On Paper Crowns: The Exclusive Jurisdiction of the Court of Justice of the EU and International Law

Table of Contents: 1. Introduction. – 2. The CJEU and the Autonomy of EU Law. – 2.1. Exclusive jurisdiction under Article 344 TFEU. – 2.2. Exclusive jurisdiction over broader interpretation of EU law. – 3. The View from Public International Law. – 3.1. Key Concepts and Structures of Public International Law. – 3.2. The Limits of CJEU Exclusivity. – 4. Conclusion.

Abstract: In recent years, several high-profile cases have seen the development and consolidation by the Court of Justice of the European Union (CJEU) of the concept of autonomy of European Union (EU) law in relation to international law. Although the wealth of CJEU jurisprudence has garnered much academic debate, its contours are still hazy and much of the discussion is EU-centric. In this context, this article focuses on one of the most developed features of the external dimension of autonomy, namely the exclusive jurisdiction of the CJEU vis-à-vis other international courts and tribunals. It seeks to demonstrate that in contrast to the jurisprudence and reasoning of the CJEU, the international law foundations of this form of autonomy are shaky at best. In juxtaposing the CJEU and international law perspectives, it questions whether the CJEU’s position on its exclusive jurisdiction has a true legal basis in international law. Instead, this article asks whether the autonomy of EU law, specifically the exclusive jurisdiction of the CJEU, is merely a paper crown, dependent on recognition by others for its authority.

Keywords: autonomy – international law – self-contained regimes – norm conflict – fragmentation – CJEU.

1.   Introduction

In recent years, several high-profile cases have seen the development and consolidation by the Court of Justice of the European Union (CJEU) of the concept of autonomy of European Union (EU) law in relation to international law. While the concept has generated a wealth of academic debate and scholarship,[1] its contours are still hazy, not least due to the evolving jurisprudence of the CJEU.[2]

Autonomy of EU law ‘is part of the very DNA’ of EU legal order,[3] having played a crucial role in defining the principles of supremacy and direct effect in foundational cases of Van Gend en Loos and Costa v ENEL.[4] While the concept featured prominently in subsequent case law concerning the internal legal order of the EU (internal dimension of autonomy),[5] It took on a life of its own in the CJEU case law concerning the relationship between the EU legal order and international law (external dimension of autonomy).[6] Autonomy has internally distinguished the EU legal order from national legal systems, while externally insulating its core features from outside influence. Fundamentally self-referential in nature,[7] autonomy has thus served to render the EU legal order apart from other legal orders receiving much criticism and occasional defense in the process.[8]

While the CJEU presents autonomy as a constitutional principle intrinsic to the EU legal order, the same principle appears differently when viewed from outside the system it seeks to constitute. In international law, the EU Treaties remain binding treaties governed by the Vienna Convention framework, and the autonomy of the legal order they establish therefore operates as a claim within, not above, the international legal order. Understanding this dual character of autonomy, internal constitutional principle yet externally an assertion grounded only in treaty law, is essential for assessing whether the CJEU’s position can be sustained on the international plane.

Within this context, this contribution critically examines the concept of autonomy as developed by the CJEU: what exactly does ‘autonomy’ mean, and what is its legal basis? The concept of autonomy, closely associated with ‘the very foundations’ or ‘essential characteristics’ of EU,[9] was most recently recited in Komstroy: autonomy of EU law pertains to ‘essential characteristics of the [EU] and its law, relating in particular to the constitutional structure of the [EU] and the very nature of that law.’[10] Specifically, this article focuses on ‘the most strongly developed’ institutional feature of the external dimension of autonomy, namely the exclusive jurisdiction of the CJEU vis-à-vis other international courts and tribunals (ICT).[11] It analyses the scope and legal foundation of the CJEU’s exclusive jurisdiction (based inter alia on Article 344 of the Treaty on the Functioning of the EU (TFEU)[12] (ex Article 292 EC Treaty)[13], under which EU Member States ‘undertake not to submit a dispute concerning the interpretation or application of the Treaties to any method of settlement other than those provided for therein’. It does so first from the perspective of the CJEU itself (Section 2), before testing the Court’s conception of its exclusive jurisdiction against public international law (Section 3). Here, we argue that the legal foundation for the CJEU’s claim to exclusive jurisdiction in international law is highly debatable, as it rests on a distinctly EU-centric conception of law that does not always correspond with the broader structures, concepts, and principles of general public international law.

