The relationship between international investment law and EU law has been a stormy one. The existence of bilateral investment treaties between the EU Member States (intra-EU BITs) has culminated in the measures terminating the BITs,1 and numerous arbitral cases in the uncertain situation of whether the enforcement of the awards is precluded by EU law.2 Simultaneously, the fate of the intra-EU applicability of the Energy Charter Treaty (ECT) remains unclear. The EU, conversely, continues to develop its external investment policy and conclude international investment agreements with third States, as the Court of Justice of the European Union (CJEU) has given a green light for their conclusion.3 In all its storminess, the conflict between the two regimes provides a fruitful insight into how the arbitral tribunals acting outside the EU legal order have viewed the EU and EU law, and how they have treated the demand for the preservation of the autonomy of EU law in the investor-State dispute settlement practice.
It is well established that the EU legal order is ‘its own legal system’ and ‘an integral part of the legal systems of the Member States’, and EU law has primacy over national laws of the Member States.4 The CJEU has summarised that
according to settled case-law of the Court, [the] autonomy of EU law with respect both to the law of the Member States and to international law is justified by the essential characteristics of the European Union and its law, relating in particular to the constitutional structure of the European Union and the very nature of that law. EU law is characterised by the fact that it stems from an independent source of law, namely the Treaties.
Thus, the EU legal order is by nature autonomous in relation to both national laws of the Member States and international law.6 Whereas internal autonomy entails the independency of EU legal order from the national legal orders of the Member States, the external dimension has been described to entail ‘the idea that the integrity of EU law and the EU legal order should not be undermined by the international action of the Union or the Member States’.7 This feature of EU law has brought about a line of cases where the CJEU has assessed the limits of the right of judicial bodies external to the EU legal order to apply and interpret EU law. According to this case law, an external judicial body cannot affect the division of competences within the EU or give interpretations on EU law binding within the EU.8 Furthermore, the Court stated in Opinion 1/17 that tribunals ‘outside the EU judicial system … cannot have the power to interpret or apply provisions of EU law other than those of the [investment treaty] or to make awards that might have the effect of preventing the EU institutions from operating in accordance with the EU constitutional framework’.9
Although the characteristics of the EU legal order are rather well established within the Union, the question of how they relate to competing obligations raised by international law remains. A prominent example of the culmination of the tensions between EU law and international law has been the surge of investment arbitration between an EU Member State and an investor from another Member State (intra-EU disputes). The investor-State arbitral tribunals are constituted based on investment treaties – either BITs, multilateral treaties such as the ECT, or numerous free trade agreements containing provisions on investment protection and dispute settlement (such as the new international investment agreements concluded by the EU (EU IIAs)). Although EU law may be relevant in a case, as will be discussed below, the tribunals are bound by the rules of international law.10 How far can, and should, the tribunals then go in respecting the special features of EU law, if that would entail a conflict with international law? Should the tribunals, acting outside the EU legal order, take the EU law issues into consideration, or are the issues merely internal to the Union?
Taking the topical questions of the principle of autonomy of EU law and the treaty conflict in intra-EU cases as well as the EU’s State aid rules in investor-State arbitration as examples, the article discusses how the EU and EU law have been viewed by the arbitral tribunals. By analysing publicly available arbitral awards,11 the article examines how the tribunals have understood the nature and the status of EU law and how they have treated EU law in practice, in questions relating either to the jurisdiction of the tribunals or to the substantive provisions of the investment treaties in intra-EU cases (Section 2).12 Then, the article turns to discuss whether there are means in EU law and international law that would enable a mutual understanding and appreciation of the central features of the other legal regime (Section 3). Lastly, it will be discussed whether any implications for the status and treatment of EU law can be drawn from the intra-EU cases and applied to the extra-EU practice (Section 4). Section 5 concludes.
2. EU law in intra-EU investment arbitration
2.1. How have the arbitral tribunals seen the nature of EU law?
The status and the effect of EU law in intra-EU investment arbitration has proved to be a pivotal question. Besides the possibility of EU law being relevant in the dispute through the applicable law provisions in the investment treaties and being a part of the relevant international law to be taken into consideration in the interpretation and application of the treaty, it can form a part of the facts in the assessment of the alleged breach of the treaty standards. Furthermore, the intra-EU investment tribunals have faced questions of whether the EU Treaties and the accession of States into the Union have rendered the investment treaties and their jurisdiction invalid. The view of the arbitral tribunals on these questions will be discussed in more detail in Section 2.2, Section 2.3, Section 2.4 and Section 2.5, but first an overview will be provided on how the tribunals have generally understood the nature of EU law.
Investment treaties usually contain provisions on the applicable law for the settlement of the merits of the dispute.13 The contents of such provisions may vary, and some treaties mention international law, some national law, and some both. These provisions are not purely decisive on the determination of applicable law; general rules of treaty interpretation also play their part in such assessment.14 An applicable law provision may also be found, for example, in Article 42 ICSID Convention, according to which
[t]he Tribunal shall decide a dispute in accordance with such rules of law as may be agreed by the parties. In the absence of such agreement, the Tribunal shall apply the law of the Contracting State party to the dispute (including its rules on the conflict of laws) and such rules of international law as may be applicable.
The application of international law and national law differs in treaty-based investment arbitration as relevant international law is taken into account together with the treaty, and national law, in turn, may come into play merely as a factual matter.16 Thus, the understanding of the nature of EU law plays a crucial role in determining whether the tribunals have to interpret and apply EU law, or merely to ‘take it into consideration as a fact’.17 This question has been particularly relevant in the assessment of whether the tribunals in fact interpret EU law and whether that conflicts with the autonomy of EU law.
