Investment Tribunals Are Too Quick to Establish the Existence of Issue and Cause of Action Estoppel in International (Investment) Law

Alexandros-Cătălin Bakos[1]

There is no denying that there is a serious backlash against investment arbitration at the moment. The signs are everywhere: from the latest discussions occurring within UNCITRAL’s Working Group III to the more recent practice of states (see the 22 European Union Member States’ declaration concerning the termination of their intra-EU Bilateral Investment Treaties); the latest ‘battlefront’ seems to be the Energy Charter Treaty, where the investment tribunals seized of disputes on the basis of this treaty consider it immune from the effects of the Achmea decision. The causes for this backlash are manifold. For present purposes, however, I would like to focus my attention on only one of the causes: incorrect decisions. And I would like to go even further and look at a very specific example of incorrect decisions: the application of the principle of estoppel by investment tribunals. I will focus exclusively on the procedural aspect of estoppel, as a bar to a claim. This seems to be its main, although not its only (para. 831), function – at least in international investment law.

Some background information on estoppel

Generally, estoppel is a very strong mechanism which has a preclusive effect against a party contradicting itself if another party has relied (usually to the latter’s detriment) on the initial position of the former (para. 231). Essentially, the party which contradicts itself is prevented from averring the contradictory fact (the subsequent one). ‘[W]hat is relevant for estoppel is that there has been a declaration, representation, or conduct which has in fact induced reasonable reliance by a third party, which means that the State, even if only implicitly, has committed not to change its course’ (idem, para. 246). Furthermore, the element which induces reliance must be unambiguous (paras. 8.46-8.47). Other tribunals refer to the fact that representations must be ‘clear and consistent’ (for example, the Chagos Marine Protected Area Arbitration, para. 438).

In international law, the application of estoppel dates back to the days of the Permanent Court of International Justice: for example, in the Legal Status of Eastern Greenland case, Norway was precluded from asserting sovereignty over Greenland, as the former had expressly recognized the latter as part of Denmark. This form of estoppel, however, seems to heavily overlap with vaguer principles – including the principle of good faith (para. 483).

There are voices in international law which argue that estoppel as such exists in a single form in international law and not in its various iterations found in the domestic common law systems (para. 436). This view, however, is not shared by all international law practitioners. Whether due to fragmentation of international law or not, this divergence becomes obvious once one analyzes arbitral practice. One example of how arbitral tribunals have looked at estoppel in its specific iterations concerns procedural aspects. There, estoppel acts as a more specific and technical mechanism designed to prevent an already litigated claim from being pursued again (similar to res judicata, although with a few important differences which will be mentioned below). The important branches of estoppel which may preclude a claim from being relitigated are: cause of action estoppel;[2] and issue (or collateral) estoppel.[3] It is important to mention that both these doctrines ‘prevent the parties from re-litigating a question that has been determined by a Court of competent jurisdiction, between the same parties or their privies, in a previous action. Once those elements have been made out, and unless there are special circumstances, the parties are precluded from raising the issues. [footnote omitted] The special circumstances which would permit the issue to be raised again include the discovery of further material relevant to issues in the first set of proceedings [footnote omitted] or fraud’.[4] The essential difference between the two doctrines, according to Griffith and Seif, is that cause of action estoppel concerns the claim itself which is precluded, whereas issue estoppel prevents relitigation of a point of law or of fact already decided by a tribunal.[5] Wilken QC and Ghaly point out that the difference is one of specificity.[6] According to them, ‘issue estoppel bites on the facts and issues required to establish the cause of action whereas cause of action estoppel looks only at the cause of action’.[7] Sheppard equates ‘cause of action’ with ‘claim’.[8]

A very important point of difference between estoppel – in both its iterations – and res judicata is that the latter requires (at least traditionally, as Judge Anzilotti mentioned in his dissenting opinion to the Factory at Chorzów case) a three-element identity between the concerned claims (the same person, the same claim and the same legal grounds); also known as the ‘three-element test’. Moreover, estoppel extends to the privies of the relevant parties, while res judicata – if interpreted strictly – does not.[9] Without going into the details of how the three elements of res judicata have been interpreted, especially in investment arbitration (as this is another subject for another date), it can be reasonably stated that estoppel is a stronger tool (than res judicata) in the arsenal of investment tribunals which can be used to prevent abusive re-litigation. The problem, however, is that the existence of such an instrument in international law is not clearly evident and tribunals seem to have taken its existence for granted.

The problems with the investment tribunals’ application of estoppel

Although not a general principle of law,[10] some arbitral tribunals seem to have applied estoppel as such. As will be seen below, however, there is at best inconclusive evidence as to the existence of a general principle of estoppel and at worst clear attempts to disregard this non-existence and apply a principle out of nothing.

At the same time, there are arbitral tribunals which may suggest or clearly determine that estoppel is a principle of law,[11] although this is usually not explained clearly and the reasoning is incomplete. As such, one is left wondering how did the tribunal uncover such a principle and whether it really exists.

For example, the Petrobart tribunal mentioned that ‘while the doctrine of collateral estoppel seems to have primarily developed in American law, other legal systems have similar rules which in some circumstances preclude examination of an issue which could have been raised, but was not raised, in previous proceedings. A doctrine of estoppel is also recognised in public international law’ (at pp. 66-67).

The tribunal, however, was unclear whether this amounted to a principle of law or not. The fact that there exist rules which establish preclusion of issues which could have been raised but were not raised and that these rules occur outside of the American legal system, as well, does not transform estoppel into a principle of law. At the same time, the tribunal did not mention in what form is estoppel recognised in public international law. It may have suggested that this would be applied as a principle, but it stopped short of fully clarifying whether such a principle indeed exists. The alternative may have been the customary law nature of estoppel, but the tribunal neither identified the underlying state practice and opinio juris nor referred to awards/ judgements in which such a custom was established. In the end, the claim preclusion argument was anyway rejected, since – among others – there was no identity between the legal grounds relied on in the relevant proceedings (at pp. 67-68).

Another example is RSM v. Grenada. There, the tribunal explicitly endorsed collateral estoppel as a general principle of law (para. 7.1.2). The tribunal noted ‘that the doctrine of collateral estoppel is now well established as a general principle of law applicable in the international courts and tribunals such as this one. [footnote omitted] (ibid.). However, it did not come to this conclusion itself, but rather relied on other tribunals’ conclusions.[12] What is surprising after looking at the cited cases is that neither of them clearly endorses estoppel as a principle of law.

For example, the Amco v. Indonesia tribunal referred to res judicata as a principle of law (paras. 26-46). One cannot exclude the possibility of this encapsulating estoppel as well, but such a conclusion is not clear. This lack of clarity is further compounded by the fact that the Amco v. Indonesia tribunal mentioned that ‘it is by no means clear that the basic trend in international law is to accept reasoning, preliminary or incidental determinations as part of what constitutes res judicata’ (idem, para. 32). As issue/collateral estoppel necessarily implies the fact that the reasoning of an award must be considered for this mechanism to arise,[13] the finding of the Amco v. Indonesia tribunal raises serious doubts as to the conclusion that estoppel was part of the principle to which that tribunal referred.

As regards the other relevant case (Southern Pacific Railroad Co. v. United States, which arose before the Supreme Court of the United States) it is true that what the cited tribunal referred to was issue estoppel (pp. 48-49). It mentioned that a general principle existed which mandated ‘that a right, question, or fact distinctly put in issue, and directly determined by a court of competent jurisdiction as a ground of recovery cannot be disputed in a subsequent suit between the same parties or their privies, and, even if the second suit is for a different cause of action, the right, question, or fact once so determined must, as between the same parties or their privies, be taken as conclusively established so long as the judgment in the first suit remains unmodified’ (ibid.). What the tribunal does not mention, however, is whether this general principle is a general principle common to all nations or whether this was a general principle specific only to the common law system.

There are tribunals which even seem to rely on estoppel, although, in reality, they are applying res judicata. This was the case with the Marco Gavazzi and Stefano Gavazzi v. Romania tribunal (paras. 164-166). In the first place, the tribunal analyzed whether an initial decision (which was alleged to preclude the claims before the forum) had ‘conclusive effects on the Parties to the present proceedings under the doctrine of res judicata or issue estoppel’ (idem, para. 164). Subsequently, it went on to mention that ‘under international law, three conditions need to be fulfilled for a decision to have binding effect in later proceedings: namely, that in both instances, the object of the claim, the cause of action, and the parties are identical’ (idem, para. 166). Although it did expressly refer to issue estoppel at one point, the tribunal referred to the conditions which were necessary to be fulfilled in order for res judicata to operate (the three-element test, as mentioned above). Moreover, it conflated issue estoppel with cause of action estoppel. As shown earlier, identity of cause of action is only necessary in the case of cause of action estoppel and not in the case of issue estoppel.

All the above examples demonstrate that estoppel as such is not applicable in investment arbitration (by virtue of international law, at least) and that tribunals seem to ignore this. There is no general principle – as understood by Article 38 (1) (c) of the Statute of the International Court of Justice, as an authoritative reflection of the sources of international law – of estoppel. At least no principle which could cover cause of action or issue estoppel. There is no evidence of a customary rule encapsulating estoppel either.[14] Moreover, not even investment treaties seem to contain this mechanism. For example, the 2012 US Model BIT – selected for being relevant to a common law jurisdiction – does not make any reference to estoppel. Neither does one of the latest UK BITs (the UK-Colombia BIT) contain any reference to estoppel – although it does allow the tribunal to address abuse of process; however, this is different than estoppel.


[1] Editor at avocatnet.ro and Associate Expert at DAVA | Strategic Analysis. This post is based on part of my thesis, submitted for the completion of an LL. M. in Law and Economics at Utrecht University. I would like to express my gratitude to Dr. Yulia Levashova, for her continuous support and for an in-depth and comprehensive feedback. In any case, I take full responsibility for the opinions and they are exclusively mine, not reflecting anyone else’s or any other institution’s.

[2] Audley Sheppard, ‘Chapter 8. Res Judicata and Estoppel’ in Bernardo M. Cremades Sanz-Pastor and Julian D.M. Lew (eds.), Parallel State and Arbitral Procedures in International Arbitration, p. 225 (hereinafter referred to as ‘Sheppard’).

