The CETA ICS and the Autonomy of the EU Legal Order in Opinion 1/17 – A Compass for the Future

Abstract

In April 2019, the Court of Justice of the EU (‘CJEU’) handed down its Opinion (C-1/17) on the compatibility of the Investment Court System (‘ICS’), that is the Investor-State Dispute Settlement (‘ISDS’) mechanism under the EU-Canada Comprehensive and Economic Trade Agreement (‘CETA’), with EU law. This article puts Opinion 1/17 in its broader (policy and legal) context, focusing on the salient issue of compatibility with the principle of autonomy of the EU legal order. It argues that the Court's openness to this judicial competitor was an acknowledgment of the need to maintain the powers of the Union in international relations. However, Opinion 1/17 should not be perceived as an automatic green light for any future investment court (such as the Multilateral Investment Court) as the autonomy test it introduces is a rather difficult one to pass.

I. INTRODUCTION

It took several years for the Court of Justice of the European Union (‘CJEU’ or the ‘Court’) to declare an external system of judicial review compatible with EU primary law, and in particular with the principle of principles, the autonomy of the EU legal order. Previously, the Court had not displayed much tolerance towards judicial competition, even though it has had a handful of opportunities in a number of different contexts. Such tolerance was not displayed, for example, in the case of the European Patent Convention creating a European and Communities Patent Court.¹ Or, rather disappointingly, in the case of the draft agreement providing for the accession of the EU to the European Convention on Human Rights (‘ECHR’),² a much-criticised example in the Court's record.³ Instead, in its recent Opinion 1/17,⁴ the Court decided to give a green (albeit very cautious) light to the Investment Court System (‘ICS’), that is the Investor-State Dispute Settlement (‘ISDS’) mechanism provided in a bilateral mixed agreement, the EU-Canada Comprehensive and Economic Trade Agreement (‘CETA’).⁵

The timing of this green light is interesting in many ways. Firstly, Opinion 1/17 was rendered a little more than a year after the Court's landmark judgment in Achmea,⁶ a case that concerned the compatibility of the ISDS clause provided in an intra-EU Bilateral Investment Treaty (‘BIT’) with EU law. Despite the different contexts (an intra-Member State BIT on the one hand and an agreement concluded by the EU on the other), the extra-EU relevance of this intra-EU case is significant.

Secondly, Opinion 1/17 was rendered almost a decade after the entry into force of the Treaty of Lisbon,⁷ that is when exclusive competence in common commercial policy (‘CCP’), including foreign direct investment (‘FDI’), was conferred on the EU, and slightly less than two years after the Court opined on the contours of this competence (Opinion 2/15).⁸ Notably, over the course of this decade, the EU emerged as a significant actor in the investment field.⁹ In the context of its bilateral negotiations with several of its trade partners, the EU introduced the idea of establishing a permanent investment court to replace the traditional ad hoc ISDS.¹⁰ Tellingly, this court proposal soon moved beyond the limits of these bilateral negotiations. The prospect of a Multilateral Investment Court (‘MIC’) is now part of the reform agenda discussed under the auspices of the United Nations Commission on International Trade Law (‘UNCITRAL’).¹¹

Thirdly, Opinion 1/17 comes at a time when the traditional ad hoc ISDS (in the form of investor-State arbitration) is subject to rampant public criticism, while debates on its reform are ongoing in different forums. Of course, the so-called backlash against traditional ISDS is not a new phenomenon.¹² Briefly, it encapsulates critiques concerning the independence and impartiality of the arbitral tribunals, the lack of systemic checks and a second degree of review of arbitral awards, as well as an alleged dearth of transparency.¹³ All these critiques often reflect a widespread perception that ad hoc ISDS is an inherently pro-investor system and, ultimately, they relate to the concerns that investor-State awards may negatively impact the states’ regulatory powers.¹⁴ These critiques are mirrored in (and to some extent have also been legitimised by) the reform solutions proposed by the EU.

Against this background, Opinion 1/17 forms part of the discussion of the impact that the EU internal (constitutional) order, and more precisely the principle of autonomy, may have on the EU's participation in international dispute resolution mechanisms. This article is structured as follows. First, for context purposes, some light is shed on the legal and policy landscape that led to CETA (Section II). The Court's reasoning can be better understood when having in mind the determined way in which the EU exercised its new competence and shaped its policy towards the creation of a permanent investment court. Subsequently, and among the various compatibility questions raised in Opinion 1/17, the focus of this article is on the salient issue of compatibility with the autonomy of the EU legal order (Section III). Accordingly, the article explores certain aspects of the Court's reasoning and dissects the Opinion 1/17 autonomy test. The potential implications of the ‘compatibility check-list/conditions’ for other extra-EU investment agreements are examined (Section IV). The final section presents a conclusion (Section V).

II. A COMPLEX LEGAL AND POLICY PATH TOWARDS A ‘NO RETURN TO OLD ISDS’ EU POLICY

‘Does Europe not, now that is finally unified, have a leading role to play in a new world order, that of a power able both to play a stabilizing role worldwide and to point the way ahead for many countries and people?’ ¹⁵

Every time the EU Members States amended the Treaties, they granted more external powers to the EU, enhancing its growing aspirations to emerge as a global actor that points the way ahead. The Lisbon Treaty incorporates the most notable expansion of EU external powers to date as exclusive competence in FDI was conferred on the Union (Article 207 TFEU). This new competence offered the legal channel for the emergence of the EU as a rule-shaper in the context of FDI and, controversially, ISDS. However, the nature and the scope of this new competence were not straightforward. The issue made its way to Luxembourg following a request made by the European Commission under Article 218(11) TFEU.¹⁶ A few quick observations on Opinion 2/15 are necessary for context purposes, taking into account that the Court's analysis in Opinion 1/17 ‘[began] … where the Court left off in its Opinion 2/15’.¹⁷

A. Shared Competence over ISDS (Opinion 2/15)

Briefly, the Court was asked to answer the question whether the Union had the required competence to sign and conclude alone the EU-Singapore Free Trade Agreement (‘EUSFTA’, one of the first new generation ‘deep and comprehensive’ trade deals which were bringing together trade and investment provisions), and what the nature of such competence was (exclusive, shared, no competence) depending on each type of provision in the agreement.¹⁸ According to the Commission's broad interpretation, its exclusive competence was covering the investment chapter of the EUSFTA, including the ISDS mechanism therein.¹⁹ By contrast, certain Member States and the Council took the view that this was a matter of shared competence.