Arguably, the understanding of the status of EU law links to how the nature of the EU itself is perceived. The tribunals in intra-EU cases have tended to understand EU law to be, by nature, international law, as well as being a part of domestic legal orders of the Member States (‘dual nature of EU law’).18 Perhaps the most well-known case engaging with the assessment of EU law is Electrabel v Hungary, where the Tribunal stated that EU law has a ‘multiple nature’ and ‘is a sui generis legal order, presenting different facets depending on the perspective from where it is analysed’.19 The Tribunal acknowledged that EU law forms a part of national legal orders of the Member States but argued that, from the perspective from which it looks at the issue, EU law should primarily be understood as international law, being based on international treaties. According to the Tribunal, this reasoning applies both to primary and secondary EU law. The Tribunal concluded that ‘there is no fundamental difference in nature between international law and EU law that could justify treating EU law, unlike other international rules, differently in an international arbitration requiring the application of relevant rules and principles of international law’. By contrast, it was also held that EU law should be taken into consideration as a part of national legal order of the Member State, and thus applied as a fact in the arbitration.20 Some later tribunals have in principle agreed with the reasoning in Electrabel, but they differentiated between primary and secondary EU law. For example, the Tribunal in Vattenfall v Germany agreed that EU law is international law as it derives from international treaties. It noted, however, that a more ‘exact way to say’ would be that EU law also contains internal rules that are not part of international law.21
The special nature of EU law as an autonomous legal order or the ‘constitutional nature’ of the EU Treaties have been seen as not preventing EU law from being international law.22 However, in recognising the nature of EU law as international law, or ‘a subsystem of public international law’,23 some tribunals have emphasised that there are limits to its application. The Tribunal in Eskosol v Italy noted that EU law is not general international law, but it can rather be characterised as ‘a special species of international law’. As such, EU law does not prevail over other fields of international law, but ‘they exist side-by-side’.24 This is a logical argument: although EU law has supremacy within the EU legal order, it is not superior in relation to other fields of international law. Similarly, it has been stated that EU law does not constitute principles of general international law.25 The understanding by the tribunals of EU law as forming a part of international law is in line with the understanding of EU law under international law generally, which can be summarised as seeing EU law as ‘a special and potentially self-contained subsystem of public international law’.26
However, the tribunals seem unanimous in the fact that, when necessary, EU law can be applied in the arbitration as domestic law, forming part of the facts applicable to the case.27 The Tribunal in Eskosol, for example, stated that ‘an ECT tribunal [can] consider EU law as a matter of fact if potentially relevant to the merits of a dispute, just as an ECT tribunal may consider a State’s domestic law as part of the factual matrix of a case’.28As noted above, this understanding of the role of national law in investment arbitration applies to international investment law in general, and not only to intra-EU cases.
When faced with questions of EU law, the tribunals tend to refer to the case law of the CJEU and follow the interpretation of EU law laid down there.29 This suggests that the tribunals recognise the CJEU’s role as the ‘ultimate interpreter’ of EU law, flowing from Article 19 Treaty on European Union (TEU).30 However, the tribunals have emphasised that this role is not without limits. The Tribunal in Eureko v Slovakia drew a distinction between a monopoly to interpret EU law and a monopoly to give ‘final and authoritative interpretation’. Whereas it recognised that the CJEU has the latter, it rejected the argument that it has both.31 Other tribunals have similarly agreed that the CJEU does not have a monopoly over the interpretation and application of EU law.32 The Tribunal in PL Holdings v Poland noted that ‘so far as the Tribunal knows, no other jurisdiction in the world has asserted a monopoly – much less succeeded in asserting a monopoly – over the interpretation and application of its law, even though it may of course claim to have “the last word” on the meaning of its law’.33 Similarly, the tribunals have acknowledged that they have no capacity to rule on the legality of State actions under EU law, but their jurisdiction is limited to assessing breaches of investment treaties.34 Thus, the autonomy of EU law and the central role of the CJEU in determining the ‘last word’ in the interpretation of EU law appears to be recognised by the tribunals, although not without limits.
2.2. Assessment of the effect of EU law on the jurisdiction of the tribunals
Perhaps the most controversial question in the interplay of EU law and investment arbitration has been – and remains – whether the investment tribunals have jurisdiction over the disputes based on intra-EU investment treaties. From a practical point of view, the answer to this has significant ramifications. From the EU law perspective, the autonomy of EU law is compromised if the jurisdiction is considered valid, while from the international investment law perspective the denial of jurisdiction entails that the investors cannot use the dispute settlement mechanism provided in the investment treaty, and thus cannot practically invoke the substantial provisions of the treaty.
The tribunals tend to take the point of view of international law in the determination of their jurisdiction and assessment of the issue of whether EU law has rendered it invalid in intra-EU cases. As the relevant Articles 30 and 59 VCLT35 apply when there is incompatibility between the treaties or the subject matter of the allegedly conflicting treaties is the same, it has been crucial to determine the scope of investment protection under EU law. Several tribunals have thus engaged in analysis – albeit sometimes rather superficial – of the protection of investments under EU law. The tribunals have generally argued that investment treaties offer broader substantial protection in comparison to EU law, especially the provisions on free movement of capital and freedom of establishment, which are of relevance here. Furthermore, the tribunals have considered that the dispute settlement mechanisms under EU law do not equate to investment arbitration. Thus, Articles 30 and 59 VCLT are not seen as applicable to the treaty conflict between EU Treaties and the BITs or the ECT, as there is no treaty conflict to begin with.36 In relation to Article 59 VCLT, it has also been necessary to assess whether the Member States have intended to replace the BITs with EU law. This argument has also been rejected by the tribunals. For example, the Tribunal in Oostergetel v Slovakia argued that the objective of EU law is to create a common market, whereas a BIT aims to provide ‘guarantees for the investor’s investment’.37
Thus, the tribunals have generally not directly rejected the relevance of EU law but instead have engaged in the assessment of the nature and content of the EU legal order, in order to determine whether the VCLT provisions on treaty conflict are applicable. Were the tribunals to conclude that EU law and investment treaties do cover the same subject matter, they would be faced with a more difficult assessment: which treaty prevails under international rules of treaty conflict? The Tribunal in Electrabel argued that ‘from whatever perspective the relationship between the ECT and EU law is examined, … EU law would prevail over the ECT in case of any material inconsistency’.38 The Tribunal in Charanne v Spain agreed with this line of reasoning,39 and similar argument has also been presented in the literature.40 Thus, the conclusion that EU law and investment treaties do not share the same subject matter has practically provided an ‘escape route’ from having to conclude – under the international law rules of treaty conflict – that EU law could prevail over the jurisdiction of the arbitral tribunals.