[3] Ibid.

[4] Sean Wilken QC, Karim Ghaly, The Law of Waiver, Variation and Estoppel. Third Edition (Oxford University Press 2012), para. 14.08 (hereinafter referred to as Wilken QC, Ghaly).

[5] Gavan Griffith; Isabella Seif, ‘Chapter 8: Work in Progress: Res Judicata and Issue Estoppel in Investment Arbitration’, in Neil Kaplan and Michael J. Moser (eds), Jurisdiction, Admissibility and Choice of Law in International Arbitration: Liber Amicorum Michael Pryles (Kluwer Law International 2018), p. 124 (hereinafter referred to as ‘Griffith; Seif’).

[6] Wilken QC, Ghaly, para. 14.09.

[7] Ibid..

[8] Sheppard, p. 225.

[9] Griffith; Seif, p. 126.

[10] Charles T. Kotuby Jr. and Luke A. Sobota, General Principles of Law and International Due Process. Principles and Norms Applicable in Transnational Disputes (Oxford University Press 2017), footnote 262, p. 200. Such a conclusion (that estoppel is not a general principle of law) is in accordance with one of the major views in international legal relations as to what constitutes a general principle of law: one ‘which can be derived from a comparison of the various systems of municipal law, and the extraction of such principles as appear to be shared by all, or a majority, of them [emphasis added]’, Hugh Thirlway, The Sources of International Law. Second Edition (Oxford University Press, 2019), p. 108.

[11] Stating that estoppel is a principle of law serves two aims: firstly, the tribunal justifies the application of estoppel by reference to a source of international law (usually, part of the applicable law). Secondly, this gives the tribunal legitimacy, as the tribunal grounds its decision to rely on estoppel on a widely-applicable source of law (whether objectively true or not is not as important).

[12] The cases to which the RSM tribunal referred were mentioned at page 27, footnote 34 of the award: Amco Asia Corporation v Republic of Indonesia, ICSID Case No. ARB/81/1, Decision on Jurisdiction (Resubmitted Case), 10 May 1988, para. 30; Company General of the Orinoco Case, 10 R.I.A.A. 184 (1905); and Southern Pacific Railroad Co. v. United States, 168 U.S. 1 (1897). The second tribunal quoted in turn the third one. As such, I will refer only to the first and third tribunals in the remainder of this part.

[13] Sheppard, p. 234; Griffith, Seif, p. 121.

[14] Christopher Brown, ‘A Comparative and Critical Assessment of Estoppel in International Law’, University of Miami Law Review [Vol. 50:369 1996], pp. 384-385;Pan Kaijun, ‘A Re-Examination of Estoppel in International Jurisprudence’, 16 Chinese Journal of International Law (2017), p. 761.

Schrödinger’s Investment: the EU’s General Court Considers that the Compensation Ordered by the Micula Tribunal is Not a Form of State Aid (Although it Might as Well Have Been)

Alexandros Catalin Bakos, LL.M. Candidate, Utrecht University

In a somewhat fortunate turn of events for the stability (or what is left of it in any case) of the intra-European Union (intra-EU) investment treaty system, the General Court of the European Union (GCEU) has annulled the EU Commission’s decision rendered against Romania for illegal state aid concerning the enforcement of the Micula arbitral award. Although the GCEU’s decision may be good news for the investors themselves, it does nothing to allay fears regarding the future of intra-EU ISDS. In the grand scheme of things, the effects which culminated with the Achmea judgement are still there.

This latest installment in the long-running saga of intra-EU investment treaties and their conflict with the EU legal order does not substantially change the paradigm. In fact, one may argue that it complicates the matters: the only certain conclusion that can be derived from the General Court’s decision is the fact that there can be no conflict between EU State Aid rules and intra-EU Bilateral Investment Treaties (BITs)/awards based on such treaties if the compensation ordered by the tribunal relates to measures which were taken prior to the entry into force of EU law. However, the Court did not analyze what is the situation of compensation which needs to be paid for measures adopted after the entry into force of EU law.

In any case, before continuing with the decision’s analysis, a short recap of the major developments in this situation is in order.

How did we get here? 

Prior to joining the EU, the Romanian state offered the Micula brothers and the companies controlled by them (the investors) certain custom duty exemptions and other tax breaks (the GCEU’s decision, paras. 5-6). Later, in 2004 and 2005, those exemptions and breaks were suddenly repealed, in an effort to ensure compliance with the EU laws on State Aid – which would become effective from 1 January 2007 (the GCEU’s decision, para 12). Because of this, the investors began ICSID arbitration proceedings, challenging the compliance of the measure with the applicable BIT (the 2002 Sweden-Romania BIT). The arbitral tribunal found in the investors’ favour and ordered Romania to pay compensation amounting to approximately €178 million. The court’s finding was based on a violation of the fair and equitable treatment standard. More specifically, on behaviour contrary to the legitimate expectations of the investors. This is of utmost importance, as what was considered to be in breach of the treaty was not the repealing of the exemptions itself, but the manner in which this occurred. The arbitral tribunal expressly found that ‘by repealing the […] incentives prior to 1 April 2009, Romania did not act unreasonably or in bad faith […] [H]owever […] Romania violated the Claimants’ legitimate expectations that those incentives would be available, in substantially the same form, until 1 April 2009. Romania also failed to act transparently by failing to inform the Claimants in a timely manner that the regime would be terminated prior to its stated date of expiration. As a result, the Tribunal finds that Romania failed to “ensure fair and equitable treatment of the investments” of the Claimants in the meaning of Article 2(3) of the BIT’ (para. 872 of the award).

Subsequently, the investors sought the enforcement of the award. However, this proved difficult because the EU Commission intervened and tried to prevent Romania from enforcing the award. The former argued that an enforcement would constitute a form of illegal state aid. After Romania, nonetheless, partially paid the award, the EU Commission officially adopted a decision against the Romanian state for breach of State Aid rules. The Commission’s argument was that this payment would, in essence, favour the investors in the same way in which the exemptions favoured them in the first place. Romania, thus, was under an obligation to stop paying the award and to recover the amount which had been paid so far. This was eventually challenged by the investors before the GCEU and the judgement analyzed here is the European Court’s decision regarding that challenge.

This turn of events determined other courts where enforcement of the award was sought to stay the proceedings until the European Court will have rendered an award concerning the challenge to the EU Commission’s decision on illegal state aid (see here for an example).

What does the GCEU’s decision entail and what does it not entail?

The GCEU found that the compensation rendered by the Micula arbitral award could not be considered illegal state aid, at least as it regards events which took place before Romania’s accession to the EU (para. 109 of the GCEU’s decision).

The essence of the GCEU’s arguments is based on a clear establishment of the temporal nexus to which the arbitral award referred (paras. 71-93 of the GCEU’s decision). To this end, the Court clarified that all the relevant issues (including the events which gave rise to the right to compensation) arose and produced effects before Romania’s accession to the EU (para. 71). In that respect, even if the arbitral tribunal’s award was rendered after EU law became applicable to Romania, it merely ‘retroactively produced definitively acquired effects which it merely ‘stated’ for the past, that is to say, effects which, in part, were already established before accession’ (para. 84 of the GCEU’s decision). Accordingly, even if the award was rendered after Romania’s accession to the EU, ‘the Commission retroactively applied the powers which it held under Article 108 TFEU and Regulation No 659/1999 to events predating Romania’s accession to the European Union. Therefore, the Commission could not classify the measure at issue as State aid within the meaning of Article 107(1) TFEU’ (para. 92 of the GCEU’s decision).

What is interesting, though, is that the GCEU referred only to a part of the compensation as not being under the Commission’s power of review. It did not exclude the entirety of the award from the Commission’s reach: ‘as regards the amounts granted as compensation for the period subsequent to Romania’s accession to the European Union, namely, the period from 1 January 2007 to 1 April 2009, even assuming that the payment of compensation relating to that period could be classified as incompatible aid, given that the Commission did not draw a distinction between the periods of compensation for the damage suffered by the applicants before or after accession, the Commission has, in any event, exceeded its powers in the area of State aid review’ (para. 91 of the GCEU’s decision). In other words, had the Commission distinguished between the pre-accession and the post-accession periods, the decision may not have been annulled after all (or may have been only partially annulled).

Clearly, the GCEU left open the possibility of finding an incompatibility between State Aid rules and the observance of an arbitral award rendered for acts which occurred after EU law became applicable. And this is what the decision does not entail: it does not clarify whether compensation payable on the basis of an arbitral award is contrary to EU State Aid rules.

It is true that the Court began an analysis of whether compensation offered on the basis of an arbitral award can be considered State Aid, but it stopped short of drawing any relevant conclusions. It limited itself to referring to the general conditions necessary for State Aid to arise (paras. 100-103 of the GCEU’s decision) and concluded that it cannot be considered that the compensation amounted to a form of illegal State Aid, at least not until the accession period. However, after the accession period, the analysis would advance to the issue of whether the objective elements of illegal State Aid were present: this, however, was not undertaken by the Court. It never determined whether the measure was imputable to Romania. And one can clearly see why the Court avoided this. It would be very hard to argue that the compensation ordered by the arbitral award can amount to illegal state aid.

Firstly, how can one impute an investment tribunal’s award to Romania? This would mean that Romania had control over the arbitrators, which is clearly not the case. Quite the opposite, as otherwise arbitration would not have been used so often in the settlement of investor-state disputes. Neutrality is one of the reasons ISDS exists. Additionally, for state aid to exist, one needs to demonstrate effective control of the state over the body which adopts the decision alleged to constitute such state aid (para. 52 of the Stardust case – France v. Commission, Case C-482/99). As shown earlier, this is clearly not the case with an investment arbitral tribunal.

Moreover, the GCEU mentioned that ‘compensation for damage suffered cannot be regarded as aid unless it has the effect of compensating for the withdrawal of unlawful or incompatible aid’ (para. 103 of the GCEU’s decision). This must be read together with the Court’s earlier judgment in the Asteris case. The basis of this case-law is that ‘State Aid […] is fundamentally different in its legal nature from damages which the competent national authorities may be ordered to pay individuals in compensation for the damage which they have caused to those individuals’ (para. 23 of the Asteris judgment). In this context, one must tread carefully before concluding that the subsequent compensation is, in fact, a hidden form of State Aid. Given the evident difference between the two, it is of utmost importance to demonstrate in-depth that in a specific case this difference is diluted.

