisputed that the claimants’ dominant and effective nationality was that of Venezuela, given the long-standing relationship that they had developed with that country for several years.Footnote ⁴¹

The tribunal rejected the objection on the premise that IIAs are ‘special law’, meaning that its jurisdiction should exclusively be based on the nationality criteria of the BIT. In making this finding, the tribunal examined the legal nature of diplomatic protection and IIAs. It observed that the rules relating to diplomatic protection—such as the dominant and effective nationality test—were developed at a time when States were the only bearers of rights and obligations under international law. However, the tribunal continued, IIAs have departed from the traditional system of diplomatic protection by enabling investors to vindicate their own rights before international tribunals. According to the tribunal, by providing direct access to international arbitration, IIAs have created a self-contained regime that derogates from the customary law of diplomatic protection. In other words, the fact that the claimants were also Venezuelan nationals and had strong connections with that State mattered little to the tribunal.Footnote ⁴² As Spanish nationals, the tribunal concluded, the claimants qualify as investors.

This decision has set a new precedent in investment treaty practice, a precedent that has opened the floodgates for claims by dual home–host State nationals. Shortly after the tribunal rendered its jurisdictional award in García Armas I, an increasing number of investors brought treaty claims against their States of nationality outside the ICSID Regime.Footnote ⁴³ These investors include other dual Venezuelan–Spanish nationals belonging to the García Armas family who were also affected by Venezuela's actions.Footnote ⁴⁴

Tribunals have already rendered decisions in some of these cases. In at least three awards, tribunals have followed the literal approach to treaty interpretation adopted in García Armas I.Footnote ⁴⁵ For instance, in Rawat v Mauritius the tribunal had to determine ‘whether the term ‘‘ressortissant’’ [or national], as used throughout the France-Mauritius BIT includes or excludes dual nationals’.Footnote ⁴⁶ In response to an objection by Mauritius, the tribunal held that silence in the treaty on the standing of dual home–host State nationals ‘would also point to the outcome of including, rather than excluding, dual nationals as protected ‘ressortissants within the ambit of the BIT’.Footnote ⁴⁷

Other tribunals have taken a different view.Footnote ⁴⁸ Most recently, in Manuel García Armas and others v Venezuela (García Armas II) the tribunal held that the Spain–Venezuela BIT should not apply in isolation from the nationality rules of the customary international law of diplomatic protection.Footnote ⁴⁹ The tribunal reasoned that ‘el derecho internacional sobre arbitraje de inversiones [no ha] evolucionado hasta el punto de admitir los reclamos de doble nacionales sin restricciones’.Footnote ⁵⁰ For the tribunal, therefore, the fact that the BIT did not regulate dual home–host State nationality justified the application of the customary rule of dominant and effective nationality, a position that completely contradicts the award in García Armas I. After considering the claimants’ links with Spain and Venezuela, the tribunal found that the claimants were predominantly Venezuelan. The tribunal's analysis did not stop there.

As previously explained, the Spain–Venezuela BIT offers ICSID and UNCITRAL arbitration as fora to resolve investment disputes under the treaty. The BIT, however, prioritises ICSID over UNCITRAL, the latter being a fallback option. Interestingly, the tribunal reasoned that, ‘debido a esta estructura que impone la prioridad u obligatoriedad de los arbitrajes bajo el Sistema CIADI’, the exclusion of dual home–host State nationals in Article 25(2)(a) of the Convention pervaded the BIT.Footnote ⁵¹ According to the tribunal, this conclusion emanates from the interpretive rules of the Vienna Convention on the Law of Treaties, which requires that the tribunal interpret the definition of ‘investor’ in its ‘context’, that is, by reference to other provisions of the BIT. Hence, ‘[m]ediante la referencia al Sistema CIADI’,Footnote ⁵² the tribunal concluded, ‘las Partes Contratantes “han implícita, pero necesariamente, excluido a los dobles nacionales [españoles-venezolanos] del ámbito de aplicación del TBI”’.Footnote ⁵³ The tribunal added that ‘[l]a consecuencia práctica de dicha conclusión es que la definición de ‘inversor’ en el Tratado no tiene diferentes significados según el foro al que se recurra’.Footnote ⁵⁴ The tribunal made this determination, and declined jurisdiction, despite the fact that the arbitration was being conducted under the UNCITRAL Arbitration Rules and that Venezuela denounced the ICSID Convention in 2012.

Inconsistency now reigns over the question whether broad nationality requirements should be interpreted as extending protection to dual home–host State nationals. Time will tell if tribunals in the remaining cases follow the permissive approach adopted in García Armas I or if, on the contrary, they apply the limitations imposed by the tribunal in García Armas II. In making this assessment, arbitrators should consider that, in addition to increasing host States’ exposure to ISDS, claims by dual home–host State nationals have other repercussions.

For one, by allowing claims of this nature, arbitrators place dual home–host State nationals in a favourable position with respect to domestic investors lacking the nationality of the home State. Indeed, as opposed to dual home–host State nationals, who can ‘pick and choose’ among two nationalities to claim against either Contracting Party or even both, domestic investors with one nationality only have access to local remedies. This creates an incentive for domestic investors to ‘internationalise’ their claims through the acquisition of a second passport. An increasing number of States facilitate this practice through ‘Golden Visa programs’, which allow wealthy individuals to obtain residency and nationality through investment.Footnote ⁵⁵

The unwillingness of investment tribunals to impose limitations on dual home–host State nationals has also resulted in the initiation of parallel ISDS proceedings under the same BIT but pursuant to different dispute resolution mechanisms. On 23 August 2018, 43 investors initiated ICSID and UNCITRAL arbitrations against Spain under the Spain–Mexico BIT for the same cause of action. In the UNCITRAL request for arbitration, eight of these investors clarify that ‘they are dual nationals holding both Mexican and Spanish citizenship, and therefore are prevented from participating in the ICSID arbitration’.Footnote ⁵⁶ Presumably cognisant of the García Armas I award, these eight investors further state with confidence that ‘their concurrent Spanish nationality does not affect their rights under the treaty’.Footnote ⁵⁷

2. Claims by dual nationals from third States

Dual home–host state nationals are not alone in benefiting from broadly worded definitions of ‘investor’. An investor who is a national of the home State party and a non-State party can also strategically rely on one or both nationalities to gain access to IIAs. Respondent States have equally opposed claims by dual home–third State nationals. Here too, tribunals have embraced investor-friendly interpretations,Footnote ⁵⁸ even in the presence of potential restrictions in the relevant IIA.