Ultimately, this article seeks to demonstrate that, in contrast to the jurisprudence and reasoning of the CJEU, the international law foundations of this form of autonomy are shaky at best. In juxtaposing the CJEU and international law perspectives,[14] it questions whether the CJEU’s position on its exclusive jurisdiction has a true legal basis in international law. Instead, this article asks whether the autonomy of EU law, specifically the exclusive jurisdiction of the CJEU, is merely a paper crown, dependent on recognition by others for its authority.

2.   The CJEU and the autonomy of EU law

Although not explicitly prescribed in EU primary law, the concept of autonomy – particularly its external dimension – derives from CJEU case law, much of it developed through the Court’s jurisdiction under Article 218(11) TFEU to assess the compatibility of envisaged international agreements to be concluded by the Union with the EU Treaties. Over decades, this case law has evolved, adapting to different international contexts and presenting a concept ‘in flux’ that manifests varying features depending on the circumstances.[15] The successive case law shows not only refinement of the concept of autonomy, but also its expansion. While the jurisprudence on external autonomy reflects the broader aim of preserving the Court’s authority over the relationship between EU and international law[16], the subsequent sections analyse closely only the case law where the CJEU’s exclusive jurisdiction under Article 344 TFEU and beyond played a central role. This is a necessary starting point for analysing the autonomy of EU law vis-à-vis other ICTs. Only once this internal perspective is established can one meaningfully assess how international law conceptualises the autonomy of ICTs.

2.1.  Exclusive jurisdiction under Article 344 TFEU

Article 344 TFEU has been described as ‘an obvious and necessary provision for ensuring the supremacy, unity, and consistency of [EU] law’.[17] This provision featured first in Opinion 1/91 on the European Economic Area (EEA) Agreement where the Court was concerned with its own functioning more broadly.[18] Specifically, the possibility of the EEA Court ruling on what constitutes a ‘contracting party’ was seen by the CJEU as a threat to its exclusive jurisdiction prescribed in the then Article 219 EEC Treaty.[19] Thus, rather than considering a possibility of an inter-Member State dispute arising in the EEA Court, the CJEU instead focused on the possibility of the EEA Court ruling on division of competences between the EC and its Member States.[20]

The only case involving a real dispute between two Member States is Mox Plant concerning an infringement action against Ireland.[21] The Commission claimed that Ireland violated EU law by bypassing the Court and submitting its dispute with the UK concerning potential environmental harm from the plant to two external tribunals: one under UNCLOS and another under the OSPAR Convention. Having found that the UNCLOS provisions relied upon by Ireland ‘are rules which form part of the EC legal order’ (namely environmental legislation and Euratom rules), the CJEU affirmed its jurisdiction and went on to clarify that its jurisdiction is exclusive under then Article 292 EC Treaty.[22] The EC judicial system has precedence therefore, and internal legal procedures (infringement proceedings) under Article 227 EC Treaty (current Article 259 TFEU) should have been relied upon to resolve the dispute and to seek interim measures.[23]

The second head of the Commission’s complaint concerned particularly Ireland’s submission of EU instruments to UNCLOS arbitral tribunal interpretation and application. Here, the proceedings before UNCLOS involved ‘a manifest risk that the jurisdictional order laid down in the Treaties and, consequently, the autonomy of the Community legal system may be adversely affected.’[24] Underscoring the importance of the autonomy of the EU legal order, the judgment highlighted that international treaties such as UNCLOS must be interpreted and applied in a manner consistent with EU law when Member States are parties to them. The CJEU thus significantly limited the Member States’ freedom to opt for alternative dispute settlement mechanisms, at the same time interfering with the functions of other ICTs requiring certain deferral of their jurisdiction.[25] In practice, however, according to de Witte, an interpretation by an external court would not have been binding on the CJEU which could have developed its own interpretation of relevant EU rules.[26] A further weakness in the CJEU’s position was its level of analysis about EU law itself, namely the absence of an inquiry into the level of harmonisation,[27] and the ultimate outcome of ‘communitarising’ of significant chunks of international law of the seas.[28] Entrenching its exclusive jurisdiction thus threatened that EU courts would be in ‘clinical isolation’ from other ICTs.[29]