One underlying premise for this difference to be able to disappear is for the EU Member State to actually be the one which formally re-institutes the illegal aid through the formal measure of compensation. The two measures – the initial state aid and, subsequently, the compensation for the withdrawal of the unlawful measure – must be seen as a whole, as having one purpose and as being able to be imputed to one entity – in this case, the Romanian state. In the Micula case, though, this was not present. The initial measure was indeed adopted by the Romanian state. The compensation, though, was decided by an objective and neutral tribunal. They are related, but they do not constitute one whole. Not to mention the fact that it can be very hard to argue that compensation on the basis of an award could offer unjustified economic advantages.

Secondly, one other condition for the compensation to be considered as re-instituting the illegal State Aid is for the compensation to be structured so as to replace the illegal measure itself. Nonetheless, this was not the case with the Micula award. One aspect must be taken into consideration in order to understand the difference between the customs and tax incentives themselves (the illegal State Aid) and the arbitral award. As mentioned at the beginning of this post, it was not the withdrawal of the incentive schemes that was considered to be the basis of compensation. What led to the present outcome was the manner in which the withdrawal took place, essentially leading to an infringement of legitimate expectations. Those are different and it is clear that, in any case, this would not be a case of re-instituting said state aid through the backdoor.

As such, the GCEU’s award is clearly not a silver lining for intra-EU ISDS, as it does not clarify – in the end – the most important aspect: can compensation rendered by an arbitral award be considered illegal state aid? In this context, when one thinks about the general scheme of things, it becomes evident that nothing has really changed: Achmea is alive (the effects have come sooner rather than later). Additionally, nobody knows its scope, especially when it comes to the Energy Charter Treaty’s (ECT) arbitration mechanism. Although arbitral practice seems to insist that Achmea does not preclude intra-EU ISDS on the basis of the ECT, what is eagerly waited is the CJEU’s position on this. After this, the CETA opinion – although reconciling ISDS with EU law when there is a third party (a party outside the EU) involved – does not mean the endorsement of intra-EU ISDS; it can clearly be seen that the EU’s position within UNCITRAL’s Working Group III is still the one we have been used to for so long: ISDS must be replaced with a standing court.


[1] LL. M. candidate in Law and Economics at Utrecht University.

Post-Achmea Energy Charter Treaty Coherence and Stability: Upheld or Hindered?

Alexandros Catalin Bakos, LL. M.*

[…] but this is not where or how it ends. Fate promises more twists before this drama unfolds…completely (in-game dialogue from the intro scene of the video game Soul Reaver 2).

The EU’s backlash against intra-EU (Bilateral?) Investment Treaties – intra-EU (B)ITs – reached its peak when the CJEU issued its decision in the Achmea (C-284/16) case. According to the CJEU, intra-EU BITs such as the one analysed in the Achmea case are contrary to EU Law because they created a parallel jurisdiction (that of investment arbitration tribunals) to that of the domestic judicial courts. Such a jurisdiction may impair especially the consistency, full effect and autonomy of EU Law because investment arbitration tribunals are not able to rely on instruments such as the preliminary question (§§ 35-60 of the Achmea decision). Although the Achmea decision has been criticised (here and here), the present analysis is not concerned with the merits of the decision itself. The object of this analysis regards the effects of the Achmea decision on the Energy Charter Treaty’s (ECT) provisions on investment. This is of high practical importance since the International Investment Agreement which is most commonly invoked in intra-EU investment disputes is the ECT. An analysis of this issue raises the following questions:

Firstly, what are the immediate effects of the CJEU’s judgement on Article 26 (3) (a) of the ECT (the ECT’S Investor-State Dispute Settlement  – ISDS – provision)? Any analysis should begin with an analysis of the meaning of intra-EU BITs and if that meaning shall extend to the ECT – a treaty to which the EU is a formal party –, as well. As will be seen, the fact that the EU is a formal party to the ECT is of high importance (1).

Secondly, if it is to be considered that the Achmea decision does refer to the ECT, as well, and, as such, that it is conflicting with the ECT ISDS provision as regards EU Member States parties to the ECT, it must be seen whether the practice of terminating the intra-EU BITs between EU Member States can be undergone in the case of the ECT, as well. As such, could EU Member States – only as between themselves – denounce – partially or in its entirety – the ECT (2)?

The last point of this analysis is whether the EU’s international responsibility under Public International Law could be engaged for the Achmea decision – provided it is considered that the Achmea decision does refer to the ECT, as well. This question arises since the EU is a formal party to the ECT and an analysis needs to be made as regards the compliance of such an act – the Achmea decision – with the ECT (3).

Before concluding, I will address a less evident but very important issue generated by the Achmea decision – again, provided it is considered to refer to the Energy Charter Treaty, as well. The issue regards the systemic effects of the decision on the International Legal Order (4).

  1. What is the meaning of intra-EU BITs? If it covers the Energy Charter Treaty – as between EU Member States –, how does it affect the ISDS provision therein?

The departing point of analysis is the Achmea decision itself. The CJEU expressly made a differentiation, within the decision between investment treaties to which the EU was a formal party and those to which it was not (see §§ 57-58 of the Achmea decision). Essentially, this differentiation was made in the context of describing the characteristics of the BIT which was thought to conflict with the EU legal order (§58). Moreover, the EU pointed out that an international agreement which sets up a dispute resolution mechanism and is binding on the EU institutions is not in principle incompatible with EU law (§57).  It is hard to conceive that the CJEU made this differentiation by accident. In fact, it can be reasonably derived from here that the CJEU wanted to limit the scope of the decision’s effects by referring expressly to treaties to which the EU was not a formal party.

As the CJEU itself referred to international agreements to which the EU was a formal party and as the ECT is such an agreement, it follows that the CJEU considers that a different legal treatment shall be applied to such agreements – in this particular situation, to the ECT – than to intra-EU BITs – to which the EU is not a formal party. It can reasonably be inferred that this differentiation is based the principle of pacta sunt servanda (Article 26 of the 1969 Vienna Convention on the Law of Treaties – VCLT) which binds the parties to an international treaty. This principle is doubled within the EU legal sphere by Article 216 (2) of the TFEU: agreements concluded by the Union are binding upon the institutions of the Union and on its Member States. As such, for the Achmea decision to be compatible with the obligations deriving from the ECT – especially the obligation contained in Article 26 (3) of the ECT – and those incumbent on the EU institutions by virtue of Article 216 (2) of the TFEU, any interpretation of the Achmea decision, in order not to be unreasonable and self-contradictory, must be made to the extent that the CJEU did not refer to the ECT in its decision. If the CJEU had wanted the Achmea decision to refer to the ECT, it would have expressly mentioned this so as not to create confusion as regards a possible infringement of Article 216 (2) TFEU – in addition to the infringement of the ECT. In other words, the CJEU must have been aware that it was under a duty, if it had wanted the Achmea decision to refer to the ECT, to actually explain why such a decision would not have contradicted Article 26 (3) of the ECT and Article 216 (2) of the TFEU. Not doing this, the CJEU basically concluded that the ISDS provision in the ECT is not contrary to the EU legal order. Moreover, in this context, it is hard to envision that the EU would have entered into ECT negotiations and would have subsequently become a party to the ECT had it considered the ECT as contrary to EU Law (RREEF Infrastructure (G.P.) Limited and RREEF Pan-European Infrastructure Two Lux S.à r.l. v. Kingdom of Spain (ICSID Case No. ARB/13/30). Decision on Jurisdiction, § 76).

At this point, although the previous conclusion seems logical and necessary, there are analyses that accept the possibility that the Achmea decision referred to the ECT, as well (here and here). Moreover, it has been argued, constantly, before arbitral tribunals applying the ECT that the ISDS provision contained within the ECT is incompatible with EU Law (RREEF v. Kingdom of Spain. Decision on Jurisdiction, §§ 40 – 50; Charanne B.V., Construction Investments S.A.R.L. v. The Kingdom of Spain. Final Award, Court of Arbitration of the Chamber of Commerce, Industry and Services of Madrid (Arbitration No.: 062/2012), unofficial translation by Mena Chambers, §§ 207 – 224; Masdar  Solar & Wind Cooperatiff U.A. v. Kingdom of Spain. Award (ICSID Case No. ARB/14/1), §§ 296 – 300, § 305 and § 325). In all the cited cases, jurisdiction was upheld by the arbitral tribunals. It was considered that there was no conflict under Public International Law between the ECT and the EU Law. However, there are some arguments relied on to support the contention that a tribunal does not have jurisdiction over intra-EU disputes based on the ECT which I would like to mention here – not exhaustively, but only as examples – in order to clarify the debate. For example, it was argued that there existed an implicit disconnection clause within the ECT as regards intra-EU ECT disputes, because of the nature of the EU Legal Order. The effect of such a clause would be that in intra-EU investment disputes EU Law would derogate from the ECT, rendering the latter inapplicable. Moreover, it was argued that there was no difference between the territory of the home state and that of the host state when both were EU parties. As such, the condition that the territories of the host state and of the home state must be different (Article 1 (10) (a) and (b) of the ECT) was not satisfied (the Charanne Award, § 214).

The argument that the ECT impliedly included a disconnection clause which rendered the ECT inapplicable as between EU Member States is flawed on different levels:

Firstly, an implied disconnection clause would run contrary to the pacta sunt servanda principle – the implied disconnection clause is nothing more than a speculation relied on to avoid treaty obligations. Moreover, the same pacta sunt servanda principle is contrary to an implied disconnection clause if such clause is not expressly contained in the ECT. This is because obligations must be observed as agreed by the parties and supposing the existence of an implied disconnection clause would actually be contrary to Article 31 (1) of the VCLT, which sets up an interpretation of the treaty according to the ordinary meaning to be given to the terms of the treaty. This latter point regarding interpretation on the basis of Article 31 of the VCLT was reinforced by the Charanne tribunal (§ 437).