The most recent example is the decision in Aven v Costa Rica. Mr David Aven, a dual United States–Italian national, instituted UNCITRAL proceedings against Costa Rica under DR-CAFTA.Footnote ⁵⁹ Article 10.28 of DR-CAFTA defines ‘investor’ as ‘a national […] of a Party [here the United States]; provided, however, that a natural person who is a dual national shall be deemed to be exclusively a national of the State of his or her dominant and effective nationality’. Notably, whilst Mr Aven invoked his United States nationality to claim under the treaty as an investor of the home State, he used his Italian nationality to establish and operate its investment in Costa Rica. Costa Rica objected to jurisdiction on two grounds. First, the State argued that, since Mr Aven maintained more links with Italy, his dominant nationality was Italian and therefore his United States nationality was not opposable to Costa Rica as per the test in Article 10.28 of DR-CAFTA.Footnote ⁶⁰ Second, Costa Rica contended that Mr Aven's convenient switch of nationality to take advantage of DR-CAFTA constituted an ‘exercise of [nationality planning], the consequence of which should be to deny him the protection of’ the treaty.Footnote ⁶¹

The tribunal disagreed with Costa Rica on both fronts. It held that, in line with the customary law of diplomatic protection, the ‘dominant and effective’ nationality test in Article 10.28 should be construed as applying in a dual home–host State nationality scenario, and not where, as in the present case, the second nationality involves that of a third State.Footnote ⁶² The tribunal then ruled that Mr Aven's use of his United States and Italian nationalities (one to make the investment and the other to access the treaty) was simply ‘the privilege of holding two passports in, at times, a hostile world’ and thus did not amount to illegitimate nationality planning’.Footnote ⁶³

The Aven tribunal is not the first to find that the ‘dominant and effective’ nationality test does not cover dual home–third State nationals. In Adel v Oman, an ICSID panel opined obiter dicta that the ‘dominant and effective’ nationality test in the United States–Oman FTA, which is identical to that of DR-CAFTA, ‘aimed at preventing claims by dual nationals of both State parties … from seeking to use the FTA to claim against their own State of … nationality’.Footnote ⁶⁴

Latest developments in international investment law concerning the nationality of individuals allow a conclusion that arbitrators have increased the range of qualified investors beyond what treaty parties could have anticipated at the time they signed their IIAs. Depending on who you ask (or perhaps more importantly, who you appoint as arbitrator), investors from both Contracting Parties enjoy the benefit of having two nationalities that can freely invoke against either party or even both. Moreover, irrespective of a potential agreement to the contrary in an IIA, nationals from the home country and a State that is not a party to the reciprocal arrangements embodied in the treaty can now opportunistically rely on one or both nationalities to benefit from these arrangements. This lenient approach to individual nationality planning exacerbates material inequities between already unequal disputing parties.

It is reasonable to expect that claims by dual nationals will continue to arise and that more tribunals will contribute to the growing and inconsistent investment jurisprudence in in this area. All States can do for now is to incorporate additional nationality criteria in their IIAs with the purpose of reducing exposure to ISDS and avoiding the legal uncertainty generated by investment jurisprudence. By doing so, States will ensure that these instruments only cover international investment disputes, that is to say, disputes between States and ‘foreign’ investors.Footnote ⁶⁵ After all, that the investor is foreign ‘is critical to the architecture of the system of international investment arbitration [and] without that criterion, the system would provide an impermissible forum for purely domestic disputes’.Footnote ⁶⁶

1. Corporate structuring

As previously explained, in addition to merely requiring incorporation in the home State party, most IIAs cover indirect investors or investments. A company within the same corporate group can therefore bring a treaty claim through an intermediate subsidiary and/or its shareholders, both minority and controlling. Likewise, a company can alter its organisational structure with the purpose of gaining access to an IIA signed by a State that offers more favourable protection than the one available in the company's original home State. The company can, for instance, incorporate a legal entity in the State Party to the targeted IIA and insert it into the corporate chain, thereby becoming a protected home State national. The company can also benefit from the treaty by purchasing shares in an existing holding company that has the nationality of the targeted home State. A domestic investor can equally employ this kind of corporate manoeuvrings, claiming, as a ‘round tripper’,Footnote ⁶⁹ the status of a ‘foreign’ investor. Indeed, domestic companies can ‘expatriate’ their investments through a legal entity located in a State that has concluded an IIA with their State of nationality.

The corporate vehicles typically used in these scenarios are shell or mailbox companies lacking an economic connection with the State of incorporation. Moreover, the company owing or controlling the corporate vehicle often remains a national of a non-treaty Party or the host State. The underlying objective of corporate structuring therefore becomes apparent: to disguise the real national origin of the claimant investor.