The next occasion for relying on Article 344 TFEU came in Opinion 2/13,[30] where the Court was branded as a ‘selfish’[31] or ‘fearful’[32] court, undermining engagement with international law, but also its commitment to human rights.[33] Although the Court considers Article 344 TFEU separately from another ground on ‘specific characteristics and autonomy’ of EU law, it is clear that the CJEU analysis of Article 344 TFEU alongside the other grounds of incompatibility discussed in the next section equally concern autonomy. In the Court’s view, the draft accession agreement did not prevent the possibility of the EU or its Member States submitting an application to the ECtHR.[34] Not directly excluding such a possibility concerning ECHR violations in conjunction with EU law was seen as undermining Article 344 TFEU.[35] The latter in itself precludes any prior or subsequent external control, and only an express exclusion of ECtHR jurisdiction on the application of the ECHR within the material scope of EU law would suffice.[36] Member States are thus prevented from entering into such agreements altogether, as requiring express exclusion effectively expands the scope of Article 344 TFEU.[37] In this way, a perceived internal problem was projected into the draft accession agreement.[38] Furthermore, although Article 344 TFEU refers only to Member States, its scope has been interpreted to also cover potential disputes between a Member State and the EU.[39] Together with Mox Plant, Opinion 2/13 thus resulted in an expansive interpretation of Article 344 TFEU.[40]

The provision reappeared in Achmea, dealing with the compatibility of the arbitration clauses in bilateral investment treaties (BIT) concluded between EU Member States.[41] Article 344 TFEU featured again despite the absence of an intra-EU dispute and the referring court questioning its relevance.[42] Although the Court’s finding of incompatibility in intra-EU BITs focused mainly on their impact on the preliminary reference procedure, the CJEU ultimately held the agreement violated both Articles 267 and 344 TFEU.[43] Article 344 TFEU was relied upon here most likely as an expression of the duty of loyalty in Article 4(3) Treaty on European Union (TEU),[44] as established in Mox Plant earlier.[45] Indeed, in para 58, the CJEU noted the risks that such external dispute settlement mechanism would have on mutual trust and the principle of sincere cooperation in the EU. According to Hindelang, in this sense Article 344 TFEU is a broader provision which encompasses all judicial proceedings and it intends to prevent their impairment from the Member States.[46] At the same time, , from the standpoint of the interpretation of EU law, Article 344 TFEU should be seen as an instance of broader CJEU jurisdiction to interpret EU law even in cases not concerning disputes between EU Member States or institutions.[47] Indeed, in those cases the Court refers to Article 344 to confirm the general rule under Article 19(1) TEU according to which the CJEU ‘shall ensure that in the interpretation and application of the Treaties the law is observed.’[48] This will be examined in the following section.

2.2.  Exclusive jurisdiction over broader interpretation of EU law

Already in Opinion 1/76, the first case on the European Community’s (EC) external autonomy, the early warnings shots appeared in terms of the Court’s determination to guard the EU legal order, including its role and legal effectiveness.[49] Although the CJEU did not make a final pronouncement on the ability of national courts to request a preliminary ruling from the Fund Tribunal, it did express its concern over potentially conflicting interpretations hoping that this would be avoided in practice as it might have an impact on legal certainty.[50] The effectiveness of legal protection was another concern here, linked also to the possible undermining of the impartiality of its own judges if the same question appears in front of both courts given that the CJEU judges were to be also part of the Fund.[51]