In addition to the analysis on the existence of treaty conflict between the two regimes, the Tribunal in Eureko recognised that the wider protection provided by the BIT in comparison to investment protection under the EU Treaties may lead to a breach of the principle of non-discrimination of EU law. It argued, however, that this possible breach did not mean that there had not been a valid consent to arbitrate by the parties.41 This shows that while the tribunals may recognise real issues under EU law following from the exercise of their jurisdiction under the investment treaties, they still consider themselves as being constituted under international law and having the obligation to respect the binding nature of valid international treaties. The Tribunal further stated that ‘it is not for an arbitral tribunal to cancel rights created by a valid treaty in order to safeguard a State party against the possibility that it might one day decide to apply the treaty in a way that could violate its obligations under one or more other later treaties’.42
In addition to the direct assessment of EU law, the tribunals allowing the Commission to submit amicus curiae may also be seen as a sign of recognition of the relevance of EU law. In JSW Solar v the Czech Republic, the respondent State had in fact not raised the intra-EU jurisdictional objection, but it was raised by the Commission. The Tribunal noted that normally non-disputing parties are not able to object to jurisdiction but this did not prevent the Commission from raising the issue. The Tribunal further stated that the issue should be reviewed ex officio.43 Both, the allowing of the Commission’s amicus curiae and the need to assess the issue ex officio could be seen as reflecting a level of acknowledgement towards EU law and the possible treaty conflict between the EU Treaties and the investment treaties. Similarly, the Tribunal in PL Holdings v Poland noted that it could have rejected the intra-EU objection as it was submitted by the Respondent too late in the proceedings, but decided to address the issue. The Tribunal argued that the issue ‘implicates important sovereign assertions not only by Poland but also by the European Union’, and thus considered it necessary to address it.44
Perhaps a more inventive solution to the intra-EU jurisdictional challenge was seen in Tomasz Częścik and Robert Aleksandrowicz v Cyprus, in an award rendered prior to the CJEU’s judgment in Achmea, where the Tribunal acknowledged that it might have to assess its jurisdiction under the BIT in the light of EU law if the CJEU decided that the EU Treaties prevailed over the BIT. The Tribunal, however, relied on its kompetenz-kompetenz and held that in the absence of such a ruling by the CJEU it was able to uphold ‘prima facie jurisdiction’ in order for it to be able to decide the case on the merits and to see whether the conflict between the BIT and EU law materialised in an award favourable to the investor.45 In other words, as the Tribunal decided the merits in favour of the State, it considered it unnecessary to further address the intra-EU objection. Although the Tribunal did not address the effect of EU law on its jurisdiction, it should be noted here, for the purposes of the current article, that the Tribunal was ready to acknowledge the authority of the CJEU’s judgment on the matter.
Altogether, the case law shows that the intra-EU investment tribunals do not generally tend to ignore EU law, but rather they have engaged in different levels of assessment of it. Rather than merely denying the effect of EU law, it has been decisive that the tribunals have not seen any conflict between EU law and the intra-EU investment treaties under international law.46 Conversely, as will be discussed below, some arguments have also been raised according to which EU law is not relevant to the jurisdiction of the tribunals.
2.3. Rejection of the relevance of EU law for the jurisdiction of the tribunals
Perhaps the most straightforward rationale behind the rejection of the intra-EU jurisdictional objections was summarised by the Tribunal in PL Holdings. It stated that it ‘determines its jurisdiction solely on the basis of the instrument that purports to found its jurisdiction and not on the norms of an entirely different legal order’.47 Similarly, the Tribunal in Eureko argued that it ‘cannot derive any part of its jurisdiction or authority from EU law as such: its jurisdiction is derived from the consent of the Parties to the dispute, in accordance with the BIT and German law’.48 Furthermore, the Tribunal in Strabag v Poland concluded that EU law could be relevant for its jurisdiction only if it formed ‘a part of international public order or such international law principles which meet the fundamental requirements of justice in international trade’.49 Thus, the tribunals tend to emphasise that they are by nature international dispute settlement bodies by nature and they operate under international law and international legal instruments.50 As such, they are not based on EU law nor derive their existence from the EU legal order. The tribunals have also argued that the implicit effect of a BIT becoming an intra-EU BIT, that is, the accession of a State to the EU, on their jurisdiction must be assessed in the light of international law and the VCLT.51
Similar logic can be seen in the rejection of the existence of an ‘implicit disconnection clause’ in the ECT. A disconnection clause would render the ECT inapplicable in the relations between EU Member States, and it has been argued by the Commission and some EU Member States parties to intra-EU ECT cases that such a clause is implied in the treaty due to EU law. In Charanne, for example, the Tribunal argued that the inclusion of an implicit disconnection clause could only be confirmed through interpretation of the ECT in the light of rules of treaty interpretation under international law. As, according to the Tribunal, the ECT is clear on the lack of a disconnection clause, no recourse to additional means of treaty interpretation was necessary.52 In other words, the Tribunal did not accept the interpretation of EU law to be able to entail the result that there was an implicit disconnection clause in the ECT, nor was there a need to interpret the ECT in the light of EU law.
2.4. Assessment of the CJEU’s Achmea judgment
Maintaining that their jurisdiction is based on international treaties unaffected by EU law, and either rejecting the relevance of EU law or conducting different levels of assessment of it, the tribunals never accepted ‘the intra-EU objection’ to their jurisdiction in the intra-EU cases prior to the judgment of the CJEU in Achmea. 53 In Achmea, the CJEU ruled that dispute settlement provisions in intra-EU BITs, ‘such as’ in the Netherlands–Slovakia BIT in question, are precluded under EU law as they are incompatible with Articles 267 and 344 TFEU.54 While the aftermath of the judgment has led to the Agreement terminating the intra-EU BITs,55 the scope of the judgment remains uncertain, especially in relation to the ECT.56 Furthermore, the effect of Achmea on the ongoing intra-EU arbitrations or on the validity of the consents to arbitrate given in the intra-EU BITs has not been clear.