Secondly, the negotiating history of the ECT shows that, although a disconnection clause was proposed by the European Community bloc and expressly rejected, the EU still became a party to the ECT. This essentially means that the parties rejected the disconnection clause and any interpretation to the contrary would be unjustified under Article 32 of the VCLT, which in this case would mean reliance on the negotiating history to confirm the initial interpretation (see, for the use of Article 32 of the VCLT as a means to confirm the interpretation made under Article 31 of the VCLT, Mark E. Villiger, Commentary on the 1969 Vienna Convention on the Law of Treaties, Martinus Nijhoff Publishers, Leiden, Boston (hereinafter referred to as Villiger), 2009, pp. 446-447).

As regards the territorial identity in the case of the host and the home state of the investor, this argument was rebutted, as well. It was found that being a state party to a Regional Economic Integration Organization (REIO) and party to the ECT while that REIO (the EU, in the present case) is a party to the ECT, as well, does not create an identity between the territory of the state and that of the REIO. This is true as long as the REIO and the state party to the REIO can both have individual standing as respondents under the ECT (Novenergia II – Energy & Environment (SCA) (Grand Duchy of Luxembourg), SICAR v. The Kingdom of Spain. Final Arbitral Award, Arbitration Institute of the Stockholm Chamber of Commerce (2015/063), § 453).

Notwithstanding all the above arguments, there was even an arbitral tribunal which ruled expressly that the Achmea decision did not apply to the ECT: the Masdar tribunal (§§ 678 – 683) effectively ruled that the Achmea decision is limited to intra-EU BITs, excluding, thus, multilateral investment treaties such as the ECT.

  1. Could EU Member States – only as between themselves – terminate – partially or in its entirety – the ECT?

If it was considered that the Achmea decision, in spite of the above, would apply to the ECT, as well, this would raise another practical issue: could the EU Member States terminate the ECT between themselves, similarly to what has been done regarding intra-EU BITs? How would this work? Would this be a partial termination – only as regards Article 26 (3) of the ECT – or a complete termination? Such questions raise issues of treaty termination by reference, especially, to the object and purpose of that treaty. In order to answer the previous questions, the analysis is divided in two parts: firstly, the issue of treaty termination as regards the possibility of only certain parties to the treaty to proceed to this end shall be addressed (a). Subsequently, it must be seen whether a partial termination of the ECT – as regards Article 26 (3) only – is indeed a real possibility when tested against the object and purpose of the Energy Charter Treaty (b).

  • The possibility of certain parties to a multilateral treaty to denounce it only as between themselves:

The ECT provides in Article 47 (1) that […] a Contracting Party may give written notification […] of its withdrawal from the Treaty. While this clarifies the general issue of withdrawal, the question remains whether the EU Parties can denounce the ECT as between themselves only. It is considered that a partial withdrawal vis-à-vis several, but not all of the other parties, is possible (Thomas Giegrich in Oliver Dörr, Kirsten Schmalenbach (editors), Vienna Convention on the Law of Treaties. A Commentary, Springer – Verlag Berlin Heidelberg, 2012, pp. 952-953, § 25). However, while this may seem possible, generally, serious issues may arise when attempting to terminate the ECT as between certain parties to it only. This can be seen when interpreting Article 47 (1) of the ECT according to Article 31 of the VCLT – in the light of the object and purpose of the ECT. As will be demonstrated within (b) of this part of the analysis, a partial termination of the ECT would effectively create micro-regimes within the ECT and this would be against the object and purpose of the ECT. Would termination between certain EU parties not have the same effects? More specifically, coherence as to measures in the energy sector would be affected if what is applied under the ECT between EU and non-EU parties would not be applicable between EU parties. Any policies, in this context, lose their cogency, because of lack of (even legal) coherence. In effect, this would defeat the object and purpose of the ECT, since coherent policies are incredibly important in the energy sector – for example, major infrastructure projects, such as pipelines, usually span over several states. Because of all this, it can be argued that Article 47 (1) of the ECT must be interpreted as precluding termination between EU Member States only.

  • Is it possible to partially terminate the ECT as regards the ISDS provision?

Partially terminating the treaty, while possible (Villiger, p. 685), is more problematic, in the present context. Article 44 (1) of the VCLT provides that withdrawal from a treaty, where provided expressly by that treaty, may be undertaken only with respect to the whole treaty unless the treaty otherwise provides or the parties otherwise agree. Accordingly, since there is no express provision as to partial termination within the ECT, the only legal basis for partially terminating the ECT as between the EU Member States would be if the ECT parties agreed. Nonetheless, even if there is no provision as to partial termination in Article 47 (1) of the ECT, an analysis of the remainder of Article 44 (2) – (4) of the VCLT – which sets out, exceptionally, the legal regime of severability, especially when there is no express provision as to partial termination – can still be undertaken (Villiger, p. 563). In this respect, Article 44 (3) provides a series of conditions which must be met – cumulatively (see the term and at the end of indent (b)) – for separability to be possible.

A problem with such an outcome is that one of the conditions provided for in Article 44 (3) of the VCLT is not met: that the clause which is sought to be terminated does not hold a high importance in the general architecture of the treaty. Or, in the words of Article 44 (3), that acceptance of the clause – which is sought to be terminated – was not an essential basis of the consent of the other party or parties to be bound by the treaty as a whole. This is essentially an indirect reference to the object and purpose of the treaty.

The importance of the ISDS provision – which is the clause which the EU Member States would want to terminate –, in this context, is fundamental. Firstly, an investment treaty lacking an ISDS mechanism would be devoid of all practical effect (Opinion of Advocate General Wathelet in the Achmea case, § 207; although the AG referred to BITs in the context of this statement, the reasoning can easily apply to any investment agreement since the importance of ISDS is the same). Secondly, it can be seen that a proper investment framework is an important element needed to attain the object and purpose of the Energy Charter Treaty. This is demonstrated by the ECT’s preamble, which repeatedly mentions the importance of a proper investment regime to the attainment of the ECT’s goals.

In this context, in order to understand the impact of a partial termination of the ECT as between EU Member States in the case of the ISDS provision of the ECT and the relationship of the ISDS provision to the object and purpose of the ECT, consider the following: if EU Member States were able to denounce the ISDS clause of the ECT as between them and leave the treaty in effect between them and the other parties – from outside the EU –, this would, effectively, create micro-regimes within the ECT system – especially since energy investment in the EU by investors from within the EU would not be covered by the ECT anymore or, at least, not by the ISDS clause. Would this not defeat the object and purpose of the treaty, since such a fragmentation would hinder the possibility of attaining the objectives the treaty was supposed to achieve? Perhaps the most important objective the ECT set out to achieve was a “level playing field for investment in the energy industry, which is notoriously complex, expensive and long-term in nature (Norah Gallagher, The Energy Charter Treaty (1994) (ECT), WORLD ARBITRATION REPORTER 2d Edition, p. 3). Additionally, proper investment in the energy sector is needed for attaining security of supply (Sanam S Haghighi, Energy Security. The External Legal Relations of the European Union with Major Oil- and Gas-Supplying Contracts, Hart Publishing, Oxford and Portland, Oregon, pp. 24-25).This is because investments in the energy sector are characterised by a Return of Investment spanning sometimes even several decades. Thus, an investor wants to be assured that the protections will be in place over such a time-span. However, lack of such protections – which refers to the existence of ISDS, as well – may disincentivise a potential investor to invest in the energy sector. And this is how security of supply may be compromised, in addition to compromising the object and purpose of the ECT, in the first place.

The previous considerations can reasonably lead to a conclusion that any termination as regards the ISDS provision of the ECT is not possible owing to the provisions of Article 44 (3) of the VCLT, because of the importance of ISDS to the object and purpose of the ECT. What this means, effectively, is that EU Parties have only one choice: termination of the ECT as between themselves – whether in its entirety or only partially – only by agreement between all the parties to the ECT. Such an outcome is hard to imagine: non-EU ECT parties’ companies have subsidiaries registered in the EU. If the intra-EU investment regime in energy matters governed by the ECT were to be affected (or even the ECT in its entirety as between EU Member States), this would effectively affect such companies. Because of this, it is highly unlikely that the other ECT Parties would agree to partial termination of the ECT as between EU Member States – either as to the entire treaty or only regarding Article 26 (3) of the ECT.

As such, it has to be concluded that EU Member States which are parties to the ECT cannot, only on the basis of their own will, terminate the ECT as between themselves – neither completely nor partially. And since it is practically very hard to envision acceptance by the non-EU parties – among them existing energy-exporting states which hold negotiating power – as regards intra-EU ECT termination, the answer must be that, for practical purposes, it is more likely that the EU Member States cannot terminate the ECT.

  1. Could the EU’s international responsibility be engaged for the Achmea decision?

Whatever the answers to the previous enquiries are, the EU is bound by the provisions of the ECT which it accepted when it signed and ratified the treaty. As such, if the Achmea decision refers to the ECT, it is  contrary to the provisions of the latter. In this context, an analysis must be undertaken regarding the responsibility of the EU for internationally wrongful acts. However, such an analysis implies two different steps: firstly, it needs to be seen whether the responsibility of international organisations for internationally wrongful conducts exists under International Law (a). If it can be demonstrated that such responsibility indeed exists, it must be seen if the Achmea decision can lead to engaging the responsibility of the EU (b).

  • Does responsibility for internationally wrongful acts exist in the case of International Organisations?

The idea behind the existence – or lack – of responsibility for internationally wrongful acts committed by IOs is a complex one and it is not my attempt, within the present analysis, to exhaustively address it. However, for clarity of the argument, before analysing the issue of the Achmea case, the following need to ascertained: firstly, is there any legal basis for responsibility of IOs? Secondly, what is the scope of such responsibility in the case of the IOs and, specifically, in the case of the EU?