The Court, after emphatically highlighting that the Opinion was rendered without prejudice to the question of the compatibility of said ISDS mechanism,²⁰ left no doubt as to the exclusiveness of the EU's competence in relation to FDI²¹ but found that ISDS is not part of the EU's exclusive competence.²²

One would expect that the procedural provisions of an agreement should have the same treatment as the substantive ones since they are of an ‘ancillary nature’.²³ To make this clearer, the general rule is that dispute resolution mechanisms are to secure compliance with substantive obligations. Therefore, they have the same legal basis as the substantive obligations and follow the allocation of competence.²⁴ Nevertheless, the Court departed from this general rule finding that competence in relation to ISDS is shared,²⁵ and this is because, notably, the ISDS is not of a purely ancillary nature.²⁶

The Court's reasoning on why ISDS is not ancillary to the substantive investment obligations is very opaque. In the relevant and short but still revelatory passage, the Court's vision of the EU judicial system emerges once again. It is worthy of being recorded in full:

Such a regime, which removes disputes from the jurisdiction of the courts of the Member States, cannot be of a purely ancillary nature … and cannot, therefore, be established without the Member States’ consent.²⁷

By deciding that ISDS provisions are a shared competence,²⁸ the CJEU did not make the Commission's negotiations any easier.²⁹ Despite having proclaimed that the question of competence is a preliminary issue that cannot be influenced by political or other similar considerations,³⁰ it seems that in this case, the CJEU did not disregard the noise surrounding these all-inclusive trade deals. The Court implicitly invited the Commission to make these trade deals less complicated, perhaps further implying that investment deals (and mainly their ISDS provisions) are politically sensitive. In light of the Court's Opinion, if all these new-era trade deals were to include an ISDS chapter, they would have to be concluded as mixed agreements, unless Member States make a political choice and offer their consent in order for the EU to act alone.³¹

The Council ‘took note’ of the Court's Opinion.³² Indeed, the EUSFTA, as well as other agreements (but not CETA), are now split into two parts: one FTA and one Investment Protection Agreement (‘IPA’).³³

With the scope of the EU FDI competence in mind and the Court's decision to exclude ISDS, we shall now move on to explore the determined way in which the EU has been developing its investment/ISDS policy.

B. A Matrix of Parallel and Intermingling Negotiations

In the early exploratory days of its new competence, the EU seemed willing to work out its participation in the existing ad hoc regime.³⁴ In fact, it had expressed its intention to explore the possibility of accession to the ICSID Convention³⁵ and thus include ICSID arbitration among the available ISDS alternatives in its International Investment Agreements (IIAs).³⁶ In one of the first documents outlining the new investment policy, the European Commission made two interesting acknowledgements: first, that not being a member of ICSID was the main obstacle to opting in to this regime,³⁷ and second, that the EU ‘has not historically been a significant actor in this field’. It then observed that the ‘current structures [were] to some extent ill-adapted to the advent of the Union’.³⁸

These realisations, combined with the backlash against investor-State arbitration, made opting for traditional ISDS clauses almost a non-option.³⁹ The EU had to come up with a new ‘modern’ ISDS policy. The gradual shift was reflected in the parallel and intermingling negotiations of the various international trade agreements the Commission has been having with trade partners around the globe.

1. Creating the CETA ICS through TTIP negotiations

In September 2015, the Commission released a draft text in the context of the Transatlantic Trade and Investment Partnership (‘TTIP’) negotiations aiming to make publicly known its positions. Interestingly, this document was originally almost a non-paper since it came with the warning that ‘this is not a formal text proposal … but an internal document of the EU’.⁴⁰ Only a few months later, the EU made public a slightly edited version which then became officially ‘tabled for discussion’ with the United States (‘US’).⁴¹ Although the TTIP Proposal had been tabled for discussion for quite some time, there had been no official reaction on behalf of the US. In light of the change of administration in 2017, it soon became clear that TTIP would remain an exercise on paper. Of course, although the reasons why the TTIP negotiations did not lead to a deal are political, it is certain that, under any administration, ISDS would be one of the points on which parties would not easily reach an agreement.⁴²

When we examine the shaping of the EU investment policy, the TTIP Proposal still remains a point of reference for a number of reasons. It was the first text in which the new EU court-based approach was crystallised and then it was transplanted to all subsequent trade deals, which all include very similar provisions in their majority. In a way, the TTIP Proposal has played informally and implicitly the same role as an EU Model BIT with the EU showing to the rest of the world what its starting point in future negotiations will be.⁴³ The Proposal was also an indication of the EU's determination and commitment to this project: being aware that the US was not positively leaning towards abandoning the crystallised ad hoc system,⁴⁴ it still did not hesitate to make it part of the deal.