To understand the frequency with which corporate structuring occurs in practice,Footnote ⁷⁰ one simply needs to look at recent trends regarding the most frequent home States of claimants in ISDS disputes. As of 1 January 2019, the total number of publicly known ISDS claims had reached 942. Purported Dutch investors have brought 108 of these claims, making the Netherlands the second most popular home State of claimants after the United States.Footnote ⁷¹ According to a study by the Centre for Research on Multinational Corporations (SOMO) ‘over 75 per cent of Dutch BIT cases were brought by mailbox companies with no real economic substance in the Netherlands’,Footnote ⁷² companies whose ultimate or controlling parent was based in a country other than the Netherlands. The Netherlands is not alone however in providing a ‘base camp’ for corporate restructuring. Claimants also frequently use Cyprus and Luxembourg as a home country of convenience.Footnote ⁷³

Corporate nationality planning has proven a fertile area for jurisdictional objections. In cases where the claimant was a shell company or was controlled by third or host State nationals, respondent States have asked tribunals to dismiss the claim on the basis that the claimant was not the ‘real’ investor ultimately benefiting from the claim. The argument commonly advanced in this context is that, when ascertaining jurisdiction, tribunals should rely on considerations that are not found in the text of the applicable IIA, including the more restrictive diplomatic protection rules on corporate nationality (ie ‘genuine connection’) as well as the principles of abuse of process and lifting the corporate veil.Footnote ⁷⁴ Similar to natural persons, however, arbitral tribunals have shown marked reluctance to apply additional nationality criteria, preferring instead to adhere to the express wording of IIAs.Footnote ⁷⁵ For these tribunals, the structuring of an investment to obtain treaty protection for future disputes is considered as a ‘perfectly legitimate goal’,Footnote ⁷⁶ a position that has also been supported by most scholars.Footnote ⁷⁷ This formalistic approach extends to determinations of nationality under Article 25(2)(b) of the ICSID Convention, for tribunals are, in principle, bound to accept the definition of corporate nationality contained in IIAs when establishing ICSID jurisdiction.Footnote ⁷⁸

An illustrative and recent example of a strict textual interpretation of the definition of corporate investor is the recent decision in Isolux v Spain.Footnote ⁷⁹ In that case, Isolux Infrastructure Netherlands, B.V. (IIN), a mailbox company constituted in the Netherlands, initiated arbitration proceedings against Spain under the ECT. IIN was jointly controlled by a Spanish company and a Canadian company. Both had acquired INN and, thus, the qualifying Dutch nationality only two months before the enactment of Spain's long-anticipated energy regulatory framework subject to the claim. Spain challenged jurisdiction on grounds of corporate formality and abuse of process, arguing that INN's ownership was acquired for the sole purpose of gaining access to the ECT. In rejecting the objection, the tribunal first held that, as an entity established in the Netherlands, INN satisfied ‘los requisitos formales del Artículo 1(7) TCE [ie incorporation]’, irrespective of that company being owned by a national of the host State (ie Spain) and of a non-ECT signatory (Canada).Footnote ⁸⁰ The tribunal then held that IIN's acquisition did not constitute an abusive manipulation of the ECT, for ‘el conflicto es posterior a la restructuración discutida y a la colocación de la inversión en una sociedad holandesa’.Footnote ⁸¹

At this stage, it is important to mention that a number of tribunals have begun to set limits to corporate nationality planning in extreme cases, where the investor structured its investment shortly after the dispute arose (eg at the date on which the State adopted the relevant measure) or when the dispute was clearly foreseeable by the investor. Yet these tribunals represent the small minority. A well-known example is the decision in Philip Morris v Australia.Footnote ⁸² This dispute arose out of the so-called plain packaging legislation that Australia passed in November 2011. The legislation affected the shares held by the Swiss branch of the Philip Morris International (PMI) group in Philip Morris Australia (PM Australia). There is no BIT between Switzerland and Australia. In a failed attempt to access ISDS, the claimant, Philip Morris Asia Ltd, acquired all shares of PM Australia, becoming prima facie a protected investor under the Australia–Hong Kong BIT. The acquisition was completed almost a year after the Australian Government announced its intention to introduce the plain packaging measures, a period during which PMI has clearly stated its strong opposition to their enactment. The tribunal upheld an objection by Australia that the restructuring of the investment amounted to an abuse of process.

Relying on past investment jurisprudence,Footnote ⁸³ the tribunal held that ‘the initiation of a treaty-based investor-State arbitration constitutes an [abuse of process] when investor has changed its corporate structure to gain the protection of an investment treaty at a point in time when a specific dispute was foreseeable’.Footnote ⁸⁴ In the present case, the tribunal concluded, the claimant changed its corporate structure at a time when ‘it was reasonably foreseeable’ that the plain packaging legislation ‘would eventually be enacted and, consequently, a dispute would arise’.Footnote ⁸⁵ All claims were, therefore, deemed inadmissible. Respondent States involved in ongoing investment arbitrations are relying on this decision in the hope of barring nationality planners from accessing their treaties.Footnote ⁸⁶

It remains to be seen whether tribunals adopt a more balanced approach to the interpretation of treaty provisions defining eligible corporate investors, as the tribunal did in Philip Morris. Admittedly, the current state of affairs suggests otherwise. As it stands, arbitral practice leads to the conclusion that, with limited exceptions, arbitrators have been unwilling to prevent the use of corporations of convenience, which has significantly increased the range of qualified investors. Corporate restructuring is not, however, the only factor increasing host States’ exposure to ISDS claims. In their existing form, IIAs provide an extra layer of procedural protection in addition to broad nationality requirements: the possibility for companies to initiate multiple proceedings.