It is accepted in legal literature that there exists a legal basis for the responsibility of IOs (Mirka Möldner, Responsibility of International Organizations – Introducing the ILC’s DARIO, in A. von Bogdandy and R. Wofrum (eds.), Max Planck Yearbook of United Nations Law, Volume 16, 2012, pp. 286-287; Konrad Ginther, International Organizations, Responsibility, in Rudolf Bernhardt (ed.), Max Planck Encyclopedia of Public International Law. International Organizations in General. Universal International Organizations and Cooperation, Elsevier Science Publishers B.V., Amsterdam, The Netherlands, 1983, p. 162). Whether this is based on custom, principle or even the international legal personality of the IO is not important for present purposes (although it is accepted that the legal source for responsibility of IOs for internationally wrongful acts could be any of the previously-mentioned sources). However, what must be mentioned, here is that it can hardly be argued that there exists a single unified regime regarding the framework of international responsibility of IOs for internationally wrongful acts (see p. 5 of the linked article). The difference between IOs, their legal characteristics – such as the principle of speciality – do not justify a single legal regime (Ibid.). As such, I will not pursue this analysis by relying on the general framework set by the International Law Commission’s Draft Articles on Responsibility of International Organizations (DARIO). They are not considered to reflect customary international law (p. 9 of the linked article) and, at the same time, they offer a general framework whereas I referred earlier to the fact that hardly any general regime can be considered to exist to this end. What I will do, instead, is look for any legal elements which could justify the responsibility of  the EU for internationally wrongful acts.

A solution can be found in one of the EU’s internal acts themselves: EU Regulation No. 912/2014 establishing a framework for managing financial responsibility linked to investor-to-state dispute settlement tribunals established by international agreements to which the European Union is party. It is provided there that financial responsibility arising from a dispute under an agreement (IIA) shall be apportioned to the Union when such financial responsibility arises from treatment afforded by the institutions, bodies, offices or agencies of the Union (a) or when such financial responsibility arises from treatment afforded by a Member State where such treatment was required by Union law (Article 3, 1. of Regulation No. 912/2014). It is true that this provision refers to financial responsibility – which entails an obligation to pay a sum of money awarded by an arbitration tribunal or agreed as part of a settlement and including the costs arising from arbitration (Article 2 (g) of Regulation No. 912/2014) – and not exactly to what is commonly understood as responsibility for internationally wrongful acts. However, such financial responsibility cannot exist in a void. Unless a violation of an IIA occurred (under Public International Law, this is a violation of the primary norms which triggers the secondary norms on responsibility and, specific to the present regulation, the norms on reparation – financial responsibility), financial responsibility would not exist. Moreover, the premises for engaging the financial responsibility of the Union is that the Union was actually the catalyst to the infringement of the IIA (under Public International Law, this would actually refer to attribution of the acts to the EU). It is doubtful that the EU would have adopted such a legally binding document on itself unless it had considered that there existed an obligation under Public International Law to provide reparation for internationally wrongful acts which could be attributed to it (this being a sign that the EU acted out of a sense of obligation when it bound itself to the triggering of its financial responsibility for internationally wrongful acts caused by it – essentially, this would be the opinio juris of the customary norm on responsibility). And because the secondary norms on responsibility for internationally wrongful acts are inextricably linked to the primary obligations of the States/ International Organisations under Public International Law and, in all actuality, cannot exist if the primary ones are not breached, one can only analyse the law on responsibility of IOs for internationally wrongful conducts in this context. As such, the entire procedure would be: firstly, an analysis of the breach of the primary norm would be made; secondly, attribution of the initial violation would be undertaken, which would result in engaging the responsibility of the perpetrator; finally, reparation would occur – which is what the financial responsibility actually means. Because of this procedure, there cannot exist reparation – financial responsibility – without attribution and, continuing the reasoning, without a breach of the primary norm. Thus, the EU actually conceded within Regulation No. 912/2014 that it considered itself bound by the customary norm on responsibility of IOs for internationally wrongful acts – and, implicitly, that this is part of International Law.

One must admit the possibility that a counterargument can be brought as regards the previous argument along the following lines: Regulation No. 912/2014 is a legally binding instrument for intra-EU relations and, as such, it does not reflect opinio juris as regards a customary norm on responsibility of IOs for wrongful acts on part of the EU. While prima facie this could seem true, a look at the context and language of the Regulation would render such a counterargument moot. Firstly, the Regulation is concerned with the EU’s external relations with other subjects of International Law. This means that it reflects the EU’s perspective on the international law of Multilateral Investment Treaties, which is, in the end, concerned with primary obligations of Public International Law. Secondly, it is expressly provided within the Regulation what shall happen when the EU is a respondent in arbitration proceedings initiated by a claimant (Article 4 of Regulation No. 912/2014). Both the previous considerations clearly state that the Regulation reflects the EU’s opinion as to its legal relations under Public International Law. Because of this, the conclusion that the Regulation No. 912/ 2014 reflects the EU’s opinio juris as regards responsibility of an IO for internationally wrongful conducts – at least those in breach of a Multilateral Investment Treaty, if it is considered that different customary rules apply (or do not exist) when different regimes are concerned – is valid.

Another counterargument which could be brought against the international responsibility of the EU is that a  custom as the one mentioned earlier is hard to ascertain due to lack of clarity regarding the practice element of a custom. In any case, the legal basis for engaging the EU’s responsibility for internationally wrongful acts does not even have to be a custom. As mentioned earlier, the source of such responsibility could be a principle of law. Moreover, an unilateral act could give rise to obligations under International Law (Wilfried Fiedler,  Unilateral Acts in International Law, Rudolf Bernhardt (ed.), Encyclopedia of Public International Law. History of International Law. Foundations and Principles of International Law. Sources of International Law. Law of Treaties, Elsevier Science Publishers B.V., 1984, pp. 517-518 and 522). Regulation No. 912/2014 is such an unilateral act  (from an International Law point of view, as it concerns the EU as a single entity) for purposes of ascertaining an obligation of reparation on the part of the EU – an obligation which could not logically exist without engaging the responsibility of the EU. What is relevant, however, is that Regulation No. 912/2014 demonstrates more than the EU’s opinion juris as regards its international responsibility for internationally wrongful acts in investment matters. The position of the EU is that there indeed is a legal obligation (this can be ascertained from the mandatory language employed within Regulation No. 912/2014 and from its binding character) and not just a sense of a legal obligation to provide reparation for internationally wrongful acts attributed to the EU in the sphere of investment agreements to which the EU is a party. In other words, the source of the legal obligation must not necessarily be the custom; it can be any of the previously-mentioned sources. And, continuing the reasoning, this demonstrated the existence of international responsibility on the part of the EU for such internationally wrongful acts.

As for the legal forum where the responsibility of the EU for violations of the ECT could be engaged: that would be an arbitral tribunal which shall rule upon issues concerning the ECT – this is supported by Regulation No. 912/2014 and was reinforced by an arbitral tribunal (Electrabel S.A. v. The Republic of Hungary (ICSID Case No. ARB/07/19). Decision on Jurisdiction, Applicable Law and Liability, § 3.21).

  • Can the Achmea decision be considered a breach of the ECT and, as such, entail the responsibility of the EU?

It was demonstrated that the EU’s responsibility can be engaged for breaches of international investment agreements and, moreover, that the EU itself acknowledges this by undertaking the obligation to repair the harm caused by its acts or by those of EU Member States generated by it. This means that there is legal basis for attribution and reparation of the wrongful act. As such, is the Achmea decision such an act? At this point in time, it is too early to tell clearly. The premise of this part is that the Achmea decision refers to the ECT. Thus, the effects would be that EU Member States, sitting as respondents in intra-EU arbitration on the basis of Article 26 (3) of the ECT, would be in breach of EU Law. Contrariwise, respecting the Achmea decision would entail breaching Article 26 (3) of the ECT. As such, if an arbitral tribunal found such a breach, attributed it to the EU and, moreover, found that such breach gave rise to an obligation of reparation (financial responsibility), this would demonstrate that the EU’s responsibility for the Achmea decision could be engaged. However, this would not be a typical investment arbitration as the tribunal would not be judging a violation of investment standards of protection. It would be effectively analysing a violation of the arbitration clause in the ECT. Nonetheless, there is the possibility that this violation could give rise to a claim based on the Fair and Equitable Treatment standard, for violation of legitimate expectations as regards dispute settlement. But even if the claim is based only on a violation of Article 26 (3), legal cause for such a claim would still exist. This is because the applicable law would be the ECT in its entirety, not only the investment standards of protection (Article 26 (7) of the ECT). What remains to be seen is whether the Achmea decision itself is enough for a claim against the EU or if EU Member States need to act on the basis of the Achmea decision in order to generate a claim against EU. Practically, it is more likely that the latter would be the case, since harm would be easier to assess in that context.

In conclusion, the Achmea decision – if it is considered that it refers to the ECT, as well – can potentially give rise to the engaging of the EU’s international responsibility for internationally wrongful acts.

  1. Systemic effects of the Achmea decision:

As mentioned in the beginning, this point of analysis is applicable only if it is considered that the Achmea decision refers to the Energy Charter Treaty, as well.

When I analysed the applicability of the Achmea decision to the Energy Charter Treaty, I referred to the negotiating history of the ECT. I mentioned, in that context, that the negotiating parties rejected a proposal by the European Commission to derogate from the rules of the ECT – even those concerning the dispute-settlement clause in Article 26 (3) – as between the EU Parties. Nonetheless, the EU still signed and ratified the ECT, essentially admitting under Public International Law that the ECT shall be applicable to the EU parties.

This has strong implications, from a systemic point of view: if the EU is concerned about rule of law standards – especially coherence –, it has to take into account the obligation under Article 26 (3) of the ECT taken together with the representations it made during negotiations to the ECT. The issues of coherence regard the relations between the EU and the other parties to the ECT – even EU Member States which, under Public International Law, are different formal parties to the ECT than the EU and their interests may not always converge. Here, coherence is a fundamental aspect of the liberal doctrine within International Relations which comes to explain the functioning of the Public International Law mechanism (Andrea Bianchi, International Law Theories. An Inquiry into Different Ways of Thinking, Oxford University Press, Oxford, United Kingdom, 2016, pp. 113-114). In other words, coherence is a fundamental pillar of the rules-based international order on which especially the Western Powers seek to rely. Since the rules-based international order manifests itself within an anarchic world where a central executive agency which could guarantee the enforcement of the rules does not exist, coherence has a special meaning in this context. And since the liberal theory of international relations which I mentioned earlier and which underpins a great number of arguments regarding the effectiveness of international law as a part of a rules-based international order, is based, among others, on international cooperation and mutual benefits, coherence is, effectively, necessary for such cooperation and trust to exist. Because of this, without coherence there would hardly be any stable international legal order. In other words, when one of the major economic actors in this system – the EU – is not coherent in its approach to its international obligations and is actually trying to enforce its views regarding the supremacy of EU Law to the detriment of International Law upon its member states, any feeling of mutual benefits and international cooperation is eroded. Trust in the actions of the major international actors becomes scarce and incentives actually appear which determine the other actors to start ignoring international rules, as well. This is basically the prisoner’s dilemma after one of the parties deflected. One can easily imagine the future responses of the other parties after experiencing the real risk of deflection and wondering whether such deflection is recurrent. And this is how cracks appear in the current international law architecture. As such, the most important actors of such an international order – the EU being among them – have a special duty to ensure that this order is maintained and that the mechanism which underpins the effectiveness of this order is properly functioning. Thus, the EU – a proponent of the rules-based international order – should reassess its approach to its international obligations and, in this specific case, to the ISDS provision within the ECT.