The Council of the EU had unanimously authorised the European Commission to negotiate TTIP (including the adoption of a new ISDS mechanism) two years before the release of the TTIP Proposal, in June 2013.⁴⁵ According to the relevant Council mandate, it was made clear that ‘the inclusion of investment protection and [ISDS] will depend on whether a satisfactory solution … is achieved’.⁴⁶ The ‘satisfactory’ nature of the solution depended inter alia on the consultation with Member States and its compatibility with EU law. In addition, the Council went further and provided some more specific features of the envisaged enforcement mechanism. Indicatively, the Council asked for an agreement aiming to ‘provide for an effective and state-of-the-art [ISDS] mechanism, providing for transparency, independence of arbitrators and predictability of the Agreement …. It should provide for investors as wide a range of arbitration fora as is currently available under the Member States’ bilateral investment agreements’. The Council also requested that ‘[c]onsideration should be given to the possibility of creating an appellate mechanism applicable to investor-to-state dispute settlement’.⁴⁷ The CETA Council Directives to the Commission were along the same lines.⁴⁸

From its end, the European Parliament in its recommendations to the European Commission on the negotiations for TTIP invited the Commission ‘to ensure that agreement on any dispute settlement mechanism regarding investment protection … explicitly state[s] the Member States’ right to regulate and under no circumstances restrict or hinder legislators from passing and enforcing laws both in the area of employment and in the area of social policy for their countries; … the inclusion of any form of private arbitration courts in TTIP must be ruled out’.⁴⁹

CETA negotiations were concluded in August 2014, and the then agreement was providing for a traditional ISDS mechanism. However, in the aftermath of the finalisation of another deal, the EU-Vietnam Free Trade Agreement (‘FTA’),⁵⁰ and the new mechanism introduced in the (then still pending) TTIP negotiations, in January 2016, the EU invited the Canadian federal government to revisit the ISDS chapter of the agreement. Almost immediately, the parties revisited the ISDS chapter and agreed to implement the new court approach to ISDS.⁵¹

The CETA ICS is envisaged as a two-tier permanent body,⁵² comprising a Tribunal of First Instance and an Appeal Tribunal.⁵³ Provisions on the composition of these bodies (including the number of members, the appointment process, term, retainers, and availability), the qualifications and ethics of their members, and a few procedural rules (including transparency of the proceedings) are also included.

The little time that was required for Canada and the EU to agree upon the new ISDS approach leaves little doubt as to the role of TTIP and the EU's strategy. Notably, Canada might have agreed to the ICS in CETA but has shown no intention to depart from the traditional arbitration solution in other instances, such as in the Comprehensive and Progressive Agreement for Trans-Pacific Partnership (‘CPTPP’).⁵⁴ On the contrary, the EU has made the ICS part of the deals it has been negotiating with Mexico⁵⁵ and other partners.⁵⁶

Representatives from both the Canadian government and the European Commission had expressed their confidence that CETA would enter into force in 2017.⁵⁷ This proved to be a rather optimistic expectation, especially after the ‘Walloon’ problem arose, leading to Opinion 1/17.

2. From Wallonia to Luxembourg

The path towards the signature of CETA (30 October 2016) was not an easy one to tread and it illustrates the difficulties that the EU faces in putting forward these mega-deals. By way of a reminder, at the time (ie before Opinion 2/15), the Commission was still arguing that it had exclusive competence to sign CETA alone as an EU-only agreement.⁵⁸ Nevertheless, in light of the reactions from a number of Member States,⁵⁹ it altered its stance. In July 2016, the Commission submitted a proposal to the Council according to which the EU would treat CETA as a mixed agreement enjoying provisional application until its entry into force.⁶⁰

The Walloon problem arose immediately after. The parliament of a Belgian region, Wallonia, rejected CETA's signature, expressing concerns about threats relating to beef imports and the ISDS provisions. This led to a series of dramatic moments marked by political and diplomatic negotiations between the Belgian federal government and the Commission on the one hand and the regional government in Wallonia on the other. Quite indicative of the climate were the statements made by the Canadian trade minister following Wallonia's veto:

It seems obvious that the EU is now not capable of having an international agreement, even with a country that shares European values such as Canada, even with a country that is so kind and patient. Canada is disappointed. I am personally very disappointed.⁶¹

The Belgian government managed to reach an agreement with Wallonia. Accordingly, the signature of CETA would be approved in Wallonia but Belgium would instantly submit a request to the CJEU for an Opinion on the compatibility of CETA with EU law. Indeed, in September 2017, Belgium kept its promise⁶² and CETA made its way to Luxembourg.

C. From Several Bilateral Courts to a Multilateral Investment Court (‘MIC’)

The CETA ICS is ‘only the first stage’,⁶³ an interim measure in the ambitious EU policy. The EU had previously foreshadowed that having multiple bilateral permanent courts (one under each specific agreement) would not be a viable solution.⁶⁴ Instead, the ultimate plan has always been the establishment of an MIC⁶⁵ for all its trade partners but also with a view to having individual EU Member States opt in to the MIC in their extra-EU BITs.⁶⁶

The envisaged transition from multiple bilateral courts to one multilateral court has been reflected in the provisions of the various agreements. For instance, as stipulated in CETA:

The Parties shall pursue with other trading partners the establishment of a multilateral investment tribunal and appellate mechanism for the resolution of investment disputes. Upon establishment of such a multilateral mechanism, the CETA Joint Committee shall adopt a decision providing that investment disputes under this Section will be decided pursuant to the multilateral mechanism and make appropriate transitional arrangements.⁶⁷

In addition, in the Joint Interpretative Instrument, the parties have also clearly stipulated that CETA ‘lays the basis for a multilateral effort to develop further this new approach to [ISDS] into a [MIC]’ and that they ‘will work expeditiously’ towards its creation.⁶⁸

Notably, the idea for the establishment of an investment court is not new. Indicatively, as far back as 1948 the International Law Association (‘ILA’) published the Draft Statutes of the Arbitral Tribunal for Foreign Investment,⁶⁹ whilst in the 1960s the alternative of a permanent court was also discussed and juxtaposed to that of ad hoc arbitration.⁷⁰ The element of permanence is also present in the establishment of the Iran-US Claims Tribunal.⁷¹ Since then, it has been part of the discussions on reforming the system in different contexts and forums, including the UN Conference on Trade and Development (‘UNCTAD’),⁷² the World Trade Organization (‘WTO’),⁷³ and the Organisation for Economic Co-operation and Development (‘OECD’).⁷⁴ Its resurgence, however, should be credited to the EU's persuasiveness as a global actor (and the world's largest exporter and importer of FDI).