2. Multiple proceedings

We have seen how, when making an investment through a chain of companies incorporated in different States, part owners (ie minority or majority shareholders) and owners at various levels of the same group will become eligible nationals under various IIAs. Each of these entities can file its own separate claim against the host State for the same treaty breach in an effort to maximise their chances of success. For instance, both a foreign company established in the home State and the local subsidiary on which it holds shares could launch proceedings under the same IIA. Another possibility is that two or more entities within the group initiate multiple proceedings against the same host State but under different IIAs.Footnote ⁸⁷ ‘In more colloquial terms’, vague definitions of ‘corporate investor’ enable investors to have ‘several bites at the cherry’.Footnote ⁸⁸ The occurrence of multiple proceedings is not uncommon in practice.Footnote ⁸⁹

For example, the Dutch company OI European Group (OIEG) and its majority local shareholder, Owens-Illinois de Venezuela C.A. (OIdV), brought two parallel ICSID claims against Venezuela under the Netherlands–Venezuela BIT. Both disputes concerned the expropriation of the two largest glass container production plants in Venezuela. In the first arbitration, initiated by OIEG, the tribunal rendered an award in favour of the claimant.Footnote ⁹⁰ As a result, Venezuela sought to suspend the second arbitration. It contended that allowing OIdV to proceed with that arbitration would potentially result in ‘double recovery’ by the same company and ‘would constitute an abuse of process’.Footnote ⁹¹ In a decision of 13 November 2017, the tribunal refused Venezuela's application without providing reasons for doing so and despite acknowledging that ‘[t]he [the first arbitration] addresses mostly the same substantive matters at issue in the [second] proceedings’.Footnote ⁹²

In the very same year, another ICSID tribunal permitted the continuation of an ECT claim against Italy brought by Eskosol S.p.A., a locally incorporated company that was 80 per cent controlled by Blusun S.A., a Belgium company.Footnote ⁹³ The claim arose out of Italy's 2011 regulatory rollback on feed-in tariffs. Italy challenged jurisdiction on the basis that Eskosol's Belgian shareholder, Blusun, had previously initiated, and lost, another ICSID arbitration under the ECT against Italy for the same regulatory action.Footnote ⁹⁴ The tribunal saw nothing objectionable about Eskosol's claim. It considered that, since Blusun ‘only’ owned 80 per cent of Eskosol, it was difficult to treat both claimants as ‘effectively the same party’.Footnote ⁹⁵

The tribunals’ treatment of multiple claims in the OIdV and Eskosol is the rule rather than the exception, a rule that undeniably exacerbates existing imbalances in the investment protection regime. Other side effects of the rule that are troublesome for respondent States include excessive costs, duplication of compensation and procedural unfairness. A further risk associated with multiple proceedings is the potential for contradictory outcomes.Footnote ⁹⁶ As one author puts it: ‘[b]eyond the manifest unfairness of allowing one party to “play ‘till you win,” the ISDS approach [to multiple proceedings] also distorts incentives on all sides at the settlement stage, and facilitates opportunistic hold ups’.Footnote ⁹⁷

One may, however, take some comfort in the fact that, here too, investment tribunals are becoming more sensitive to this type of nationality planning. In a departure from earlier jurisprudence, the recent award in Orascom v Algeria relied on the doctrine of abuse of process to reject a claim resulting in parallel proceedings under different IIAs.Footnote ⁹⁸ In that case, a company registered in Luxembourg, Orascom TMT Investments, instituted ICSID proceedings under the Algeria–Belgium and Luxembourg BIT against Algeria. The company's ultimate shareholder had previously sued Algeria under other BITs for the same treaty breach by invoking the nationality of several subsidiary entities in the corporate chain. The tribunal held that, although structuring an investment to access an IIA ‘is not illegitimate’, an investor ‘may commit an abuse’ if it sues a host State ‘multiple times [through] various entities under the same control that are part of the vertical chain in relation to the same investment, the same measures and the same harm’.Footnote ⁹⁹

The Indian Government will likely invoke the decision in Orascom as part of its jurisdictional objections in two parallel UNCITRAL arbitrations initiated by the Vodafone Group against the State, one by Vodafone's Dutch subsidiary under the Netherlands-India BITFootnote ¹⁰⁰ and the other by Vodafone's UK companies under the UK-India BIT.Footnote ¹⁰¹

As a decision that imposes restrictions on nationality planning, Orascom is a positive development that sympathises with the objective for a more balanced regime. Once again, however, this decision represents an exception to the prevailing view that arbitrators should refrain from relying on non-treaty sources of general international law. In addition, although the principle of abuse of process can be seen as a useful tool to limit the use of nationalities of convenience, its application has been ‘rooted in highly abstract terms and raises its own problems of certainty and consistency’.Footnote ¹⁰² Put another way, international investment law lacks a consistent line of jurisprudence clarifying the dividing line between legitimate nationality planning and abuse of process.Footnote ¹⁰³

The foregoing discussion encapsulates a rather obvious, yet fundamental, statement: any effort to combat the practice of nationality planning ultimately lies in the hands of States. States are, after all, the ‘masters of the treaties’ and are accordingly advised to take all necessary steps to reduce the exposure created by the expansive nationality requirements with the aim of recalibrating the international investment regime.