Unsurprisingly, the arbitral tribunals have tended to interpret Achmea in a restrictive manner and denied its effect on their jurisdiction. Gáspár-Szilágyi and Usynin have divided the reception of the judgment by the intra-EU tribunals into two groups: (i) some tribunals have engaged in the assessment of the effect of the judgment on their jurisdiction but rejected it; and (ii) some tribunals have rejected Achmea based on procedural arguments.57 As the latter group has little importance for the current topic, the following will concentrate on the assessment of the arguments falling in the former group. Out of the cases engaging in the assessment of the relevance of Achmea, the Tribunal in Watkins Holdings v Spain has recognised three different lines of argumentation: (i) there is no incompatibility between EU law and international investment law; (ii) Achmea only applies to bilateral treaties, and thus not to the ECT; and (iii) there is no conflict between the EU Treaties and the investment treaties, and they exist and should be applied in parallel.58
In the ECT cases, the rejection of the applicability of Achmea has been argued to have reached the status of jurisprudence constante among the tribunals.59 Arguably, this should not be a surprise, bearing in mind that even the EU Member States have not been unanimous as to the effect of Achmea on the intra-EU applicability of the ECT.60 In Masdar v Spain, the Tribunal rejected the relevance of Achmea simply by arguing that it does not apply to multilateral treaties and that it is silent on the ECT.61 The Tribunal in RWE v Spain, in turn, argued that if EU law were to deprive it of its jurisdiction, it would have to be explicitly provided in the text of the ECT, which is the instrument constituting its jurisdiction. This was followed by an analysis of the wording of the ECT interpreted in the light of the VCLT – an analysis similar to the ones conducted by several tribunals prior to Achmea, as discussed above. The Tribunal concluded by rejecting the intra-EU argument and emphasising the principle of pacta sunt servanda. It rejected the relevance of Achmea by stating that the EU is a party to the ECT but not to the BITs, and that the dispute does not require the determination of EU law issues.62 The Tribunal in InfraRed v Spain, in turn, rejected the relevance of Achmea as it is ‘enacted solely within the legal sphere of the European Union and, as such, has not met with the agreement of all ECT Contracting Parties (including non-EU member states Contracting Parties) nor can [it] be qualified as a practice establishing such an agreement between all ECT Contracting Parties’. The Tribunal, however, engaged in an assessment of the conflict between the ECT and EU law.63 Thus, besides the rejection of the relevance of Achmea to the ECT cases, the tribunals have generally continued to assess the intra-EU objections in light of the investment treaty and international law, as was done before Achmea.64
Assessing the concept of consent to arbitration, the Tribunal in RREEF v Spain noted that a judgment of a court is not able to undo a consent already given.65 Similarly, the Tribunal in Micula II v Romania emphasised that its jurisdiction is based on international treaties, and even if Achmea has changed the domestic law of Sweden, it does not affect the consent given in the international treaties.66 This argumentation is understandable, bearing in mind that in practice Achmea means that an international court of one international regime (the CJEU) tells international dispute settlement bodies of another international regime (the tribunals) that they do not have jurisdiction due to internal rules of the former regime.67 However, one should recall the nature of the Achmea judgment: as it was a preliminary ruling on the interpretation of EU law, by definition it merely clarified the content of the law that already existed. Thus, Achmea could, in principle, have stronger effect on the jurisdiction of the tribunals in a treaty conflict analysis – in the event the tribunals were to find the subject matter of the treaties to be the same, as discussed above in Section 2.2.
Another specific jurisdictional objection which has gained importance, especially after Achmea, is the argument that the intra-EU awards are unenforceable under EU law as they are incompatible with the legal order and public policy of the EU and the Member States. This argument has, however, been clearly rejected by the tribunals.68 Some tribunals have made a distinction between the questions of jurisdiction and the enforceability of awards, and argued that the fact that the award could not be enforced does not imply that the tribunal has no jurisdiction.69 In PV Investors v Spain, the Tribunal stated that it ‘must rule on the basis of the ECT and not on projections of hypothetical consequences of the award’.70 Furthermore, the ad hoc Annulment Committee in Antin v Spain, discussing a possible stay of enforcement of the award, stated:
[W]hile the Committee appreciates the difficulties faced by the Applicant in resolving its legal quagmire, it is reluctant to find that this constitutes a prejudice requiring the granting of a stay. In the Committee’s mind, it would be unfair to the Claimants if the Award was stayed due to a legal conundrum of the Applicant’s own making. Insofar as the Applicant willingly chose to undertake international obligations that may conflict with each other, it cannot thereafter complain of prejudice once these conflicts arise.
Thus, it seems that, although the hardship of the States in relation to complying with an award may be recognised, it is not accepted that the incompatibility of the award with EU law would lead to the invalidity of the jurisdiction of the tribunals.72
Some tribunals have also criticised Achmea. The judgment has been seen as inconclusive and leaving its actual effect uncertain, as well as employing an ambiguous line of interpretation. For example, the Tribunal in 9REN Holdings v Spain stated that ‘there is much to understand’ in the ‘truncated’ judgment of the CJEU.73 The Tribunal in Hydro Energy v Spain, in turn, argued that it is unclear whether the ‘preclusion’ of BITs under EU law means that the BITs are invalid and inapplicable, or that there is an obligation of the Member States to modify them. Furthermore, the Tribunal argued that ‘it is hard to read the Achmea ruling as a normal case of treaty interpretation’ and that ‘[t]he Achmea ruling is a decision on the constitutional order of the EU in support of the policy of European integration, rather than an orthodox application of the rules of treaty interpretation’.74 Tribunals have indeed emphasised that Achmea – although being authoritative interpretation of EU law – remains silent about the treaty conflict in terms of international law.75 This has led some tribunals to reject the relevance of the judgment to their jurisdiction, while perhaps still respecting the status and authority of the ruling under EU law. For example, the Tribunal in Hydro Energy stated:
As such the ruling of the CJEU is entitled to the greatest respect from an international arbitral tribunal. But such a tribunal is not in any sense bound by the ruling. Nor, consequently, can the Tribunal find that, on any normal basis of interpretation under customary international law codified in the VCLT, the dispute resolution provisions of the ECT are incompatible with Articles 267 and 344 TFEU.