In 2017, UNCITRAL initiated its work on ‘ISDS Reform’, entrusting the Working Group-III (‘WG-III’) with the relevant mandate.⁷⁵ The EU is not a member of UNCITRAL (but all its Member States are).⁷⁶ It has the status of an observer relying on its Member States’ duty of loyalty.⁷⁷ As noted in the Council Negotiating Directives, the Union and its Member States ‘shall fully coordinate positions and act accordingly throughout the negotiations’.⁷⁸

With this background in mind, and particularly the leading role that the EU has played in the global reform initiatives, we shall now analyse in more detail the Court's approach in Opinion 1/17.

III. THE SALIENT ‘AUTONOMY OF THE EU LEGAL ORDER’: UNRAVELLING THE OPINION 1/17 TEST

EU law is not simply the law of the Union. It is the law that makes the Union. The principle of autonomy, as construed by the Court, thus serves existential purposes. It is the glue that makes the Union a legal order, a rules-based self-contained regime. As such, it vindicates its own claim of primacy vis-à-vis the other structural principles emerging as the principle of principles.⁷⁹ Premised on Articles 19 TEU, and 267 and 344 TFEU, the principle of autonomy grants the Court the exclusive right to ensure that EU law is observed. This can be translated into the Court's exclusive right to determine questions of competence between the EU and its Member States, the legality of EU law,⁸⁰ and the relative powers and functions of the institutions.

Against this background, the salience of the autonomy of the EU legal order in the context of dispute resolution mechanisms in international agreements that the EU and/or its Member States conclude can hardly come as a surprise. It also explains why this article has focused on this aspect of the Opinion. In passing, in addition to the autonomy concerns (the question of the compatibility with the exclusive jurisdiction of the CJEU and the autonomy of the EU legal order),⁸¹ Belgium also raised concerns about the compatibility with the principle of equal treatment⁸² and the requirement of effectiveness of EU law (discrimination and effectiveness concerns),⁸³ as well as the right of access to an independent and impartial tribunal (independence concerns).⁸⁴ The Court examined all these concerns and found the CETA ICS compatible with EU law.⁸⁵

In what follows, after we glimpse the EU dogma of autonomy as recapped didactically by the Court in Opinion 1/17, we shall move on to unravel and critically discuss the two-step autonomy test it introduces.

A. A Systematic Recap of the EU Dogma

The Court, unsurprisingly and according to its usual analytical, almost academic approach, began its analysis with a general section recalling the general principles.⁸⁶ The Court reiterated the presumption of, in principle, compatibility for all international agreements providing for the creation of a court ‘responsible for the interpretation of its own provisions’.⁸⁷ The CETA court (and, eventually, the MIC)⁸⁸ may be compatible with EU law only if it has no adverse effect on the autonomy of the EU legal order.⁸⁹ The Court then devoted two paragraphs to listing the ingredients of this autonomy.⁹⁰

Autonomy has a dual aspect (internal and external) and results from the essential characteristics of the EU and its law. Of course, the Court's case law is replete with such references to the essential characteristics or to similar, synonymous, albeit slightly altered terminology.⁹¹ In fact, it was in the counterpart to Opinion 1/17, Opinion 1/00, that the Court, when asked to opine on a highly integrationist agreement between the European Community and non-Member States on the establishment of the European Common Aviation Area (‘ECAA’), declared the preservation of the essential characteristics as the first requirement for the preservation of the EU (then Community) legal order.⁹² The second requirement that the Court set was the lack of a binding effect of the rules provided therein for the resolution of disputes.⁹³

In Opinion 1/17, the Court expressly mentions three such essential characteristics of EU law: primacy, direct effect, and the fact that it stems from an independent source of law (ie the EU treaties). Furthermore, these characteristics have led to a structured network of principles and, ultimately, the autonomy of the EU legal order is premised on the fact that the EU possesses a unique constitutional framework.⁹⁴ The preservation of all these characteristics and the autonomy of the legal order is the raison d’être of the EU judicial system as premised in Articles 19 TEU and 267 TFEU.⁹⁵

Following this concise recap of the EU dogma of autonomy, the Court framed the test it has to apply. The main question is this: does the envisaged ISDS mechanism (ie the CETA ICS, a mechanism that ‘stands outside’ the EU judicial system and the judicial system of Canada) prevent the Union from operating in accordance with its unique constitutional framework as defined by the CJEU? ⁹⁶ For the question to be answered in the negative (and thus for the CETA Court to be found compatible with EU law), we need to ask two further sub-questions:⁹⁷

1. (1) Does the CETA ICS have the power to apply and interpret EU law or only the provisions of CETA and the ‘rules and principles of international law applicable between the parties’? (first limb of the test);

2. (2) Even if there is no possibility that the CETA ICS applies and interprets EU law, is its jurisdiction determined in such a way that the awards it will issue may have the effect of preventing the EU institutions from operating in accordance with the EU constitutional framework? (second limb of the test).