1. The rule of non-responsibility (dual home–host State nationals)

Recognised as customary international law, the rule non-responsibility posits that, as a corollary of the principle of equality of States, a State is not responsible for injuries caused to their own nationals.Footnote ¹⁰⁹ The ICSID Convention incorporates this rule in Article 25(2)(a) which, as explained above, provides that jurisdiction does not extend to nationals of the respondent State. A very small number of IIAs include the rule of non-responsibility.Footnote ¹¹⁰ One of these treaties is the Mauritius–Egypt BIT, which defines ‘investor’ as ‘a natural person who derives his or her nationality in virtue of the laws of one of the Contracting Parties and who is not simultaneously a national of the other Contracting Party’.Footnote ¹¹¹

States that decide to incorporate the rule of non-responsibility in their IIAs should also consider specifying the time at which the investor has to hold the nationality of a State party other than the respondent State. The ICSID Convention, for instance, provides a double test of time in that regard: the investor must not be a national of both Contracting Parties ‘on the date the parties consented to submit the dispute to [arbitration] as well as on the date on which the request was registered’ by the Centre.Footnote ¹¹² This timing of nationality requirements, which mainly focuses on the date of the presentation of the claim, may result in an opportunistic renouncement of the nationality of the respondent State after the events given rise to the dispute have occurred. This is precisely what happened in Pey Casado v Chile, where the claimant, a dual Spanish–Chilean national, bypassed the dual nationality restriction in the ICSID Convention by renouncing his Chilean nationality only two weeks before instituting ICSID proceedings.Footnote ¹¹³ Lenient timings requirements may also enable a domestic investor to ‘internationalise’ their claims through the ex post facto acquisition of a second, qualifying nationality.

To counter the above-described opportunistic behaviours, States should consider incorporating additional dates of nationality, requiring that, besides the date of the submission of the claim, the investor must have the nationality of a Contracting Party other than the respondent State on the date when the injury giving rise to the claim occurred. An illustrative, and rare, example of such dates of nationality can be found in Article 1(3)(a) the Slovak Republic–Iran BIT:

These timing requirements, which reflect the position under customary international law,Footnote ¹¹⁵ are justifiable. Indeed, logic dictates that, to claim substantive protection under an IIA, an investor should also qualify as a national at the time when the treaty obligation was allegedly breached (eg by government regulatory measures affecting the investment). This date will normally precede the date of the presentation of the claim.

Another date that States are advised to consider is when the investor made the investment in the host State. In this regard, States should be mindful of the underlying purpose of IIAs, namely to promote and protect foreign investments, that is, investment made by nationals of the home State in the host State. As Dolzer and Schreuer rightly state: ‘the foreignness of an investment is determined by the investor's nationality’.Footnote ¹¹⁶ In this regard, it is difficult to see how an investment made by a dual home-host State national will qualify as ‘foreign’,Footnote ¹¹⁷ especially if the investor has benefited from their status as domestic investor to make the investment or if they have developed a stronger connection with the host State.

In brief, States may prevent nationality planning by dual home–host State nationals by incorporating the rule of non-responsibility in their IIAs. That rule should be accompanied by three relevant dates on which the investor must have the nationality of a State Party other than the respondent State: (a) the time the investor submits the dispute to arbitration; (b) the time the alleged interference with the investment occurred or commenced; and (c) the time the investment was made.

2. The rule of dominant and effective nationality (dual home–host State nationals and dual home–third State nationals)

States that wish to limit claims by their own investors in a non-absolute manner can incorporate the customary rule of dominant and effective nationality into their IIAs. As explained above, this rule provides an exception to the rule of non-responsibility in cases where it can be established that the aggravated dual home–host State national maintains stronger connections (personal, economic, political, etc) with the home State. Although the rule is limited to dual home–host State nationals in the realm of diplomatic protection, this does not mean that Sates are prevented from extending its scope of application to dual home–third State nationals.

Some IIAs contain the rule of dominant and effective nationality as part of the definition of protected natural persons.Footnote ¹¹⁸ The standard formulation can be found in the above-mentioned Article 10.28 of DR-CAFTA:

The wording of this provision is problematic on at least three levels. First, it fails to indicate the category of dual nationals who fall under its scope of application, a drafting defect that may result in unintended interpretations. As previously explained, the tribunal in Aven rejected Costa Rica's argument that Article 10.28 of DR-CAFTA applied to both dual home–host State nationals and dual home–third State nationals. The tribunal reasoned that: ‘[t]hrough inclusion of the expression “dominant and effective nationality” in Article 10.28, the DR-CAFTA Parties intended to incorporate … customary international law for the treatment of multiple nationality in diplomatic protection cases’.Footnote ¹¹⁹ This reasoning sits poorly with the lex specialis maxim and the terms of Article 10.28.

One must first note that the dispute between Mr Aven and Costa Rica was not a diplomatic protection case, but rather a dispute between a State and a private investor arising from an investment treaty. Consistent with this observation, nothing in Article 10.28 suggests that the parties ‘intended’ to apply the customary law version of the rule of dominant and effective nationality. That provision simply refers to ‘a natural person who is a dual national’, without specifying the States of nationality of that person. If we follow the reasoning of the tribunal in Rawat v Mauritius, silence of Article 10.28 on the category of dual nationals would seem to point to the inclusion, rather than the exclusion, of investors who hold the nationality of the home contracting party and a State other than the respondent State.

To avoid interpretations such as that in Aven, States are advised to clarify the categories of dual nationals that fall within the scope of application of the dominant and effective nationality test. This is what Iran and the Slovak Republic did in their BIT. Besides excluding dual home–host State nationals through the rule of non-responsibility, the BIT provides that ‘[a] natural person that is a dual national of either Contracting Party and any non-Contracting Party, shall be deemed to be exclusively a national of the State of his or her dominant and effective nationality’.Footnote ¹²⁰ In contrast, the Netherlands has proposed to limit the test to dual home–host State nationals.

On 22 March 2019, the Dutch Ministry of Foreign Affairs published the final version of its new model BIT, which the Netherlands intends to use for the renegotiation of 78 Dutch BITs with non-EU countries.Footnote ¹²¹ To the surprise of many, the Dutch model BIT contains provisions that aim at preventing individual and corporate nationality planning. With respect to individuals, Article 1(b)(ii) provides that ‘[a] natural person who has the nationality of the Kingdom of the Netherlands and the other Contracting Party is deemed to be exclusively a natural person of the Contracting Party of his or her dominant and effective nationality’.Footnote ¹²²

Should States wish to apply the rule of dominant and effective nationality to both categories of dual nationals, they should follow the treaty drafting approach of the Iran–Slovak Republic BIT.