Thus, returning to the logic that the tribunals operate primarily under international law, a judgment by the CJEU, which assesses the issue merely from the point of view of EU law, is generally not accepted to have authority as being binding on the tribunals and affecting their jurisdiction.
2.5. Recognition of the impact of EU law on the merits: the case of State aid
A separate issue to the use of EU law to object to the tribunals’ jurisdiction is the use of EU law in assessing the merits of the case and determining whether there has been a breach of the investment treaty by a State. A good example of this are cases concerning State aid law. The Tribunal in AES v Hungary, for example, took the EU’s competition law regime into consideration as a fact in the assessment of the merits of the case but noted that States are not allowed to breach their obligations under international law due to their domestic law.77 This argument flows from Article 27 VCLT, according to which a ‘party may not invoke the provisions of its internal law as justification for its failure to perform a treaty’.78 In contrast, the Tribunal in AES stated that the obligation to act under EU law may be taken into consideration in the assessment of ‘the “rationality”, “reasonableness”, “arbitrariness” and “transparency”’ of the disputed measure. However, the Tribunal emphasised that EU law should be taken into consideration merely as a fact. Hungary pleaded that one of the reasons for it taking the disputed measure was the fact that the EU Commission was conducting State aid investigations and there was likely to be an obligation on Hungary to recover State aid. The Tribunal argued that as the Commission had not yet rendered its decision, and thus Hungary had no obligation to act, the measures taken by Hungary were not motivated by ‘a rational public policy’ objective.79 This would imply that had there been a binding EU law obligation – that is, the Commission’s Decision – EU law might have been accepted as a justification for the measure.
Similar argumentation to that in AES can be seen in relation to the Spanish renewables cases, where the question has risen whether the awards constitute illegal State aid under EU law. The Tribunal in Charanne rejected the relevance of the possibility of the award constituting illegal State aid for its jurisdiction. It noted, however, that the State aid issue ‘would be a matter of public policy which the Arbitral Tribunal shall take into account when deciding the dispute on the merits’ but emphasised that no decision had yet been rendered by the Commission.80
Several tribunals have decided cases also after the Commission rendered a Decision in 2017 regarding the Spanish renewable energy support scheme, according to which any compensation awarded for the investors based on the modification of the original State aid scheme would constitute State aid under EU law. As the tribunals do not have the competence to determine the legality of such aid under EU law, the awards would have to be notified to the Commission under Article 108 TFEU and are subject to a standstill obligation while the Commission assesses their legality under EU law.81 There is thus no specific decision yet rendered by the Commission which would declare the awards to constitute illegal State aid. The Tribunal in PV Investors argued that the 2017 Decision was not relevant in its case and that the ‘Tribunal must rule on the basis of the ECT and not on projections of hypothetical consequences of the award’.82 The relevance of the Decision was rejected also by the Tribunal in RWE v Spain on the basis that the Decision did not concern the State aid regime being contested in the case, and that there had been no evidence implying that the disputed measure was taken by Spain due to the EU’s State aid regime.83 Thus, even if some of the tribunals have acknowledged that a State aid decision could be relevant in determining the case, they have emphasised the lack of a specific decision on the legality of the aid regime under dispute in the specific case.84
In Electrabel the State aid had already been confirmed illegal under EU law.85 The Tribunal, in turn, seemed ready to recognise the relevance of such decision:
Where Hungary is required to act in compliance with a legally binding decision of an EU institution, recognized as such under the ECT, it cannot (by itself) entail international responsibility for Hungary. Under international law, Hungary can be responsible only for its own wrongful acts. The Tribunal considers that it would be absurd if Hungary could be liable under the ECT for doing precisely that which it was ordered to do by a supranational authority whose decisions the ECT itself recognises as legally binding on Hungary.
Thus, there appears to be a division between two types of cases on the recognition of the relevance of EU State aid law: the cases where there is no binding decision on the illegality of the State aid, and the cases where the Commission has adopted such a decision. This suggests that the clarity of the EU law rules plays an important role in order for EU law to be taken into consideration in the determination of the merits of the cases.
3. Mitigating conflicts through harmonious interpretation?
3.1. The (lack of) interpretation of international investment law in the light of EU law
Although the intra-EU BIT saga is coming to an end with the termination of the BITs,87 the conflict between EU law and international investment law is far from being over. Essentially, the fate of the intra-EU applicability of the ECT remains unresolved, and numerous cases are pending under the ECT, but also under the BITs. Thus, the following will turn to discuss whether the conflict could be alleviated through interpretation of the respective treaties, also paying closer attention to the other regime.
The tribunals have been criticised for not taking enough account of EU law. According to Koutrakos, instead of showing the acknowledgement necessary in ‘the parallel adjudication of interlocking sets of transnational law’, some arbitral tribunals have displayed a ‘distinct lack of engagement with the practical realities’ of the conflict between EU law and international investment law. For example, Micula v Romania could be criticised for not properly assessing the consequences of the accession of Romania to the EU and the applicability of the EU State aid rules, and thus not taking into consideration the compatibility of the award with EU law.88 Furthermore, it has been claimed that the arbitral tribunals are ill-equipped to determine the core issues of EU law, and there might be a need for the tribunals to show ‘self-restraint and deference’ in relation to the questions of compatibility of investment treaties with EU law. According to Dimopoulos, the determination of the core issues of EU law should be left to the EU Courts, even when this might be contrary to the nature of investment arbitration.89 This idea is in line with the view taken by the International Tribunal for the Law of the Sea (ITLOS) in the Mox Plant case, where the Tribunal noted that the dispute between Ireland and the UK included questions on the division of competences in the EU and the exclusivity of the CJEU’s jurisdiction. Taking into consideration judicial comity and mutual respect between judicial institutions, the Tribunal maintained that these questions should be decided within the EU in order to avoid the risk of competing decisions, and suspended the proceedings.90 It should be noted, however, that the United Nations Convention on the Law of the Sea (UNCLOS) explicitly supports this conclusion,91 whereas the BITs do not contain such provisions.