The concerns about the ability of the CJEU judges to perform their duties resurfaced subsequently in Opinion 1/91 on the EEA Agreement.[52] Most importantly, however, Opinion 1/91 clearly demonstrated the CJEU’s unease regarding the potential of other international courts to interpret EU law. Given that the EEA legal order would replicate EC law, the possibility of interpreting identical legal provisions divergently would threaten the uniformity of application of law: while the EEA courts had a duty to interpret the EEA agreement in light of the CJEU rulings adopted prior to the signature of the agreement, there was no such obligation concerning future judgements.[53] This threat to the ‘very foundations’ of EC law would not be alleviated by the fact that CJEU judges will sit on the EEA Court as the judges will have to be guided by different approaches, methods and concepts.[54] Legal certainty was also at risk, as CJEU judgments on EFTA court questions would be advisory, rather than binding. This will create uncertainty for courts of the EC Member States and might spread over to the judgments issued as part of EC preliminary reference procedure undermining its ‘proper operation’,[55] hence effectiveness. Ultimately, in light of the Court’s claim to exclusive interpretative jurisdiction in Opinion 1/91,[56] the EEA Court was replaced by a ‘politically neutered’ EFTA Court.[57]

Unlike the EEA judicial system, the dispute settlement mechanism in the agreement on the Establishment of European Common Aviation Area (ECAA) was deemed to be compatible with the autonomy of EU legal order in Opinion 1/00.[58] Repeating Opinion 1/91 in that mechanisms under such agreements must not bind EC and its institutions ‘to a particular interpretation of the rules of Community law referred to in that agreement’ in their internal dealings,[59] here the CJEU decisions were to be binding unlike under the EEA Agreement. This was in relation both to preliminary rulings from national courts, as well as in resolving differences that the parties brought to the CJEU in case the procedure before the Joint Committee established under the draft agreement has failed.[60] Besides, the uniformity of EU law remained intact: CJEU preliminary rulings were binding, and the ‘single pillar’ structure, along with the Commission’s veto power in the Joint Committee, ensured consistent interpretation.[61]

The preliminary reference procedure took a central stage in Opinion 1/09 concerning the draft agreement on European and Community Patents Court (ECPC),[62] prompting reflection on the role of the national courts and intra-EU judicial dialogue.[63] The ECPC was deemed to be incompatible with the Treaties as it would replace the national courts’ exclusive jurisdiction, depriving them of the power to request preliminary rulings from the CJEU in the relevant field (national courts being the sole courts capable of requesting such rulings).[64] The intactness of the preliminary reference procedure was essential for ‘the correct application and uniform interpretation’, and thus ‘indispensable to the preservation of the very nature of [EU] law.’[65] The Court’s focus on the national courts here was seen as ‘disingenuous’ in that it ignored the provisions of the draft agreement which would have secured its role – the ECPC would have been bound by CJEU case law.[66]

Others linked the case and the need for uniform interpretation of EU law to the protection of rights in the EU.[67] In this respect, one should also note the CJEU’s additional dissatisfaction with the draft agreement in that it did not provide for a mechanism to hold the ECPC accountable if it breached EU law, nor did it allow for effective remedies under the principles established by EU law.[68] Ultimately, the judgement was one of a significant stretch not only into the Member States’ domain over their national judicial systems,[69] but also the operation of other ICTs which do apply EU law in practice.[70]

This reach of the Court into the domain of Member States transpires also in Achmea and Komstroy. The latter case, concerning the dispute resolution system under the Energy Charter Treaty, is the latest pronouncement on autonomy vis-à-vis ICTs.[71] In both cases, the carving out of certain jurisdictional space from the preliminary reference procedure is of central concern. Achmea was a result of a preliminary request by the German Federal Court of Justice in relation to a dispute between the Slovakia and Achmea BV, a Dutch insurance group. Given that the adjudication of the tribunal under an intra-EU BIT might involve issues pertaining to EU law and the fact that the arbitral tribunal is not part of the EU judicial system, the CJEU came to a conclusion that this would undermine the effectiveness of EU law: arbitral awards cannot always be appealed in national courts which could in theory request a preliminary ruling from the CJEU.[72] The matter here is not about protecting the CJEU exclusive jurisdiction vis-à-vis an external court as the case concerns an intra-EU BIT dispute settlement mechanism.[73] It is rather the role that national courts play in ensuring the effectiveness of EU law: the tribunal will not only be interpreting EU law, but also applying it,[74] sidelining national courts without the possibility of triggering the preliminary rulings procedure. In this respect, Achmea offered a ‘broader concept of autonomy’ which ‘antagonizes’ EU law with international investment law.[75]