The second drafting defect of the standard formulation of the rule of dominant and effective nationality is the absence of criteria of dominance. The relevant enquiry here is to establish, by reference to factual ties with each country, which of the two ‘conflicting’ nationalities (home/host State, or if applicable, home/third State) is the prevalent one. More specifically, the test should focus on establishing whether a dual national is ‘foreign’ enough, that is, sufficiently connected to the home State, to render ‘international’ a dispute with the respondent State.

In Ballantines v Dominican Republic, another dispute initiated under DR-CAFTA, a tribunal offered some guidance regarding the standard to determine the dominant and effective nationality in the absence of an express agreement in the treaty. This time the claimants were dual home–host State nationals from the United States and the Dominican Republic. The tribunal rendered an award on 3 September 2019 deciding by a majority that the claimants’ dominant and effective nationality was that of the respondent State and, thus, that it lacked jurisdiction under DR-CAFTA.Footnote ¹²³ To the best of the author's knowledge, this is the first tribunal that has interpreted de facto the dominant and effective test under an IIA. In line with the reasoning adopted in Aven, the tribunal first noted that ‘the expression “dominant and effective” [in DR-CAFTA] is [undoubtedly] rooted on customary international law’, thereby accepting that the test's scope is limited to dual home–host State nationals.Footnote ¹²⁴ It then considered that, since ‘the treaty does not prescribe’ the criteria of dominance, it is necessary to ‘take guidance from customary international law’.Footnote ¹²⁵

After engaging in its own exegesis of non-ISDS cases on dual nationality, such as Case No. A/18 brought before the Iran–United States Claims Tribunal, the tribunal held that relevant factors to assess the effective and dominant nationality of investors include ‘the State of habitual residence, the circumstances in which the second nationality was acquired, the individual's personal attachment for a particular country, … the center of the person's economic, social and family life’.Footnote ¹²⁶ The tribunal added that ‘the specific context of DR-CAFTA’ required the examination of other relevant factors', such as ‘the investment itself, the status of investor as well as other circumstances surrounding those elements’.Footnote ¹²⁷

Although not exhaustive, the tribunal's list establishes an objective factual inquiry that can serve as a point of reference for States that want to incorporate the dominant and effective nationality requirement into their IIAs. The 2019 Australia–Hong Kong BIT contains a similar list.Footnote ¹²⁸ However, it is advisable that, when negotiating the criteria of dominance, States include the facts surrounding the investment as a strong indicator of dominant nationality and, in particular, the status upon which the investor relied to obtain the benefits deriving from the investment. It is doubtful whether an investor who, for instance, presents him or herself as a domestic investor to establish and operate the investment in the host State (see García Armas I) should qualify as investor having home State nationality.

Lastly, States are encouraged to define the relevant times in which a tribunal should conduct any dominance analysis. Here, this article recommends that States refer to the same dates as for the rule of non-responsibility. As such, the home State nationality should be the dominant one on the date the investor submitted the dispute to arbitration, the date the alleged treaty breach and the date the investment was made. With these timings of nationality requirements, States will reduce the possibility that a dual national disguises his or her ‘real’ dominant or effective nationality before bringing the claim under the treaty.

1. Codification of ‘pre-existing/foreseeable dispute’ jurisprudence and ‘self-invoking’ denial of benefits clauses

Drafting approaches to prevent corporate structuring can vary.Footnote ¹²⁹ A rare, innovative approach found in some recently concluded IIAs is the codification of arbitral decisions declining jurisdiction on grounds of abuse of process. As an illustration, Article 16(3) of the Dutch Model BIT provides:

This provision is a welcome attempt on the part of the Netherlands to incorporate the essence of the test applied by the tribunal in Philip Morris. The clause de-legitimatises a change of nationality with the purpose of accessing the treaty if there was a pre-existing dispute or a future dispute was foreseeable. The second sentence reinforces the application of this test to domestic investors that channel their investments through a State Party to an IIA in order to sue their country of origin. While Article 16(3) is a welcome effort to combat corporate nationality planning, its wording is not without problems.

For one, it does not define the term ‘dispute’ and its materialisation, nor does it prescribe the elements necessary to categorise a dispute as ‘foreseeable’. Does, for instance, a dispute arise at the time the investor knew of the State's intention to adopt the legislation harming the investment? If so, should the investor also oppose the enactment of that legislation for the dispute to materialise? Moreover, Article 16(3) fails to define the factors that the tribunal should consider when determining whether the purpose of the investment structuring was to ‘gain access to the treaty’. States willing to follow the approach adopted by the Netherlands are advised to clarify these issues in order to limit the interpretive discretion of arbitrators, ensuring that the treaty will be applied in accordance with the intention of the parties.

Another, more effective alternative to prevent corporate structuring is the inclusion of a self-invoking DOB clause. Feldman writes that ‘[w]hen available, a [DOB] provision, rather than the principle of abuse of right, should be a respondent's first choice when facing a claim brought by a company that appears to lack any genuine connection to its purported home State’.Footnote ¹³¹

The function of DOB clauses is twofold. First, they exclude from treaty coverage shell companies that are controlled by third or host State nationals. Second, and as a result, DOB clauses prevent the use of corporations of convenience through corporate structuring.Footnote ¹³² While having gained increasing popularity in investment treaty making, DOB clauses are not a typical feature of IIAs.Footnote ¹³³ In addition, as presently drafted in IIAs, DOB clauses have created undesirable room for interpretation by arbitrators.