Nevertheless, the possibility for harmonious interpretation has been raised in several intra-EU arbitrations. A basis for such an interpretation could be found from Article 31(3)(c) VCLT, according to which ‘any relevant rules of international law applicable in the relations between the parties’ ‘shall be taken into account, together with the context’ when interpreting a treaty.92 In Oostergetel v Slovakia, the Tribunal noted that it ‘will seek to interpret both the BIT and applicable EU law in a manner that minimises conflict and enhances consistency’.93 Similarly, the possibility of EU law affecting the way the ECT is interpreted has been raised. In RREEF v Spain, the Tribunal noted that ‘EU Law reflects the common understanding of 28 countries in such important matters as legitimate expectations’ and that the Tribunal cannot ‘disregard’ that.94
The Tribunal in Electrabel also argued that there is no ‘general principle of international law compelling the harmonious interpretation of different treaties’.95 It emphasised, however, that the situation was special in the intra-EU ECT case, when taking into consideration the close involvement of the EU and its Member States in the ECT as well as the origins of the Treaty. Thus, according to the Tribunal, the interpretation of the ECT should be ‘in harmony with EU law’. While the Tribunal did not see the need for harmonisation of the dispute settlement provisions of the ECT with EU law as it did not consider there to be any inconsistency with EU law, it did engage in the assessment of what the (in the Tribunal’s view hypothetical) inconsistency between the two regimes would entail in practice. It noted that there is no obligation to interpret conflicting treaties harmoniously and conducted the assessment in the light of the international rules on treaty conflict.96 Similarly, the Tribunal in Eskosol argued that, as it is a tribunal acting under international law, it must resolve a treaty conflict based on the rules of international law and cannot replace those rules with ‘a very different conflicts analysis derived from EU law’.97 This suggests that the understanding of the need for harmonious interpretation reaches its limits in the question of jurisdiction.98 Similar argumentation is put forward in the International Law Commission’s Report on Fragmentation of International Law, according to which harmonisation might not be a suitable option for the resolution of a treaty conflict, if clear-cut rights or obligations laid down in the treaties are involved in the bargain.99
Conversely, the reach of harmonious interpretation, or perhaps the idea of comity, to the jurisdiction of the tribunals has been supported by some commentators. In relation to Achmea, it has been argued that the tribunals should either decline their jurisdiction or, alternatively, dismiss the cases as inadmissible. In practice this would enhance ‘the effective administration of justice’.100 The idea is attractive when taking into consideration the procedural turmoil and uncertainty the parties face in the enforcement stage of the arbitration, if the tribunal renders an award based on an intra-EU investment treaty. Whether the tribunals would follow such a suggestion, and not continue to emphasise their mandate under international law, is another question.
An issue closely linked to the harmonious interpretation has been the possibility of the arbitral tribunals to refer questions of EU law to the CJEU for preliminary ruling, which was raised in the early intra-EU cases.101 The tribunals have, however, rejected such requests.102 Although there is no requirement and, even more importantly, no possibility for the investment arbitral tribunals to request a preliminary ruling from the CJEU on the interpretation of EU law,103 another question is whether they should wait for such a ruling if a relevant case happens to be pending in the CJEU at the time of the arbitral proceedings taking place. There is no legally binding obligation under international law to do so, and the nature of arbitration as an independent, case-specific method of dispute settlement speaks against that. Furthermore, such a practice could potentially prolong the already rather lengthy arbitral proceedings. The Tribunal in PL Holdings found, in fact, that the pendency of Achmea in the CJEU did not remove the obligation of the Tribunal to decide the case in front of it, nor did it impose an obligation to address the possible consequences of the CJEU’s ruling.104 In contrast, such occasions could be seen as a possibility for comity, and as a way to avoid further conflicts between the two legal orders, as was recognised by the Tribunal in Eureko. Although it did not consider it necessary to wait for the conclusion of a pending infringement case as the facts of the case were not similar enough, it argued that it wished ‘to organise its proceedings with full regard for considerations of mutual respect and comity as regards other courts and institutions’. Furthermore, the Tribunal stated that, were the facts of the cases to become ‘so close as to be a cause of procedural unfairness or serious inefficiency, the Tribunal will reconsider the question of suspension’.105
A similar question has arisen also in relation to the State aid issues in Spain’s renewable energy cases: should the tribunals wait for the Commission to render its decision on the compatibility of the support scheme with EU law? For example in RWE, the Commission stated that alternatively to declining its jurisdiction due to intra-EU objection ‘should the Tribunal find that it has jurisdiction, the EC invites the Tribunal to stay the proceeding until the EC has issued its decision regarding the Respondent’s State aid under the Special Regime’.106 As noted above, the tribunals have tended not to accept the mere possibility of a future breach of EU law to be relevant to their decisions. Arguably, this could be a situation where a certain level of comity could mitigate the conflicts. If the tribunals were to wait for a binding decision from the Commission, the risk of their rendering an award in conflict with EU State aid law would be lower.107 The rendering of awards regardless of their compatibility with the EU State aid law practically moves the conflict to the enforcement of the award. This, in turn, not only prolongs the dispute, but also further aggravates the conflict.