B. The First Limb: Does the CETA ICS Apply or Interpret EU Law?

This first question,⁹⁸ in short, requires a determination of the applicable law in CETA. Article 8.31 CETA incorporates the applicable law clause and provides for five (a–e) guarantees introduced to keep CETA ICS separate from EU law.

Accordingly, under the provision of Article 8.31, the ICS (a) applies only ‘this Agreement’ (ie CETA), ‘as interpreted’ in accordance with the VCLT⁹⁹ ‘and other rules and principles of international law applicable between the Parties’. Furthermore, (b) the CETA tribunal ‘shall not have jurisdiction to determine the legality of a measure, alleged to constitute a breach of this Agreement, under the domestic law of a Party’. For ‘greater certainty’, (c) the CETA ICS, in determining the consistency of a measure with CETA, ‘may consider, as appropriate, the domestic law of a Party as a matter of fact’ and, in doing so, (d) it ‘shall follow the prevailing interpretation given to the domestic law by the courts or authorities of that Party’, whilst (e) ‘any meaning given to domestic law by the Tribunal shall not be binding upon the courts or the authorities of that Party’.¹⁰⁰

Clearly, the drafters of the provision were very careful and seemingly had in mind this potential moment of scrutiny. The Court examined this applicable law clause under the microscope and concluded that the guarantees offered therein suffice to exclude any possibility that the ICS will apply or interpret EU law. In the Court's ‘applicable law’ analysis, there are four noteworthy features to which we will now turn. More precisely, three points that the Court made (B1–B3) and one point that the Court remained reticent about (B4).

1. Navigating extra-EU waters: Reciprocity in lieu of mutual trust (Opinion 1/17 v Achmea)

First, the Court had to determine how the CETA applicable law clause differs from the applicable law clause in earlier relevant cases. The Court started with the applicable law clause contained in the draft agreement for the Patent Court (ie the envisaged agreement in Opinion 1/09).¹⁰¹ Distinguishing CETA from Opinion 1/09 was an almost frictionless endeavour, as the envisaged Patent Court's jurisdiction was extending to future EU regulation on patents, other related EU law instruments, the general principles of EU law and fundamental rights. Hence, EU law was intentionally part of the applicable law clause.¹⁰²

Inevitably, the Court also had to distinguish the case of the CETA ICS from Achmea ¹⁰³ and the relevant ISDS clause in the Netherlands-Slovakia BIT, namely an intra-EU agreement ‘which was concluded not by the EU but by Member States’.¹⁰⁴ Both Achmea and Opinion 1/17 are cases in which the Court, while ruling on the compatibility of two different but, at the same time, similar mechanisms, continues to paint the picture of the autonomy of the EU legal system, a construction that serves existential needs in that it creates the EU legal order. Although the Court only decides upon the case it has before it on each occasion, it is true that the heavily criticised ruling in Achmea was handed down at a time when Opinion 1/17 was already pending before it. Unsurprisingly, a number of diverging views came to light, ranging from warnings about the potential far-reaching Achmea impact on any ISDS¹⁰⁵ to suggestions that the Court might have offered a boost to the Commission's permanent court solution.¹⁰⁶

However, the only proposition for which Achmea serves as a clear authority is that Articles 267 and 344 TFEU have the combined effect of precluding an intra-EU ISDS clause having the specific characteristics of the one provided in Article 8 of the Netherlands-Slovakia BIT. Beyond this starting point of certainty, the implications of the judgment for other types of ISDS, such as intra-EU cases under the Energy Charter Treaty (ECT), intra-EU BIT cases under ICSID or extra-EU BITs, remain open to debate. Indeed, such debate is prompted and permitted by the characteristic high level of abstraction of the Court's reasoning. That being said, it is true that the Court's construction of autonomy in Achmea was so expansive that no one can put their heads on the block and state what is (or is not) Achmea's reach, and if the Achmea test were to be taken literally, no dispute settlement mechanism could escape from it.

The Court found that the CETA ICS differs from the ISDS mechanism provided in the intra-EU BIT at issue in Achmea on three grounds. First, in Achmea, the tribunal ‘would be called upon to give rulings on disputes that might concern the interpretation or application of EU law’.¹⁰⁷ This was the case because, according to the wording of the applicable BIT, the arbitral tribunal could ‘[take] into account’ among others ‘the law in force of the Contracting Party concerned’.¹⁰⁸ Second, Achmea differs in that it concerned ‘an agreement between Member States’¹⁰⁹ and, third, the principle of mutual trust does not apply in relations between the EU and a third state.¹¹⁰

The Court's reference to the principle of mutual trust is another indication of the role it played in its reasoning in Achmea, where it seemed to have approached the intra-EU ISDS clause at issue as a public international law expression of mistrust between new and old EU Member States originating from (and allegedly belonging to) a different era.

Hence, in Opinion 1/17, the Court drew a line between the sphere of intra-EU relations (where mutual trust applies) and operating in an extra-EU context (where reciprocity comes into play and reigns).¹¹¹ In the specific case of the ISDS, there is a ‘demand for neutrality’¹¹² that the ICS comes to meet. So long as this projected body completely ‘stands outside’ the EU legal system, reciprocity allows EU's participation.

2. EU law as a fact: ‘Examination’ versus interpretation

In the context of the inquiry into the applicable law, the Court also evaluated the exercise that the CETA tribunal will be called to undertake when determining the consistency of a contested measure with CETA. It ‘may consider, as appropriate, the domestic law of a Party as a matter of fact’. Further, in the context of this consideration, it ‘shall follow the prevailing interpretation given to the domestic law by the courts or authorities of that Party and any meaning given to domestic law by the Tribunal shall not be binding upon the courts or the authorities of that Party’.¹¹³

Interestingly, the Court drew a distinction between the examination of EU law as a fact on the one hand and the legal exercise of interpretation of EU law on the other. The challenge against a measure brought by an investor will ‘inevitably’ lead to ‘an examination of the effect of that measure’. Hence, ‘on occasion’, this may ‘require that the domestic law of the respondent Party be taken into account’. However, the Court held that this inevitable exercise cannot ‘be classified as equivalent to an interpretation’ by the CETA tribunal.¹¹⁴ Put differently, the Court seemed to acknowledge the realistic necessity that arises before all courts and tribunals to ‘take into account’ laws over which they have no jurisdiction.