In the same vein, the dispute settlement system under the Energy Charter Treaty in Komstroy was also found to be incompatible with the autonomy of EU law if it involves a dispute between an investor of one Member State and another Member State concerning EU law. As in Achmea, the CJEU distinguishes the tribunal from ordinary commercial arbitration: the Energy Charter Treaty removed certain disputes from the jurisdiction of national courts, therefore from the preliminary reference procedure.[76] For the CJEU, this means that the uniform interpretation, hence the consistency and the full effect of EU law, will be affected.[77]

One must, however, return in time to Opinion 2/13 to form a full picture of the CJEU exclusive jurisdiction to interpret EU law. Besides Article 344 TFEU discussed above, the CJEU’s claimed absolute jurisdiction over EU law through the additional grounds concerning the preservation of the EU judicial order.

The first ground entitled ‘specific characteristics and autonomy of EU law’ can be broadly summarised as related to the ‘exceptionalism’ in the field of Area of Freedom, Security and Justice.[78] Among other features troubling the Court here, one aspect concerned the preliminary reference procedure: the possibility that requests for advisory opinions of the ECtHR might adversely affect the EU’s own preliminary ruling procedure.[79] Autonomy therefore became part of the ‘specific characteristics’ of EU law distilled here to ‘being about protecting the ‘essential character’ of the Court’s own powers’.[80]

The remaining grounds were similarly linked to the CJEU’s preoccupation with its own prerogatives. Thus, the co-respondent mechanism fell foul as it would affect the division of powers if the ECtHR was to decide whether the EU or its Member States should get involved given that the matter of allocation of competences should be left to CJEU interpretation,[81] even though this for instance happens in the context of WTO law.[82] As for the procedure for the CJEU’s prior involvement, the agreement did not exclude the possibility of ECtHR giving a ruling on a matter of EU law before the CJEU had a chance to do so.[83] In particular while the procedure allowed for the CJEU to rule on the validity of EU law, it failed to secure the Court’s exclusive jurisdiction over ‘the definitive interpretation’ of EU law.[84]

The last ground, probably most astonishing of them all, related to the limited CJEU jurisdiction over Common Foreign and Security Policy (CFSP) matters: since the CJEU itself lacked jurisdiction over CFSP matters except for certain derogations, it was incompatible with the EU legal order that the ECtHR would have such jurisdiction.[85] Halberstam links this to the CJEU’s ‘consolidating function of EU law’: ECtHR’s interpretation of CFSP will trickle down to national courts.[86] More generally, Moreno-Lax and Ziegler have noted the expansion of autonomy in Opinion 2/13 where the CJEU no longer insisted that it is only when the ICT interpretation is binding on EU internally that it will contradict the autonomy of EU law: any interpretation of EU law seems to be incompatible now.[87] In sum, the CJEU’s fervent defense of its exclusive jurisdiction over EU law – all of it – was a signal to the ECtHR to engage in judicial dialogue, particularly before assessing potential ECHR breaches or subsequent remedies involving EU action.[88]