Regarding the intra-EU BITs and the issue of the validity of the jurisdiction of the tribunals, one could say that the ship has already sailed, and ‘no amount of balancing of conflicting interests by intra-EU tribunals will gain constitutional acceptance by the Court of Justice’.108 However, the issue of intra-EU applicability of the ECT remains unresolved, and there is a remarkable number of cases pending at the time of writing. Thus, the question of interpretation of the ECT in harmony with EU law remains relevant. Conversely, the issue might be more complicated in the context of multilateral treaties, such as the ECT. The Tribunal in Vattenfall emphasised the need for a uniform interpretation of the treaty text in light of the principle of pacta sunt servanda and argued that the interpretation of the ECT should not differ depending on which States are parties to specific disputes. Thus, the legal obligations – such as EU membership – into which each State has entered outside the ECT regime should not affect the interpretation of the provisions of the ECT in the light of international law.109 Thus, even if harmonious interpretation might be a useful tool to ease the tensions between the two fields to a certain extent, it seems rather unlikely that the tribunals under the ECT would fully engage in it.
3.2. The (lack of) interpretation of EU law in the light of international law
The EU is committed to ‘the strict observance and the development of international law’, according to Article 3(5) TEU. Similarly, the CJEU has stated that the EU ‘must respect international law in the exercise of its powers’.110 Thus, it can be argued that the principle of autonomy of EU law cannot operate in isolation from the acknowledgement of international law.111 It has, however, been claimed that the CJEU sees the autonomy of the EU legal order as ‘an absolute value’ and not as something that allows balancing of interests.112 Similarly, the CJEU’s insistence on protecting the autonomy of EU law and the adjudication system of the EU legal order is alleged to be ‘a little selfish’, and not doing much to promote the development of international law.113 According to Klabbers, the mitigation of treaty conflicts by claiming the superiority of EU law leads to ‘the Court’s policy [being] that of the ostrich, aiming to avoid problems by sticking its head in the sand and pretending they do not exist’.114
A point of criticism in the conflict between investment law and EU law has, in fact, been the CJEU’s point of view being strictly limited to EU law. As noted above, some intra-EU tribunals have criticised Achmea for neglecting the analysis of international law implications for the case and merely assessing the issue from the point of view of EU law.115 Thus, a question arises whether, for the sake of justifying its decisions also to the wider world, the CJEU should take better account of international law in matters dealing with the autonomy of EU law.116 As Odermatt argues, the ‘[a]utonomy is currently defined primarily in terms of exclusion, even isolation, and not in terms of engagement and integration’.117 Arguably, it could also be asked whether there would be a place for a certain level of ‘proportionality’ of the claim of the autonomy of EU law. Especially dealing with the intra-EU BIT issue from the point of view of EU law has been characterised by a lack of discussion on whether the co-existence of the two regimes could have been established, and whether the termination of the treaties in fact is the only way to resolve the conflict.118
Another possibility for the international fora to be taken better into consideration by the EU legal order could be the clarification of internal rules. As discussed above in relation to the Spanish renewables cases, the Commission has not rendered a decision on the illegality of the support scheme under dispute but rather has subjected the awards to its State aid assessment. This has been challenged by Abaco Energy SA v Commission in the General Court of the EU with the argument that it enables the Commission to control the awards, as well as to move the intra-EU arbitral cases under the control of EU law.119 It is rather easy to see that a binding decision on the compatibility of the scheme with EU law – instead of a case-by-case assessment of the awards – would lower the level of uncertainty that now looms over the cases.120
Perhaps a solution could be a certain degree of flexibility from both sides. Based on the Bosphorus case,121 Tietje and Wackernagel have suggested that the claims of superiority of both EU law and international law should be understood ‘as a principle rather than a rule’, which would enable the other legal order to be taken into consideration as far as ‘factually and legally possible’.122 Thus, the flexibility could be used as long as it does not lead to a contra legem situation in either of the legal orders. However, both regimes appear to hold tight to certain claims – be it the international law basis of the tribunals’ jurisdiction or the autonomy of EU law – and it seems unlikely that harmonious interpretation, comity, or other ways of consolidation of the conflict would entirely resolve the issues.
4. Implications for extra-EU cases
Although the most significant confrontations in the interplay between EU law and international investment arbitration have taken place in an intra-EU setting, the questions of the relevance of EU law cannot be ruled out in extra-EU cases either. Similar issues of the effect of EU law on the jurisdiction of the tribunals should not arise123 but some examples can already be found where EU law has been relevant in the substantive assessment of cases resolved under BITs between an EU Member State and a third State.
The status and nature of EU law in an extra-EU setting is clearer compared to the intra-EU setting discussed above. Whereas the tribunals in intra-EU cases tend to see EU law as having a dual nature as both international and domestic law, and EU law as forming a part of ‘the applicable international rules in force between the contracting parties’,124 this should logically not be the case in an extra-EU setting. In the latter, EU law should be understood more purely as domestic law of Member States – albeit that it originates from a supranational source. As such, EU law may become relevant in the assessment of the merits of the extra-EU cases.
The effect of EU law on the assessment of a breach of the substantive standards was discussed by the Tribunal in Cargill v Poland. The Tribunal noted that EU law may be taken into consideration as ‘municipal law’ as a fact, alongside the domestic law of Poland, and that EU law may be relevant in the assessment of whether the State’s international responsibility materialises in the case. It was concluded that EU law imposed no obligation on Poland to take the disputed measures before its accession to the EU (the facts of the case pre-dated Poland’s accession), and thus EU law could not be used as a justification for the measures.125 EU law was relevant also in Maffezini v Spain, where an Argentine investor argued, among other claims, that an allegedly public entity in Spain had pressured it to make an investment before final clearance of an Environmental Impact Assessment (EIA) had been given by the authorities. The Tribunal noted that EU law provides ‘strict requirements’ for EIA, and according to the relevant Directive, an EIA has to be concluded before the start of the project. The Tribunal concluded that the investor’s claim should be rejected.126 The Tribunal did not engage in detailed analysis of the content of EU law, but merely quoted the relevant provision of the EU Directive.