Generally, and particularly in the contemporary landscape, all courts at some point have to deal with issues originating in other (than their own) legal orders as a matter of course. The CJEU is not an exception. Its jurisdiction is confined only to EU law. Yet, in an Article 267 TFEU reference case, the CJEU still takes into account the national law of a Member State without (at least theoretically) interpreting or applying it, in the sense of giving the national court an authoritative interpretation. Certainly, this exercise, labelled by the Court ‘examination’, is not at odds with the approach investment tribunals (as well as other international courts) have expressly taken in the past.¹¹⁵ As a matter of fact, the legal exercise arbitral tribunals conduct is not aimed strictly to interpret and apply EU law in order to produce binding effects beyond resolving the specific dispute.¹¹⁶

Rather quickly, the Court also dealt with the applicable law of the appellate tribunal, the second degree in the CETA ICS. Indeed, the relevant provision in CETA, lists among the grounds on the basis of which the appellate tribunal may ‘uphold, modify or reverse’ the award, ‘manifest errors in the appreciation of the facts, including the appreciation of relevant domestic law’.¹¹⁷ The Court, however, found that this reference to domestic law does not amount to an indirect inclusion of EU law in the applicable law. Instead, the Court noted that it was ‘clear’ in light of the other provisions ‘that it was in no way the intention of the Parties to confer on the Appellate Tribunal jurisdiction to interpret domestic law’.¹¹⁸ This is another instance of the CJEU interpreting the jurisdiction of the CETA tribunal.

3. Determination of the respondent

Another consideration that strengthened the Court's conclusion that there is not even a mere possibility of the CETA ICS interpreting EU law was drawn from the CETA provision on the determination of the respondent (Article 8.21). In more detail, it is the EU that determines, in light of the contested measures and the internal division of powers, who should be the respondent in each case (the Union or the Member State). Such determination is binding upon the CETA tribunal and, therefore, the Court retains exclusive jurisdiction to determine the division of powers.¹¹⁹

4. An ICS standing outside the EU judicial system and the removal of disputes from domestic courts

The Court concluded its analysis of this first limb by making reference to the lack of direct effect of CETA and the finality of the awards rendered by the ICS.¹²⁰ The lack of direct effect has been used both as an argument in support of the inclusion of the ISDS mechanism in CETA and as evidence of complete separation, namely that the ICS will not touch EU law.¹²¹

In addition, in light of its finding that there is not even a mere possibility that the CETA ICS can apply or interpret EU law,¹²² the Court found unproblematic the absence of any provision allowing or requiring the CETA ICS to make requests for preliminary references to the Court. This is a logical and consistent finding. If a tribunal does not apply or interpret EU law, there is no reason for requests for preliminary rulings to the Court.¹²³

Nevertheless, it should be stressed that the Court remained conspicuously reticent about the fact that the domestic courts of the Member States are deprived of some of their jurisdiction. To put it differently, had CETA had direct effect and had it not provided for an ICS, any Canadian investors’ claims should have been brought before domestic courts of the host Member State. Of course, reciprocity plays its part here. Still, the Court has repeatedly emphasised the ‘rhetorical shield’ of the domestic courts in previous cases.¹²⁴ In Achmea, impliedly, mutual trust meant trust in the system of remedies that the EU is postulated on and in their effectiveness, as well as mutual trust in the EU judicial system as a whole. Predominantly, in Opinion 1/09, the Court found troublesome that the domestic courts were ‘divested’ of some of their jurisdiction ‘retain[ing] only those powers which are not subject to the exclusive jurisdiction of the [Patent Court]’.¹²⁵ In Opinion 1/17, EU Member States courts are not the protagonists.

C. The Second Limb: ‘Effects of the Award’ and EU Regulatory Autonomy

The Court then moved on to the second limb of the test.¹²⁶ What is the concern here? The concern is that, despite the separation and lack of direct effect, the CETA tribunal ‘examines’ the contested measures, which might relate to laws and regulations, in light of the facts of the case. EU law (secondary EU law but also primary law that formed the basis for the adoption of a measure) is one aspect of the facts.¹²⁷ Thus, the CETA tribunal will ‘often’ decide on the effects of these EU law measures (facts) and breaches of the substantive standards of CETA, such as the fair and equitable treatment (FET) (Article 8.10), expropriation (Article 8.12), and freedom of transfers (Article 8.13).¹²⁸

In addressing this concern, the Court made four preliminary observations on the features of the jurisdiction of the CETA tribunal.¹²⁹ It observed the broad definition of covered investments¹³⁰ but the limitation of covered investors to those ‘who have a real link with Canada’,¹³¹ as well as the compulsory jurisdiction of the CETA tribunal, in the sense that the respondent's consent is unequivocal.¹³² Thereafter, the Court made the observation that the ICS only grants compensation for the damage that the investor suffered because of the breach, and therefore it does not have the power to request a change to the regulatory framework or impose a penalty on the state.¹³³

It was after these observations that the Court set the negative condition for compatibility: the ICS shall not ‘call into question the level of protection of a public interest that led to the introduction of such restrictions by the Union’.¹³⁴ Otherwise, the Court held, the Union would have to abandon that level of protection ‘in order to avoid being repeatedly compelled’.¹³⁵