Against the background of the CJEU absolutist approach to its jurisprudence, there was an apprehension over its position on trade agreements with third countries as not only legal, but also policy reasons would play a role.[89] In this respect, Opinion 1/17, concerning the compatibility of investor-state dispute settlement (ISDS) mechanism in the Canada-EU Free Trade Agreement (CETA), stands out due to the Court’s reconciliatory tone, content with the ‘procedural safeguards’ in the dispute settlement.[90] In particular, it distinguished the CETA tribunal both from the ECPC in Opinion 1/09 and the tribunal under the intra-EU BIT in Achmea as the latter would inevitably engage in the interpretation of EU law.[91] On the contrary, a tribunal under CETA may consider the domestic law of the party only ‘as a matter of fact’.[92] It is however not clear where the line lies between this and an interpretation of the law.[93] Further, in its second limb, CETA was let off the hook since neither the tribunal, nor the appellate tribunal could call into question the allocation of powers within the EU as they could not decide whether to bring the case against the EU or a particular Member State.[94] Such leniency of approach and the trust the CJEU placed with the CETA tribunal stand in a stark contrast with its position in Opinion 2/13,[95] appearing to adapt autonomy for the sake of efficient external action,[96] and undermining legal certainty.[97]

To sum up, autonomy has become the channel allowing the CJEU to claim exclusive absolute jurisdiction over EU law risking ‘isolation of the EU at the global level,’[98] or even ‘undermining international law’.[99] This concludes the analysis in Section 2, which traced how the Court constructs autonomy internally and the doctrinal techniques it employs to protect that construction. The contrasting perspective under public international law will be discussed next. Section 3 therefore shifts the analytical lens outward: it examines how the EU Treaties are situated within the general international legal order, how rules such as those in the Vienna Conventions apply to them, and how other international courts and tribunals may approach questions involving EU law. By moving from the Court’s internal reasoning to the external legal framework that governs the Treaties internationally, the next section highlights the tension between these two perspectives and sets the stage for assessing the limits of the exclusivity claim.

3.   The view from public international law

The second part of this article situates the concept of the autonomy of EU law as developed by the CJEU within the international legal landscape. The aim is to assess the CJEU’s claim to exclusive jurisdiction over the interpretation and application of EU law from an international law perspective. In short, we argue that the legal basis of the exclusive jurisdiction of the CJEU in international law is questionable at best, relying as it does on a quintessentially EU-focused vision of law that does not necessarily align with the overarching structures, concepts, and principles of general public international law. It begins by outlining key relevant foundational concepts and structures within international law, before examining the potential legal bases and arguments upon which the exclusivity of CJEU jurisdiction might be founded, and highlights the limitations of such arguments in the international law context.

It is worth noting as a starting point that the EU Treaties themselves are, from the perspective of international law, ‘international agreement[s] concluded between States in written form and governed by international law’.[100] They are therefore ‘treaties’ within the meaning of the Vienna Convention on the Law of Treaties 1969 (VCLT), no different in principle or legal form from other treaties concluded between states. Despite their uniquely complex and detailed character, therefore, for an international lawyer at least, the EU Treaties are merely part and parcel of the broader international legal landscape.[101]

3.1.  Key concepts and structures of public international law

3.1.1.   Norm conflict in international law

With very limited exceptions,[102] there is no system of normative hierarchy in international law. Conflicts between norms, namely situations where two applicable and binding legal rules require distinct and incompatible outcomes,[103] are, for the most part, resolved by means of interpretation.

Where at least one of the conflicting norms is found in treaty law, the relevant framework is found in the general rules for treaty interpretation codified in the VCLT.[104] Article 31(3)(c) of the VCLT is designed to provide a tool to avoid norm conflict as far as possible by requiring the interpreter to ‘take into account’ any relevant external rules of international law.[105] Where this is not possible, however, it may be necessary to determine which of the conflicting norms is to take precedence in a particular case or situation. Interpretive principles like lex specialis and lex posterior can be applied in such cases, but their utility and outcome are far from determinative and their relevance for international legal conflicts is often debateable, and always highly context-dependent.[106]

Certain further considerations arise when determining whether a particular treaty may be said to exclude the application of other, conflicting rules of international law. The first is the identity of the parties to the treaty and to any related dispute. In accordance with Article 34 VCLT, ‘[a] treaty does not create either obligations or rights for a third State without its consent.’[107] As such, a treaty can only alter the rights and obligations of parties to that treaty, and does not impact the rights and obligations of other states in their relations with treaty parties.