Furthermore, according to the Commission Report on the application of ‘the Grandfathering Regulation’,127 the Commission has intervened or tried to intervene in three cases under the Member States’ BITs with third countries since the initiation of the Grandfathering mechanism. These interventions were made in order ‘to clarify the EU legal framework and procedures in connection with the facts of the disputes’ and related to State aid rules, EU single resolution mechanism, and the implementation of the EU energy policy framework.128 This confirms that there is a possibility for the extra-EU investment arbitral tribunals to also need to interpret and apply EU law, and the EU has an interest in this securing that the uniform interpretation and application adopted within the EU legal order is followed.
As noted above, the CJEU stated in Opinion 1/17 that the possibility of arbitral tribunals to take EU law into consideration ‘as a fact’ as domestic law of either the Member States or the EU is not in conflict with EU law.129 The requirements of EU law to be treated as a fact, as well as a limitation of the effect of the awards within the EU legal order, have been included in the texts of the EU IIAs. For example, Article 8.31(2) CETA provides:
For greater certainty, in determining the consistency of a measure with this Agreement, the Tribunal may consider, as appropriate, the domestic law of a Party as a matter of fact. In doing so, the Tribunal shall follow the prevailing interpretation given to the domestic law by the courts or authorities of that Party and any meaning given to domestic law by the Tribunal shall not be binding upon the courts or the authorities of that Party.
Similar language is not generally included in the ‘older generation’ extra-EU BITs concluded by the Member States with third countries.131 However, new BITs and the amendments of the existing ones should comply with the EU’s new investment policy.132 According to the Commission Report on the application of the Grandfathering Regulation, the inclusion of ‘provisions on the applicable law ensuring preservation of the autonomy’ in investment agreements is part of the EU’s new investment policy, and the new investment policy, in turn, ‘is reflected as much as possible in the Commission’s decisions authorising the conclusion of new bilateral investment agreements of Member States with third countries’.133 Thus, ‘the new generation treaty language’ is likely to become more common in the Member States’ BITs as well.134 The negotiations of BITs are, however, long processes, and the old-generation treaties are likely to stick around for the time being, regardless of the Commission’s encouragement for the Member States to take measures ‘replacing older bilateral investment agreements with new ones reflecting modernised standards in line with the EU’s reformed investment policy’.135
The specific treaty language is already visible also in the EU’s proposal for the revisions of the ECT. The proposal contains similar language to CETA and other EU IIAs on the status of domestic law as a fact, as well as on the interpretation of domestic law. Furthermore, it is suggested that a new article be included in the ECT, according to which an obligation flowing from membership in a Regional Economic Integration Organisation (i.e. the EU) to suspend the granting of subsidies (i.e. State aid) for an investor or to request the investor to reimburse them should not be considered a breach of the investment protection standards in the ECT or entail that the State would have to compensate the investors. Similarly, a breach of an obligation to compensate would not take place if such a suspension or request followed from an order of ‘a competent court, administrative tribunal or other competent authority’. ‘Other competent authority’ is further clarified to mean the Commission or a court of a Member State.136 This appears to be a direct answer to the uncertainty of the State aid cases discussed above, and it appears that the EU is increasingly taking measures to tackle the conflict between investment arbitration and the autonomy of EU law by including specific provisions in the investment treaties.
Even though the assessment of national law as a fact, or the limited effect of the interpretation by arbitral tribunals of national law provisions are not new ideas to the investment arbitration practice (as was seen above in Section 2), the inclusion of specific treaty language in the treaties arguably enhances legal certainty. Furthermore, it has been claimed that the acceptance of such treaty language in the EU IIAs by third countries shows that external actors are open to recognising the specific features of the EU’s legal order and the EU as an international actor.137 Thus, the Commission’s encouragement for the Member States to include the ‘new generation treaty language’ also in their older BITs is favourable in order to ensure the consistency of the arbitral tribunals’ understanding of the role and status of EU law and to mitigate the risk of arbitral practice running contrary to the principle of autonomy of EU law. In the meantime, it remains to be seen whether any issues will arise in the relationship between EU law and the older-generation extra-EU BITs, and how such issues would be handled by the tribunals.
The fact that the intra-EU arbitral tribunals have often engaged in rather detailed assessment of the treaty conflict between the intra-EU investment treaties and EU law shows that there is a certain level of recognition of the specific nature of the EU and its legal order in investment arbitration. The tribunals have, however, maintained that they are primarily international tribunals, and prioritised the point of view of international law, while rejecting the claim that EU law could render their jurisdiction invalid. A central aspect of this has also been the fact that the tribunals have not recognised that there is a genuine conflict in the lines of international treaty law between the investment treaties and EU law.138 Whereas the tribunals could be criticised for not showing sufficient recognition of EU law, from the viewpoint of international law the demand from the EU for the jurisdiction to be denied appears to be rooted in the internal structures of the EU legal order. A comparison of this could, arguably, be drawn with the primacy of EU law in relation to national laws of the Member States. However, as international law operates primarily in a horizontal manner, there is little ground for claiming that EU law would prevail over other fields of international law if such a conclusion cannot be drawn from established rules of international law or treaty conflict.139
Arguably, for the deprivation of jurisdiction to be acceptable also in the eyes of the tribunals, the specific features of EU law should be regulated in a manner that is valid also under international law and not only under the auspices of EU law.140 In addition to the adoption of the agreement terminating the intra-EU BITs,141 this is something that should be borne in mind in relation to the ECT as well. Indeed, it has been argued that as there are currently no general international rules covering the practice of international dispute settlement where a party is ‘a composite legal subject’, such as the EU, the specific rules required by such participation should be provided in the relevant legal instruments.142 Furthermore, the need for clear rules arises also when EU law is relevant in the determination of the merits of the disputes. Whereas it is generally accepted that EU law can be taken into consideration as domestic law, the relevant EU law rules need to be clear.
Arguably, the issue has its roots in the EU being, due to its constitutional structure, somewhat of a mismatch with international law. As international law traditionally deals with States and international organisations, the EU as a ‘hybrid’ in between143 does not always fit the picture seamlessly. Whereas harmonious interpretation or a practice of comity could alleviate the conflicts to some extent, there is a clear need for the relationship between EU law and international (investment) law to be regulated through clear and binding instruments of international law.