So, how is the Court reassured that the CETA ICS, which does not apply or interpret EU law, still will not present a threat to EU regulatory autonomy? Three reassurances have entertained the concerns. First, the expressis verbis recognition of the right to regulate followed by a long indicative list of legitimate policy objectives.¹³⁶ Second, and rather strikingly, the Court relied on the commitment contained in the CETA Joint Interpretative Statement that the standards and regulations of each Party will not be lowered.¹³⁷ Third, the Court considered the formulation of the substantive standards contained in CETA and the restrictive manner in which they are drafted. In particular, it took note of the clarification on the notion of indirect expropriation¹³⁸ and the exhaustive enumeration of situations in which a breach of FET may be found.¹³⁹

The Court's reasoning contains a few novel elements. Leaving aside the reliance on the Joint Statement (which after all is a political statement), what is striking is the Court's attempt to predetermine the jurisdiction of the CETA tribunal (that is an international and completely separate tribunal). Put differently, and borrowing the Court's terminology from another part of the Opinion,¹⁴⁰ a risk to the EU regulatory autonomy would be ‘unimaginable’ if the CETA tribunal interprets its powers correctly, and that is in the way the Court suggests.

However, it seems that the Court takes its chances and departs from the suspicious stance it had adopted in the past. The CETA ICS, as any other tribunal, has kompetenz kompetenz, so it will define its own jurisdiction independently. Moreover, there is nothing in CETA that prevents the ICS from doing so and, where applicable, it might rule that the respondent, say an EU Member State or the EU, breached one of the substantive standards in CETA (eg by falling into the scope of the exceptions provided therein) and this breach was due to the implementation of an EU law.

Is this the beginning of a new era of ‘EU law objections to jurisdiction’ raised before the CETA ICS? In other words, shall we expect the respondent EU (Member State) to argue that the CETA ICS lacks jurisdiction because the contested measure is taken to protect the public interest? Is this reference to the ‘effects’ simply the CJEU's way of emphasising the importance of EU regulatory autonomy, while acknowledging deep inside that there is no immunity for any measures?¹⁴¹ The intra-EU experience has shown that these questions are not of a pure academic interest. Assuming arguendo, that the awards rendered by the ICS are to be enforced under the New York Convention,¹⁴² and there is an award (unimaginably but conceivably) that finds a breach of CETA on the basis of a measure that was taken in view of a significant EU public interest, will Opinion 1/17 serve as the legal basis for the award debtor's non-compliance/denial of enforcement?¹⁴³ Such objection, in this scenario, would be brought in the enforcement proceedings before domestic courts.

Overall, following the Court's analysis, it seems that had CETA not come with these reassurances as understood by the Court (or if, in the future, an agreement comes without similar ones), ‘it would have to be concluded that such an agreement undermines the capacity of the Union to operate autonomously within its unique constitutional framework’.¹⁴⁴

D. An Autonomy Checklist of Full Separation

Regardless of the view one may take on the Court's findings and/or reasoning in Opinion 1/17, it seems that we have a dispute resolution mechanism that pleased the Court, and thus (perhaps for the first time) a test of compatibility with the principle of autonomy for future reference. Is this an easy test? Certainly not. On the contrary, it is a difficult one to pass (especially in light of its second limb). Nevertheless, it is the closest we have to a workable checklist applicable to both first instance and appellate tribunals forming part of an ICS-type dispute resolution mechanism (ICS-DRM). In a nutshell, the components of this checklist can be summarised as follows.

First, the ICS-DRM should not have the power to apply and interpret EU law. The mere possibility that EU law might be brought in the proceedings as applicable law is eliminated by (1) the inclusion of an express and clear applicable law clause;¹⁴⁵ and (2) a provision that only the EU will exclusively determine the respondent in each case.¹⁴⁶ These guarantees shall be, broadly speaking, accompanied by evidence, including (3) the lack of direct effect of the said agreement (ie no direct action by investors in national courts) and, coherently, the absence of a provision for the possibility to send requests for preliminary reference rulings (ie no recourse from the ICS to the EU judicial system).¹⁴⁷

Even if there is no possibility that the CETA ICS applies and interprets EU law, the inquiry does not stop there. Its jurisdiction should be determined in such a way that its awards do not have ‘the effect of preventing the EU institutions from operating in accordance with the EU constitutional framework’.¹⁴⁸ Under which circumstances could the ICS awards have this unconstitutional effect? When they could have a regulatory chill, is the answer.¹⁴⁹ And how can the drafters prevent this risk? At least in the following three ways, the Court held: (1) by including an express, broad and declaratory provision on the parties’ right to regulate;¹⁵⁰ (2) by granting to the ICS the power to order compensation instead of any other remedy;¹⁵¹ and (3) by drafting narrowly the substantive standard clauses (such as the FET clause in CETA).¹⁵²

With the above-sketched checklist in mind, we will now turn to reflect on the implications it signals for other extra-EU scenarios in the future, namely ISDS provisions contained in both extra-EU BITs and other EU agreements, including notably a future Convention establishing an MIC.¹⁵³

IV. FUTURE IMPLICATIONS FOR OTHER EXTRA-EU SCENARIOS

A. ISDS in Other EU Agreements Including the Future Convention Establishing an MIC

The envisaged opt-in Convention establishing the MIC (MIC Convention) is for now an exercise on paper. In the absence of a final sketch of the MIC Convention, it is at least somewhat premature to fully discuss its compatibility with EU law.¹⁵⁴ Atop the complexities arising out of the compatibility of the MIC (and ICS) with existing international law instruments,¹⁵⁵ and other practical considerations,¹⁵⁶ any MIC Convention (having the EU on board) will have to take into account the CJEU's construction of autonomy and compatibility requirements (as set forth in Opinion 1/17).