The second – and perhaps most fundamental – consideration is whether this aligns with the intentions of the parties which underlie the treaty itself. This is in line with the overarching interpretive perspective of the VCLT and its focus on the will of the parties.[108] It is acknowledged in international law that states may agree to generate ‘self-contained’ legal regimes which exclude the operation of particular rules of general international law.[109] Whether or not a particular treaty has this effect will come down to whether, on interpretation of that treaty, it can be said that parties intended to exclude the operation of those rules of general international law.[110] The notion of ‘self-contained regimes’ is dealt with in more detail below.

In sum, from an international law perspective, the EU Treaties hold no superior hierarchical position in relation to other rules of international law. The extent to which EU law takes precedence over or excludes the application of other rules of international law will depend on a specific interpretation of the norm in question, in accordance with the interpretive rules in the VCLT. Furthermore, in accordance with Article 34 VCLT, the EU Treaties in and of themselves only create rights and obligations for Member States, and not for third parties.

3.1.2. Absence of Precedent in International Law

In principle, there is no system of precedent in international law. The decisions of ICTs are generally binding upon the parties to a particular dispute,[111] but there is no obligation for judges or arbitrators to follow the same interpretation of the law in a future case.[112] This arises in part out of a reluctance by states and other international actors to grant the power to create law to such bodies, as a formal rule of precedent essentially amounts to the recognition of judicially-generated law.[113] Despite the absence of a formal rule of precedent, there is a strong tendency to follow prior jurisprudence in the interests of consistency and coherence of international law.[114] Those interests do not, however, completely prevent situations arising where an ICT comes to a decision in a particular case that amounts to a conflicting interpretation of the law in relation to a previous decision made by the same ICT as discussed next.[115]

3.1.3.    Absence of institutional hierarchy between international courts and tribunals

Just as there is no hierarchy of norms in international law, there is also no formal institutional hierarchy between ICTs.[116] There is no rule in international law that places the interpretation of one adjudicative body above that of another. Thus, there can arise conflicting interpretations of the same rule by different courts, with no clear means by which to determine the single ‘correct’ interpretation.

This is so even where one adjudicative body is supposedly the authoritative interpreter of a particular treaty or area of law. Take, for example, the UN human rights bodies, which are understood to be authoritative interpreters of their respective treaties.[117] Nevertheless, this power of authoritative interpretation has not stopped other ICTs from coming to divergent interpretations of the rules in those treaties.[118] In Qatar v United Arab Emirates, the ICJ took a very different view to the International Committee for the Elimination of Racial Discrimination (ICERD) on the interpretation of the Convention for the Elimination of Racial Discrimination (CERD).[119] Contrary to an earlier finding by the ICERD, the ICJ held that it was not bound by the ICERD’s assessment, and that ‘national origin’ did not include current nationality for the purposes of the Convention.[120] The ICJ has consistently held that it will take the interpretation of bodies like the ICERD ‘into account’ and that it will ‘ascribe great weight’ to such findings, but that this is not determinative.[121]

Despite the central role of the ICJ in the international legal system and its position as the ‘principal judicial organ’ of the UN,[122] as a matter of law the ICJ’s interpretation cannot overrule or otherwise invalidate the conflicting interpretation of the CERD. The other side of this coin is that the ICERD, despite being recognised as the authoritative interpreter of the CERD, cannot claim to have legal superiority over the ICJ in such a case of interpretation. Both interpretations exist; both are in certain terms valid, despite being in conflict with one another.

The ICJ also famously disagreed with the International Criminal Tribunal for the former Yugoslavia (ICTY) with regard to the test for the attribution of conduct of individuals to a state, amounting to a disagreement in interpretation of the rule codified in Article 8 of the Articles on the Responsibility of States for Internationally Wrongful Acts (ARSIWA).[123] The ICJ continues to adopt a stricter ‘effective control’ test in place of the ICTY’s ‘overall control’ standard.[124] It is clear that the ICTY and ICJ’s interpretations of the rule of attribution are different tests; different interpretations of the same rule.[125] Today, there therefore exist two possible and distinct interpretations of Article 8 ARSIWA and the notion of ‘conduct directed or controlled by a state’: both arguably valid, neither overruled by the other. This stands as yet another example of conflicting interpretations by different ICTs.