  1. Conclusion:

There is no clear and predictable answer as to what will happen after the Achmea decision as regards the Energy Charter Treaty. The variables are numerous and they are generated by decisions adopted by different actors: the ECJ – in its future decisions and opinions on the issue of ISDS, such as the opinion requested by Belgium regarding investment provisions contained in CETA; the EU Member States – who have yet to decide what will their approach to the ECT be after some of them decided to terminate intra-EU BITS; and, finally, arbitral tribunals who are faced with challenges to their jurisdiction, requests to reopen proceedings or to review their decisions or simply the fact that such tribunals are faced with an additional element which must be accounted for when adopting decisions to the ECT cases which are still pending before them. However, what is certain is that the whole issue has gone past the point where the catalyst to future evolutions was the CJEU. While the CJEU still plays an important role in this whole issue – perhaps the most important one –, it is not alone anymore in influencing the final outcome. Nonetheless, it can still find a way to balance the interests and recreate a framework of relative predictability. But in order to do this it must account for several considerations: it must understand that the EU legal regime is not a self-contained one and any future decision on the part of the CJEU as regards ECT Arbitration has several implications which are, in essence, produced under the framework of Public International Law and not only under the framework of EU Law. Ignoring those aspects may lead to severe hidden consequences which nobody would desire: weakening of the ECT and more unpredictability in the energy sector; questions of responsibility under Public International Law for internationally wrongful acts; and, perhaps most important, problems of coherence both at International and EU regional level. In the end, all the above create problems of legitimacy and one can ask himself: are the consequences really bringing more benefit than harm?


* I am aware of the European Commission’s latest Communication on Protection of intra-EU investment (19.7.2018). And while it is true that the Commission expressly referred to the Energy Charter Treaty investor-State arbitration mechanism established in Article 26 of the Energy Charter Treaty as being covered by the Achmea decision (pp. 3-4 and 26 of the cited Communication), the present analysis is still relevant, as are the arguments herein. This is because of two reasons: firstly, the aforementioned Communication is, essentially, the EU Commission reinforcing its traditional position as regards intra-EU investment arbitration. However, this is not a new position on the part of the EU Commission, as it has repeatedly argued against the incompatibility between intra-EU Investment Treaties (including the investment provisions of the Energy Charter Treaty) and the EU Legal Order. Secondly, the EU Commission’s Communication does not clarify the Achmea decision itself. Unless the CJEU expressly and unambiguously considered the Energy Charter Treaty as being contrary to the EU Legal Order, the questions regarding the scope of the Achmea decision and its applicability to the Energy Charter Treaty would still exist.

The Relationship between EU State Aid law and Obligations Arising under Investment Treaties

by Alexandros Catalin Bakos, LL.M 

I. Introduction:

In recent years, a series of debates have emerged in regard to the relationship between the EU State Aid law[1], on the one hand, and obligations arising under Investment Treaties (to which the EU is not a formal party)[2], on the other hand. Those debates manifest themselves at different levels and have powerful implications: firstly, they clarify the scope of State Aid law and its relationship with one of the most important fields – that of Investment Law. Secondly, they clarify – or complicate, depending on the vantage point from which one analyses the issue – the relationship between EU law and Public International Law[3]. And, thirdly, they raise questions of interpretation of EU law, especially from a historical interpretation point of view and from a teleological point of view – this is a great tool to understand the limits of EU law (the real limits, not the attempts to politically force an interpretation which extends the limits of EU law beyond what the Member States had envisioned initially).

Needless to say, the practical importance of the relationship between State Aid law and obligations arising under ITs can hardly be overstated. One can only look at the recent Micula affair[4] and the unenviable position in which Romania finds itself: on the one hand, it is faced with severe opposition from the European Commission as regards the observing of certain obligations arising under ITs (more specifically, the obligation to pay compensation to the Micula brothers as the final award against Romania dictates). On the other hand, Romania cannot outright ignore the legal framework set by the ITs (including the binding effect of the awards within this field) and show total disregard to the interests (and even rights) of investors.

As such, I endeavour in this study to provide an analysis of this relationship between State Aid law and obligations arising under ITs. I will focus my attention only on the first tier of this issue – the relationship between State Aid law and obligations arising under ITs themselves – and, as such, I will not analyse more general issues such the relationship between General Public International Law and EU law. Moreover, I will ignore general issues of interpretation of EU law. However, those issues will be touched upon where relevant for the analysis conducted through the present study.

II. Analysis:

Before starting, it should be stated that this analysis is composed of two parts. Firstly, I will analyse the relationship between State Aid law and obligations arising under ITs signed by EU Member States (intra-EU ITs) (1). Subsequently, I will analyse what the relationship between State Aid law and obligations arising under ITs when the ITs’ signatories are both from within the EU and from outside the EU (2).

  1. The relationship between State Aid law and obligations arising under ITs signed by EU Member States (intra-EU ITs):

(a) Scope of analysis:

The problem with intra-EU ITs and State Aid law seems to be that compensation given to investors by member states, as a result of an investment tribunal award, is considered illegal state aid, in cases such as the Micula one.[5] As such, the analysis should address the following: can an investment award rendered by an investment tribunal on the basis of an intra-EU IT be considered illegal State Aid? If so, when can it be considered as such (b)? Following, the next question should be: notwithstanding specific issues of whether enforcement of an investment arbitral award can be considered illegal State Aid, is it justified to ever argue for the termination of an intra-EU IT relying on State Aid law? In other words, can the intra-EU IT, by itself, be considered as violating EU rules on State Aid (c)?

(b) Can an investment award rendered by an investment tribunal on the basis of an intra-EU IT be considered illegal state aid? If so, when can it be considered as such?

In order to address this question, the first issue which must be clarified is what exactly is considered illegal state aid[6]: under Article 107 of the Treaty on the Functioning of the European Union[7], state aid refers to any aid granted by a Member State or through State resources in any form whatsoever which distorts or threatens to distort competition by favouring certain undertakings or the production of certain goods in so far as it affects trade between Member States.[8]

In other words, a number of four conditions must be met in order for a measure to be considered State aid: the State must intervene through that measure (and that measure must be imputable to the state[9]); the beneficiary of the intervention must be conferred an advantage; competition must be distorted; and the intervention must be likely to affect trade between member states.[10] As such, (when) is an investment award rendered by an arbitral tribunal on the basis of an intra-EU IT considered State Aid? In order to answer the question, a qualification of an investment award which leads to an obligation on the State to pay compensation to a wronged investor must be made (including its subsequent enforcement).[11] In simple terms, an award is a final judgement or decision, esp. one by an arbitrator or by a jury assessing damages.[12] Continuing, the enforcement of an award is the act or process of compelling compliance with a law, mandate, command, decree, or agreement.[13]

Of course, there is a question which arises, at this moment: supposing an arbitral tribunal renders an award against an EU member-state based on an intra-EU IT and the State enforces it, what is the legal basis for that? The answer can be found in the two most relied-on arbitration frameworks: the ICSID[14] Convention on the Settlement of Investment Disputes between States and Nationals of Other States[15] which provides that the award given under its framework shall be binding on the parties[16] and the New York Convention on the Recognition and Enforcement of Foreign Arbitral Awards[17] which provides that each Contracting State shall recognize arbitral awards as binding and enforce them[18]. Thus, an obligation to respect and enforce arbitral awards arises for the state under Public International Law.[19]

This leads to an important question regarding State Aid law: is the condition that the measure must be imputable to the State met? The issue of imputable conduct has been defined by the CJEU as follows: this condition is met when the decisions of Member States by which, in pursuit of their own economic and social objectives, they give, by unilateral and autonomous decisions, resources to undertakings or other persons or procure for them advantages intended to encourage the attainment of the economic or social objectives sought.[20]

Can arbitral awards which are enforced as an obligation under Public International Law be considered as being unilateral and autonomous? Enforcement is, in the end, the latter part of a judicial proceeding. There is nothing autonomous or unilateral about it. And it would be artificial to separate the arbitral decision itself from the enforcement. Just because there is no central executive authority to enforce decisions rendered under Public International Law does not mean that the enforcement of those awards can be separated from the framework in which they arose: obligations owed to other legal subjects which, usually[21], have as their ultimate basis a treaty or customary relationship (a bilateral or multilateral relationship).

This view is shared by two other authors who base their approach on the fact that, when enforcing arbitral decisions, states do not act in their sovereign powers, but as agent(s) of the international community[22]. And there is undeniable merit to this view: that States, when undertaking obligations in their sovereign capacity, are giving up part of their sovereignty.[23] Moreover, even EU official bodies have constantly repeated this fact.[24] In other words, because States give up part of their sovereignty when undertaking legal obligations, they can no longer act unilaterally and autonomously within the fields and situations where they are under certain obligations.

Therefore, it can be concluded that, under EU law, as a rule, a state acting on the basis of an investor-state arbitration award, is not acting unilaterally and autonomously and, as such, the analysed measure regarding a possible issue of illegal State Aid cannot be imputable to the State – a condition for a measure to be considered State Aid.

In the following, I shall research whether there are exceptions to this rule and, as such, whether certain cases of compensating investors as a consequence of an arbitral award can be considered illegal State Aid. A practical case may offer better insights into this question.