Generally, DOB clauses in IIAs contain wording such as ‘[e]ach Contracting Party reserves the right to deny’, as in Article 17(1) of the ECT, or ‘[a] Party may deny the benefits’, as in Article 10.12(2) of the United States–Peru FTA:

Assuming that the substantive conditions of a DOB clause (eg substantive business activities and control) are fulfilled, the standard language of the clause has raised the question whether the denying State can benefit from the automatic application of the provision or, to the contrary, whether the it must deny benefits before the putative investor initiates proceedings. It seems logical to assert that a respondent State should be entitled to deny treaty benefits at the time such benefits are sought by the investor through a request for arbitration. Several tribunals, most of them constituted under the ECT, have however held that host States must proactively deny treaty benefits and give notice before a dispute arises, leaving the utility of such clauses in doubt.Footnote ¹³⁵

The tribunal in Plama v Bulgaria was first in holding that a DOB clause has mere prospective effect.Footnote ¹³⁶ In that case, Bulgaria exercised its right under Article 17(1) of the ECT to deny the substantive investment protections of the treaty to Plama Consortium Limited, a mailbox company established in Cyprus and controlled by nationals of a non-ECT party. Bulgaria claimed that Article 17(1) operated retroactively from the time the investor initiated arbitration. The tribunal disagreed, finding that Bulgaria's exercise of the DOB right would not accord with the object and purpose of the ECT to ‘promote long-term co-operation in the energy field’.Footnote ¹³⁷ On this point, the tribunal reasoned that accepting a retrospective denial from the time proceedings are commenced would prevent the investor from planning in the long term for such an effect and would lure the investor with ‘false’ legitimate expectations.Footnote ¹³⁸ According to the tribunal, before becoming a respondent, the host State should ‘properly’ notify the investor ‘of the potential effect of Article 17(1) […]’ through, for instance, ‘a statutory provision in investment or other laws’.Footnote ¹³⁹

ECT tribunals have so far followed this approach, the most recent being the tribunal in Masdar v Spain.Footnote ¹⁴⁰ The claimant in that case was Masdar Solar & Wind Cooperatief U.A., a Dutch ‘special purpose vehicle’ controlled by an Abu Dhabi company. Abu Dhabi is not a party to the ECT. In an award of 16 May 2018, the tribunal found that, concordantly with ‘a majority of tribunals’, the host State must exercise its DOB right before the dispute arises, ‘whether by adopting legislation denying benefits generally (or to a specific sector or sectors) or by promulgating measures directed at specific investors’.Footnote ¹⁴¹

It should be noted, however, that the position in favour of the retroactive application of DOB clauses extends beyond ECT arbitrations.Footnote ¹⁴² In Ampal v Egypt, an ICSID tribunal held that Egypt's pre-arbitration exercise of the DOB clause in the US–Egypt BIT did not vitiate its jurisdiction. Referring to Article 25(1) of the ICSID Convention, the tribunal found that ICSID jurisdiction ‘is to be assessed at the time […] the investor's Request for Arbitration is registered by the Centre’, and ‘there cannot be an embedded conditionality in the Treaty which could be triggered after’ that date.Footnote ¹⁴³

A requirement for a pre-arbitration invocation of a DOB clause appears somewhat detached from realities on the ground. As a practical matter, it is difficult to see how the host State will be able to assess the investor's compliance with the substantive conditions of a DOB clause before a dispute with the investor has arisen. It is in fact when the host State becomes a respondent in an investment arbitration that it becomes aware of the corporate structure and economic activities of investors investing in its territory. When restructuring their investments through third States or from the host State to access treaty protection, investors should simultaneously accept the risk envisaged in a DOB clause. States that agree with this view and wish to avoid the disabling of DOB clauses by investor-friendly interpretations, are advised to consider drafting DOB clauses in a way that treaty benefits can be denied automatically (eg ‘benefits shall be denied’ or ‘benefits shall not be available to an investor’).

Strikingly, according to UNCTAD, only nine out of 2,577 mapped IIAs contain this restrictive wording. The new Belgium–Luxembourg Economic Union (BLEU) Model BIT is among those treaties. Its DOB clause states that ‘[f]or avoidance of any doubt, the benefits of this Agreement shall be denied if the preconditions set down in paragraph 1 are fulfilled at time when the claim is submitted pursuant to Article 19(D)’.Footnote ¹⁴⁴ The Serbia–United Arab Emirates BIT contains similar language.Footnote ¹⁴⁵A provision of this nature will guarantee not only that the investor will be denied treaty benefits after the initiation of the proceedings but also that the denial will apply retroactively from that date.

The potential inclusion of a self-invoking DOB clause should be an item for discussion in the context of the negotiations for the reform of the ECT. The Energy Charter Conference, the intergovernmental body that oversees the ECT, issued a decision on 6 November 2019 approving a mandate, procedure and timeline for negotiations on the modernisation of the treaty.Footnote ¹⁴⁶ Negotiations begun in December, following a wide-range of proposals sent by a majority of the 55 Members of the Conference.Footnote ¹⁴⁷ A cursory read of these proposals shows that only two States (Albania and Georgia) are concerned about the position adopted by investment tribunals in favour of the retroactive application of the DOB clause in Article 17(1) of the treaty. Albania notes that ‘the Treaty does not define when and how a Contracting Party can apply this clause’.Footnote ¹⁴⁸ Similarly, Georgia highlights that, as practice has shown, Article 17(1) ‘does not apply automatically and requires additional action from the Contracting Party to “reserve” [its] right [of denial]’.Footnote ¹⁴⁹ Both States then propose ‘to clearly define’ the procedural requirements for the application of Article 17(1), but fail to formulate concrete amendments to the clause.

As previously argued, instead of embarking on the difficult task of defining the procedure for the innovation of Article 17(1), ECT Signatories should consider making the operation of the clause automatic, by replacing the discretionary wording ‘reserves the right to deny’ with the mandatory language the ‘benefits shall be denied’. This mandatory language eliminates the interpretive discretion of tribunals, thereby ensuring that the purpose of Article 17(1) is no longer rendered meaningless.