It should be stressed from the outset that compatibility of the CETA ICS does not automatically guarantee the compatibility of the MIC with EU law. It will largely depend on the content of this future agreement. What can be noted already at this stage is that a future MIC Convention will have to take into account the EU's sui generis character¹⁵⁷ and include a minimum set of guarantees. One such guarantee should incorporate a very carefully drafted applicable law clause that will exclude the interpretation/application of EU law (as law), provide for its compulsory jurisdiction and grant to the EU the exclusive power to determine the respondent when it comes to a case brought against an EU Member State. Importantly, it will have to include a disconnection clause to exclude an inter se application if it is to be concluded as a mixed agreement.¹⁵⁸

This will be the easy part. However, a future MIC Convention will also have to introduce sufficient safeguards for the states’ regulatory powers. This last point (that is meeting the second limb of the test) might prove particularly challenging. For instance, in CETA, the parties drafted a very narrow FET clause. However, a FET clause is a substantive standard. Intentionally, substantive provisions are not part of the MIC negotiations. The MIC Convention is envisaged to determine procedure only,¹⁵⁹ premised on the idea of the ICSID Convention, and aimed to be used by reference in a large number of different IIAs. Such IIAs will have their own substantive provisions and references (or, often, not) to the states’ regulatory powers. It is these IIAs that will have to comply with the above-sketched Opinion 1/17 ‘effects’ test. Will the CJEU be at ease declaring compatible with the EU legal order an agreement comprising of procedural only provisions and, thus, take the risk of it applying potentially non-compliant substantive provisions contained in different texts?

B. Member States’ Extra-EU BITs

Extra-EU BITs touch upon the role of Member States as international actors. Member States should not conclude an extra-EU BIT that might require it to breach its EU law obligations. In the aftermath of Opinion 1/17, it can be said with certainty that extra-EU BITs are not discriminatory. However, conclusions in relation to their compatibility with the autonomy of the EU legal order cannot be drawn with the same degree of certainty.

Different extra-EU BITs include different applicable law clauses. In some (especially if concluded before the Grandfathering Regulation),¹⁶⁰ the applicable law clause might be similar to the applicable law clause in Article 8 of the Netherlands-Slovakia BIT (the applicable ISDS clause in Achmea) and, thus, bring in the parties’ domestic law and EU law. Certainly, an extra-EU BIT (that pre-dates CETA) containing a clear-cut exclusion of EU law (in the way CETA does) is not common. On the other hand, extra-EU BITs raise different concerns than intra-EU BITs. For example, in an extra-EU BIT scenario, the EU is not bound by any interpretation rendered by an arbitral tribunal,¹⁶¹ whilst, certainly, they do not raise the mutual trust concerns that the intra-EU BITs arguably do.

Moving to the link between autonomy and the effects of the awards that the Court drew in Opinion 1/17,¹⁶² can we argue that an extra-EU BIT (which does not contain the elaborated FET/indirect discrimination clauses that the new-era treaties tend to include) cannot live up to the Opinion 1/17 test? Even under an extra-EU BIT (to which the EU is not a party), an EU measure might relate to the breach claimed. Does this call into question the ‘level of protection of a public interest’? And if yes, is this problematic in the context of an extra-EU BIT (as the Court held it is in the context of an EU IA)?

The question will remain open until (and if ever) the Court is called to rule upon another extra-EU scenario, this time not on the basis of an extra-EU agreement but on the basis of an extra-EU BIT. For instance, this could arise in the context of an enforcement of a non-ICSID award before an EU domestic court. The party objecting to the enforcement could argue that the award has the effect of breaching EU law (eg a breach of state aid laws, for instance in an extra-EU solar case) or is a breach of the EU constitutional law/principle of autonomy within the meaning the CJEU attributed to it. As the effects-limb of the autonomy test in Opinion 1/17 is so flux, the concern is that it could convert the intra-EU objection to jurisdiction into a general EU-autonomy-objection invocable in extra-EU cases.¹⁶³

Overall, the extent of the implications of Opinion 1/17 (read together with Achmea) for the extra-EU BITs remains to be seen.

V. CONCLUSION

In Opinion 1/17, the Court safeguarded the ability of the Union to enter into reciprocal commitments with partners around the globe and saved, for now, its credibility as an international actor. The complex (from a legal, political, and policy perspective) pathway towards the conclusion of CETA and the EU's leading role in shaping the debate on ISDS reforms globally played a pivotal role. A ‘no’ to the ICS from the Court would have amounted to a global embarrassment for the Commission and the EU. The Court's explicit acknowledgment of the ‘need to maintain the powers of the Union in international relations’¹⁶⁴ (and hence the need to allow the Union to be a subject of international law), leaves no doubt that the Court was mindful of the context and stakes.

Mindful as it may have been, the Court still took the opportunity to build upon its constitutional framework.¹⁶⁵ The autonomy test it introduced (premised on the principle of reciprocity that underpins external relations) stresses the importance of keeping this body ‘that stands outside’ the parties’ (and thus the EU's) legal system as separate as conceivably possible. Importantly, the EU regulatory powers form part of the autonomy test. In finding that the CETA ICS presents no threat to the EU regulatory autonomy, the Court relied primarily upon the formulation of the substantive standards included in CETA and, strikingly, on its own interpretation of the projected CETA tribunal's jurisdiction.

Overall, the Court, unable in the past to tolerate the mere possibility that a challenge, even in its remotest form, to the autonomy of EU law could occur, explained in Opinion 1/17 how autonomy cannot be a problem in reciprocal external relations. Still, in light of the conditions set, this openness should not be taken for granted, as no guarantee of automatic compatibility of a future body (such as the projected MIC Convention, most likely to be a procedure-only text) is offered.