Perhaps the most important case for the present analysis which could demonstrate whether there can exist exceptions to the rule that enforcement of arbitral awards rendered within the framework of ITs is not illegal State Aid is the Micula case. In short, this is what happened: the Micula brothers were handed, on the basis of a BIT (the Romanian – Swedish Bilateral Investment Treaty) certain custom duty exemptions. This happened before accession to the EU by Romania. Subsequently, close to the moment of accession (in 2004), Romania repealed the said exemptions, as a compliance mechanism with EU State Aid rules. On the basis of this measure, the Micula brothers challenged the measure in an arbitral tribunal under the relevant IT.

During proceedings, the EU Commission intervened as amicus curiae, effectively arguing against any reinstatement of the exemptions since that would amount to illegal State Aid. However, the arbitral tribunal ignored the Commission’s arguments and found against Romania, effectively ruling that a breach of the claimants’ legitimate expectations occurred, awarding damages.[25] As such, the claimants sought the enforcement of the arbitral decision in Romanian courts. They succeeded although the Commission had once again intervened looking to oppose the enforcement. This led the Commission to open a formal investigation into what they argued could constitute illegal State Aid. It was eventually decided by the Commission that the enforcement of the award (the payment of compensation) constituted illegal State Aid[26] and this bore upon Romania the obligation to recover the awarded compensation.[27] Moreover, the Micula brothers challenged the Commission’s decision in the CJEU, the case pending before the Court at the moment (case T-646/14).[28]

The entire Micula case complicates the matters. In order to better analyse the whole issue, two elements should be separated from the facts: on the one hand, there is the issue of the initial exemption, itself. Romania effectively considered the initial exemption to be illegal State Aid and, as a consequence, repealed it. Moreover, it seems that a formal analysis into the whole exemption leads to the same conclusion: this is an act imputable to the State, which offers the beneficiary an economic advantage, distorts competition and can affect trade between Member states. What complicates matters is that the relevant BIT, on the other hand, protects the legitimate expectations, not a specific exemption such as the custom duties exemption offered by Romania. Of course, once a specific benefit offered to the investors generated legitimate expectations, the standard of legitimate expectations set by the BIT becomes applicable and enforceable. However, there is a difference between the exemption and the obligation to guarantee legitimate expectations.

But the questions which have to be addressed now are: firstly, at the moment of the granting of the exemption, which was fairly close to the moment of Romania’s accession to the EU, were the exemptions granted by the Romanian State to be considered as having generated legitimate expectations? And secondly, what is the basis for ignoring the investment tribunal’s award, by the European Commission? I have already mentioned that I do not consider the EU Legal Regime as being totally autonomous. Thus, it is not separated from the framework of general Public International Law. It is just a system which, from a Public International Law point of view, is in (apparent) conflict with another system: that of Foreign Investment Law. Thus, which is to be considered as having primacy and why?

I will now address the first question: I argue that legitimate expectations indeed existed. I base my claim on two elements: firstly, the arbitral tribunal’s decision to award compensation to the Micula brothers, as a result of the repealing of the custom duties incentive scheme[29]. Secondly, it has been found in case-law regarding Foreign Investment that if a benefit awarded by a State to an investor was presented, by representations made to the investor, as having been in compliance with the legal requirements of the host state, the investment must be awarded the expected degree of protection (by respecting the awarded benefits on which the investor relied), even if, in reality it conflicts with the host state’s law.[30] It is true that in the previously-mentioned case the claimant relied on the principle of estoppel[31], which is a specific application of the legitimate expectations doctrine[32], but the principles applied in the Kardassopoulos v. Georgia case clarify the issue for the Micula case, as well.

Thus, it is not for the investor to bear the risk of an investment which is non-compliant with the legal rules of the host state, when the host state created the expectation of conformity. And this is what happened in the Micula case, because the investment had been protected for a few years, before the repealing of the incentive scheme, creating the proper expectations of legitimacy and legality. Therefore, legitimate expectations existed and were violated by Romania through its repealing act.

Subsequently, there is the issue of analysing the (apparent) conflict between obligations arising under BITs and ones arising under EU law. As mentioned earlier, those are two conflicting legal systems with no apparent hierarchy between them (neither the obligations under foreign investment law nor the ones under EU State Aid law can be considered jus cogens – norms of a peremptory character under International Law, from which no derogation is admitted; in other words, norms of a superior value). Therefore, because those norms are considered to be of equal value, the (apparent) conflict must be settled by relying on the principle of lex posterior derogat (legi) priori.[33]

Under this framework, it can be argued that the subsequent legal regime implemented by the EU State Aid legal regime would derogate from schemes of custom duties exemption such as the one presented earlier. Such exemptions constitute State Aid and since both Romania and Sweden are part of the EU Legal Regime, it can be considered that they have derogated from the possibility to implement such State Aid. As such, if there had been an obligation under the ITs to grant such an economic advantage to the investors, there could have been a real conflict between the ITs and the EU State Aid rules.

However, there did not exist any such obligation. And this is why I mentioned earlier that there was just an apparent conflict between those two international legal regimes – the EU legal regime and the Foreign Investment legal regime – and not a real one. The obligations under ITs are not conflicting with the EU State Aid Rules. What is conflicting is the effective benefit given by the Romanian State to the Micula brothers. While this is inherently linked to the IT, it is not identified with it. This may be a nuance, but it is an important one. It demonstrates that, at least as to the relationship between obligations arising under intra-EU ITs and EU State Aid rules there is no formal conflict.

Thus, this custom duties exemption is a different thing to the protection of legitimate expectations – the ones which are actually protected by the Romanian – Swedish BIT –, expectations which had been created before the EU laws prohibiting State Aid became effective. The legitimate expectations, as mentioned earlier, must be protected, even more so when the investors acted in good faith by relying on the representations of the Romanian officials.

Moreover, it has been proved that payment of compensation as a result of an investment tribunal’s award is not a form of State Aid, because it is neither unilateral nor autonomous, as needed for measure to be imputable to the State and, as such, to be considered State Aid. Therefore, the answer to the second question must be that, while the custom duties incentive scheme constitutes State Aid, the protection of legitimate expectations – through granting compensation –, especially when they were created at a moment when the conflicting rules on State Aid had not been effective, is not State Aid.

As such, there is no obligation for Romania, neither under International Law nor under EU law to recover the paid compensation. However, there are two authors who argue that a compensation rendered as an enforcement of an arbitration award under Foreign Investment Law can constitute a violation of Article 107 of the TFEU if the action leading to an obligation to compensate consists of repealing benefits that are themselves illegal state aid under Article 107 of the TFEU[34]. While I do agree, in principle, with the authors and the opinion of Advocate General Ruiz-Jarabo Colomer in Joined Cases C-346/03 and C-529/03 that if an entitlement to compensation is recognized, the damage cannot be regarded as being equal to the sum of amounts to be repaid, since this would constitute an indirect grant of the aid found to be illegal and incompatible with the common market[35], I cannot agree that this is applicable always, as an absolute rule.

The issue should be assessed on a case by case basis. The authors’ and the Advocate General’s statements do not take into account the legitimate expectations created to the investor. And from an Investment Law point of view, not only does the investor have locus standi under arbitral proceedings in Foreign Investment proceedings, but he is considered a subject of Public International Law[36]. As such, who is to make a hierarchy between the investor’s interests – the protection of his legitimate interests – and those of another subject of Public International Law, the EU – where the fundamental interest is that of the effectiveness of the Internal Market, through a proper competition framework which underpins the functional trade between the EU Member States? From a public international law perspective, the EC legal system remains a subsystem of international law.[37]

Thus, I find such a hierarchy between a subject of Public International Law within the Investment Law field, on the one hand, and the officials of the EU, on the other hand, arbitrary and in violation of basic principles of Public International Law. Moreover, as has been stated somewhere else regarding a similar issue of EU law: just because something is mentioned repeatedly does not turn it into reality.[38]

(c) Notwithstanding specific issues whether enforcement of an investment arbitration award can be considered illegal State Aid, is it justified to ever argue for the termination of an intra-EU IT on the basis of that treaty violating the rules on State Aid prohibition?

Can the intra-EU IT, by itself, be considered as violating EU rules on State Aid? The answer, in my view, is in the negative. This is because investment treaties govern issues such as what constitutes an investment[39], admission of investments[40] or what protection does an investor receive once an investment has been made[41]. In other words, such ITs govern the abstract rules applicable to all investments, not referring to a certain specific investment (of course, the specific investment will benefit from the protection, but on the basis of fitting the framework set by the treaty, not by other means). As such, there is no obligation, ipso jure, to grant an economic advantage which can be considered State Aid under EU rules. The choice to grant that advantage is an economic/ political choice of the State, not an obligation under ITs.

Therefore, it can be concluded that investment treaties are not prohibited under State Aid rules. However, this doesn’t render the issue of the validity of intra-EU Bilateral Investment Treaties obsolete. On the contrary, this is a different discussion, which takes into consideration the common trade policy set by the 2009 Lisbon Treaty[42] and various issues such as questions of jurisdiction of investment tribunals when the parties are an EU Member and an investor from another EU Member[43]. Nonetheless, this is a different issue, going beyond the scope of the present analysis.[44]

  1. What does the relationship between State Aid law and obligations arising under ITs entail when the ITs’ signatories are both from within the EU and from without the EU:

This issue should be easier to analyse, given all that has been presented until now. One of the main problems with the relationship between the obligations arising under intra-EU ITs, on the one hand, and EU State Aid rules, on the other hand, was that the parties to the TFEU were, at the same time, parties to those intra-EU ITs. This situation complicated matters because the conflict was noticeable (although from a different perspective than the one of this analysis – that of the validity of intra-EU ITs). In the case of EU State Aid rules and obligations arising under ITs concluded with third parties, the simple answer is that, because the third party is not a party to the EU, the rules on State Aid are not opposable to it.[45] This does not mean that the EU party granting benefits to an investor from a third State, after the TFEU became effective, is not liable for violations of Article 107. But, at the same time, this will not have any bearing on the earlier obligations arising under the ITs. In that case, if the EU member state decides to repeal any State Aid benefit, it may be in compliance with EU rules on State Aid, but its responsibility will be engaged under customary international law for violating a legal obligation arising under ITs. As a consequence, this breach of international obligations gives rise to an obligation of reparation[46], which does not constitute State Aid – as has been proven in the first part.