The European Union and its Member States, and more particularly Spain, which has faced over 40 ECT claims, are advised to follow this treaty drafting approach, and to propose it during the negotiations for the modernisation of the ECT. This advice is of utmost importance here if we consider that, as it was the case in Masdar and Isolux, investors often use Dutch empty shells to obtain protection against Member States under the ECT. In its latest report, the Corporate Europe Observatory and the Transnational Institute indicate that ‘an extraordinary 23 out of 24 supposedly ‘‘Dutch’’ investors who filed ECT-lawsuits by the end of 2017 [were] mailbox companies’.Footnote ¹⁵⁰

2. Waiver provisions (multiple proceedings)

The text of most IIAs are not oriented towards mustering the risks associated with multiple proceedings. These risks include procedural unfairness (ie the host State having to defend multiple claims related to the same investment and State measure), the potential for contradictory outcomes, additional costs and double recovery. More crucially, in the thirty-sixth meeting of the UNCITRAL Working Group III, participants pointed out that ‘multiple proceedings distorted the balance of rights and interests of relevant stakeholders’ and ‘undermined predictability more generally’.Footnote ¹⁵¹ The Working Group III further noted that the text of most IIAs lack ‘a framework to address multiple proceedings’ and considered whether concerns relating to the risk associated with this phenomenon ‘warranted some form of reform’.Footnote ¹⁵² Reform options that are often suggested to prevent multiple proceedings include ‘the use of lis pendens [and] res judicata’ and the ‘proactive use of consolidation’.Footnote ¹⁵³ These mechanisms have certain shortcomings.

The applicability of the lis pendens and res judicata doctrines in investment arbitration proceedings is questionable.Footnote ¹⁵⁴ This is because, as we have seen, investment tribunals tend to disregard rules and principles of national and international law not found in the text of the treaty, no matter how auspicious or appropriate they may appear. With respect to lis pendens, for instance, Yannaca-Small explains that ‘very few investment arbitration tribunals have expressed an opinion on parallel proceedings and the applicability of the lis pendens principle’.Footnote ¹⁵⁵ Even if an investment tribunal was willing to consider the doctrines, it will be rather hard for the respondent State to show that the requirements of lis pendens and res judicata (identity of the parties, cause of action, and object) have been satisfied. On this point, the UNCITRAL Working Group III itself has acknowledged that the ‘triple identity requirement may make it difficult to apply the doctrine[s] to concurrent proceedings’ in investment treaty arbitration.Footnote ¹⁵⁶ This is particularly true for the condition ‘identity of parties’, which has been narrowly construed by investment tribunals. The tribunal in Eskosol expressed the overall view here, that is, a company and its shareholders are legally distinct entities and consequently cannot be treated as the same party for the purposes of lis pendens and res judicata.Footnote ¹⁵⁷

As a procedural device that aims at aggregating two or more closely related disputes, consolidation may prove effective in limiting multiple proceedings, provided this mechanism is available under the applicable IIA. IIAs rarely contain provisions on consolidation, neither do UNCITRAL or ICSID, the most frequently used arbitral fora for investment disputes.Footnote ¹⁵⁸ A typical consolidation provision can be found in the Dutch Model BIT. Article 19(7) provides that ‘when two or more claims have been submitted separately to arbitration under’ the treaty ‘and the claims have a question of law or fact in common and arise out of the same events or circumstances, either party to the dispute may seek a consolidation order at either Tribunal’.Footnote ¹⁵⁹

Existing consolidation mechanisms only address situations where different investors launch claims with respect to the same State measure and the same IIA. However, they do not establish the possibility of cross-treaty claim consolidation, that is, consolidation of multiple proceedings arising out of different IIAs, an eventuality that was present in Orascom and Vodafone. This may be explained by the fact consolidation based on two or more IIAs is problematic because, as distinct sources of law, the relevant treaties might ‘contain differing substantive obligations, as well as diverging time limits, procedural obligations and dispute settlement forums’.Footnote ¹⁶⁰ Moreover, it is important to note that, regardless of the potential for (cross-treaty) consolidation, a decision to consolidate proceedings does not ultimately lie in the hands of the respondent State but the arbitrators.

A better jurisdictional device to limit multiple proceedings is a treaty provision that requires the claimant investor, including its shareholder and subsidiaries, to waive its right to seek further remedy or to pursue any pending claim, under the same or different IIAS. At this stage, it may not come as a surprise that a very limited number of IIAs contain a waiver provision, also known as ‘no-U-turn’.Footnote ¹⁶¹ One of the more comprehensive and recent waiver provisions can be found in Article 9.17 of the EU–Singapore FTA, which, in its relevant parts, provides:

The effects of this provision seem potentially far-reaching. First, the clause requires that the claimant investor, including its shareholders and subsidiaries, ‘withdraws’ any parallel claim ‘concerning the same treatment as alleged to breach’ the treaty (OIdV v Venezuela scenario). Second, it prevents the claimant investor from initiating arbitration proceedings if an ‘award concerning the same treatment as alleged to breach’ the treaty had previously been rendered in a claim brought by one of the claimant's shareholders (Eskosol v Italy scenario). Third, it requires that the claimant investor, including its shareholders and subsidiaries, ‘withdraws any pending claim’ under ‘any other treaty’ when those claims concern ‘the same treatment as alleged to breach’ the treaty (Orascom v Algeria and Vodafone v India scenario). In sum, the waiver provision contained in Article 9.17 of the EU–Singapore FTA provides a good illustration of the kind of mechanisms that can be used to avoid multiple proceedings.