III. Conclusion:

The debate over whether there exists a conflict between the legal regime instituted by ITs (excluding those where the EU is a party), on the one hand, and the EU’s legal regime, on the other, is neither straightforward nor devoid of political and economic implications. Through this study, I have analysed a part of this debate: the relationship between ITs, on the one hand, and the legal regime of State Aid law, on the other. I demonstrated, firstly, that obligations specifically arising under ITs are not, by themselves, in conflict with State Aid rules, because there is no ratione materiae identity.

In this context, I made a differentiation between the measures which can be considered illegal State Aid and the ITs (and their provisions such as the ones related to the protection of legitimate expectations), which, although inherently linked to such measures (such as in the Micula case) are, in the end, different. Continuing, I demonstrated why something which tends to be considered an absolute truth – the supremacy of EU law – must be qualified in the international sphere: there is no legal basis under Public International Law to consider such a supremacy when the EU legal regime is in conflict with other international legal regimes. And, finally, I analysed the situation where an IT has signatories both from within and without the EU There, I made a clear differentiation between what can amount to liability of an EU State for violations of Article 107 TFEU and responsibility of the same State under the Customary Law on State Responsibility for violations of obligations contained within ITs. I have shown how an EU State can infringe both legal regimes, at the same time, and why the EU legal regime is relative (and opposable) to the EU States only.


[1]Hereinafter referred to as State Aid law

[2] Hereinafter referred to as IT

[3] For a critique of the concept of self-contained regimes (the idea that supranational or international regimes, such as the EU, are self-contained and cannot be influenced by Public International Law rules, such as Treaty Law or the Law on State Responsibility), see Bruno Simma and Dirk Pulkowski, Of Planets and the Universe: Self-Contained Regimes in International Law, The European Journal of International Law, Vol. 17, no. 3, 2006, hereinafter cited as Simma and Pulkowski. This aspect of mutual influence between those regimes is of utmost importance to the present study, since the starting premise of the present study is that there is a mutual link between semi-autonomous regimes – such as the EU –, on one hand, and general Public International Law rules, on the other. See, for the opposite view (that the EU Legal Regime is an autonomous legal order which is not influenced by Public International Law), Laurens Ankersmit, Is ISDS in EU Trade Agreements Legal under EU law?, https://www.iisd.org/itn/2016/02/29/is-isds-in-eu-trade-agreements-legal-under-eu-law-laurens-ankersmit/ (last visited on 10/02/2018, at 20:58)

[4] Kelyn Bacon, BIT arbitration awards and State aid – the Commission’s Micula decision, http://uksala.org/bit-arbitration-awards-and-state-aid-the-commissions-micula-decision/ (last visited on 10/02/2018, at 20:53)

[5] Christian Tietje, Clemens Wackernagel, Outlawing Compliance? – The Enforcement of intra-EU Investment Awards and EU State Aid law, Policy Papers on Transnational Economic Law, June 2014, p. 2, hereinafter cited as Tietje, Wackernagel

[6] Although a thorough analysis should begin with what constitutes an undertaking, under EU law – since those entities are the beneficiaries of state aid –, I will not undertake such an analysis, for reasons of brevity. Thus, the analysis is considered to refer, implicitly, to such elements.

[7] Hereinafter, referred to as The TFEU

[8] EC (European Commission): Communication from the Commission: Draft Commission Notice on the notion of State aid pursuant to Article 107 (1) TFEU, § 5, p. 4

[9] Tamás Kende, Arbitral Awards Classified as State Aid under European Union Law, ELTE Law Journal 2015/1, p. 40

[10] http://ec.europa.eu/competition/state_aid/overview/index_en.html (last visited on 10/02/2018, at 02:55)

[11] It is true that the situation varies from one case to the other, but I have decided to begin the analysis by qualifying an investment award in abstracto, in order to assess its legality under EU State Aid law and only afterwards I shall address specific cases.

[12] Bryan A. Garner (Editor in Chief), Black’s Law Dictionary. Ninth Edition, WEST. A Thomson Reuters business, St. Paul, MN, USA, 2009, p. 157

[13] Idem, p. 608

[14] The International Centre for Settlement of Investment Disputes

[15] Hereinafter referred to as The ICSID Convention

[16] Article 53 of The ICSID Convention

[17] Hereinafter referred to as The New York Convention

[18] Article III of The New York Convention

[19] According to Article 26 of the 1969 Vienna Convention on the Law of the Treaties, every treaty in force is binding upon the parties to it and must be performed by them in good faith.

[20] Case 61/79, Amministrazione delle finanze del- lo Stato v Denkavit italiana [1980] ECR 1205, § 31

[21] Under Public International Law, the sources of legal obligations can include unilateral conduct or general principles of law, as well (see, for a comprehensive analysis, Malcolm N. Shaw, International Law. Seventh Edition, Cambridge, United Kingdom, 2014, Chapter 3: Sources, pp. 49-91). But, for the present study, this is not important, since obligations arising in Investment Law are based mostly on investment treaties (more specifically, obligations to respect and enforce arbitral awards are the relevant ones for the present analysis), while those based on general principles of law cannot be considered to have appeared from a consensual relationship between the parties to a dispute in a specific dispute. Anyway, from a strictly technical point of view, no matter the source of obligation, the ensuing legal relationship is, in the end, always (at least) bilateral (the correlative existence of the right and of the duty): Arthur L. Corbin, Rights and Duties, 33 Yale Law Journal 501, 1923-1924, p. 502. But what is important, as a bottom-line, is that an arbitral award (and the subsequent obligation of enforcement) arose under a legal relationship outside the scope of the State’s discretionary powers.

[22] Tietje, Wackernagel, p. 7

[23] Sir Hersch Lauterpacht, The Function of Law in the International Community, Oxford University Press, Oxford, United Kingdom, 2011, pp. 3-4

[24] For example, see Case 26/62, Van Gend en Loos [1963], p. 2, § 2 and Case 6/64, Costa v. Enel [1964], p. 594

[25] Ioan Micula, Viorel Micula and others v. Romania: Final Award (ICSID Case No. ARB/05/20)

[26] Article 1 of Commission Decision (EU) 2015/1470 of 30 March 2015 on State aid […] implemented by Romania – Arbitral award Micula v. Romania of 11 December 2013

[27] All the factual information regarding the Micula affair mentioned so far has been gathered from: Kelyn Bacon, BIT arbitration awards and State aid – the Commission’s Micula decision, http://uksala.org/bit-arbitration-awards-and-state-aid-the-commissions-micula-decision/ (last visited on 10/02/2018, at 20:53)

[28] Volterra Fietta, Further attempts by the European Commission to eradicate intra-EU BITs, https://www.volterrafietta.com/further-attempts-by-the-european-commission-to-eradicate-intra-eu-bits/ (last visited on 10/02/2018, at 20:54)

[29] Ioan Micula, Viorel Micula and others v. Romania: Final Award (ICSID Case No. ARB/05/20)

[30] Ioannis Kardassopoulos v. Georgia: Decision on Jurisdiction (ICSID Case no. ARB/05/18, §§ 191-192)

[31] Andreas Kulick, About the Order of Cart and Horse, Among Other Things: Estoppel in the Jurisprudence of International Investment Arbitration Tribunals, The European Journal of International Law, Vol. 27, no. 1, p. 119

[32] I will not get into a detailed discussion of what constitutes estoppel and what is the difference between it and other institutions, such as the one of legitimate expectations. For a detailed analysis of estoppel in International Law and its application by the International Court of Justice, see Alexander Ovchar, Estoppel in the Jurisprudence of the ICJ. A principle promoting stability threatens to undermine it, Bond Law Review, Volume 21, Issue 1.

[33] http://www.oxfordreference.com/view/10.1093/acref/9780195369380.001.0001/acref-9780195369380-e-1282 (last visited on 10/02/2018, at 20:55)

[34] Tietje, Wackernagel, p. 7

[35] Joined  Cases C-346/03 and C-529/03. Opinion of Advocate General Ruiz-Jarabo Colomier, delivered on 28 April 2005, § 198

[36] See Robert McCorquodale, The Individual and the International Law Legal System, in Malcolm D. Evans (ed.), International Law. First Edition, Oxford University Press, Oxford, UK, 2003, pp. 299, 311-314 and 321-322

[37] Simma and Pulkowski, p. 516

[38] Ibid.

[39] Matthias Herdegen, Principles of International Economic Law. Second Edition, Oxford University Press,  Oxford, United Kingdom, 2016, pp. 444-446

[40] Idem, pp. 448-450

[41] Idem, pp. 448-477

[42] Francesco Montanaro and Sophia Paulini, United in Mixity? The Future of the EU Common Commercial Policy in light of the CJEU’s recent case law, EJIL: Talk! Blog, https://www.ejiltalk.org/united-in-mixity-the-future-of-the-eu-common-commercial-policy-in-light-of-the-cjeus-recent-case-law/ (last visited on 10/02/2018, at 20:55)

[43] Emanuela Matei, The love-hate story of arbitral jurisdiction  over claims against states in the EU, EFILA Blog, https://efilablog.org/2016/10/25/the-love-hate-story-of-arbitral-jurisdiction-over-claims-against-states-in-the-eu/ (last visited on 10/02/2018, at 20:56)

[44] For an analysis of this aspect, see: Nikos Lavranos, The Lack of Any Legal Conflict Between EU law and intra-EU BITs/ECT Disputes, EFILA Blog, 25 February 2016, https://efilablog.org/2016/02/25/the-lack-of-any-legal-conflict-between-eu-law-and-intra-eu-bitsect-disputes/ (last visited on 10/02/2018, at 20:55)

[45] The 1969 Vienna Convention on the Law of the Treaties provides, in Article 30 (3.) (b) that when the parties to a later treaty (as is the case when various ITs had existed before the TFEU became effective) do not include all the parties to the earlier one […] as between a State party to both treaties and a State Party to only one of the treaties, the treaty to which both States are parties governs their mutual rights and obligations.

[46] See Article 31 of the ILC Articles on State Responsibility