It must be acknowledged that such conflicting interpretations are relatively unusual. Generally, ICTs often do defer to the interpretation of another body. It is nonetheless clear that such conflicting interpretations do indeed occur, and that the avoidance of such conflict is a question of institutional approach rather than being based on a rule of international law.

In sum, the absence of institutional hierarchy between ICTs, even where one body has particular substantive expertise in the relevant area (like the ICERD), means that international law does not recognise a single authoritative interpretative power by a particular court; nor does it have a centralised or single means for resolving competing claims to adjudicative or interpretive authority. International law is thus familiar with such conflicting interpretations. Whether a particular ICT respects the power of another ICT will largely come down to its own assessment of the respective jurisdictional limits of each ICT, and the degree of deference it is willing to give to the judgment of the other ICT.

3.1.4.   Fragmentation and self-contained regimes

The preceding sections illustrate an international legal landscape that is familiar (if not quite comfortable) with normative and interpretive conflict and inconsistency. The existence of conflicting norms, conflicting interpretations of the same norms (by the same ICT, and by different ICTs) is part and parcel of the international legal landscape.

This prospect of normative incoherence and competition in international law lies at the heart of fears related to the so-called ‘fragmentation’ of international law.[126] Fragmentation, or ‘the ‘f’ word’, was a popular discussion topic in the 1990s and early 2000s following a speech by the then-president of the ICJ, President Schwebel, in 1999, warning in particular that the proliferation of adjudicative bodies raised ‘a serious risk of conflicting jurisprudence as the same rule of law might be given different interpretations in different cases.’[127] As Koskenniemi has argued, the proliferation of specialised regimes inevitably produces centrifugal tendencies in international law, each asserting autonomy while contributing to systemic fragmentation.[128]

It may be said that modern international legal commentary and jurisprudence has ceased to be quite so afraid of fragmentation and its impact on international law.[129] The unifying elements of secondary, general international law rules, including the law of treaty interpretation and of state responsibility, as well as interpretive techniques such as systemic integration, are generally viewed as sufficient to prevent unacceptable incoherence and inconsistency within the legal order.[130]

A daughter-concept of fragmentation, so-called ‘special regimes’ or ‘self-contained regimes’, go further than the normative uncertainties discussed above in that it serves as to exclude elements of this general international law from particular legal regimes.[131] There are terms used to describe legal regimes, usually established by treaty, that provide for their own procedures and rules for the secondary processes of international law, particularly the settlement of disputes or the enforcement of primary norms.[132] In this sense, the general secondary rules of international law (those relating to state responsibility, treaty law, and similar) are understood to be residual, and thus states are free to establish alternatives within a particular treaty regime.[133]

The characterisation and legal effect of such regimes is dependent, as noted above, on the interpretation of the treaty that purports to set up such a self-contained regime. The key question is not merely whether alternative mechanisms for dispute settlement or other secondary rules are provided for in the new regime, but whether these mechanisms and the terms of the treaty were intended by states to exclude the application of general secondary rules of international law.[134] The existence of such mechanisms does not in and of itself imply that general secondary rules are inapplicable; they could also be viewed as alternatives or as a ‘fall-back’ option.[135] It must therefore be determined, through interpretation of the treaty at hand, whether states intended specifically to exclude the operation of other rules of international law. This is not necessarily an all-or-everything question: despite the reference to self-contained regimes as a whole, the core issue is whether particular rules of general international law have been excluded.

The preceding sections have outlined a number of elements that are key to understanding the position of EU law and specifically the CJEU in international law. It has established that there is a great deal of scope for normative and interpretive conflict in international law, and that international law as a discipline acknowledges and accounts for this lack of consistency. One means by which it does so is through the operation of common, general, secondary rules of international law, including the rules related to treaty law and state responsibility. From an international perspective, this is how interpretive and normative conflict is dealt with in international law. This would suggest that, if the EU Treaties are to be treated as ‘normal’ treaty arrangements under international law, prima facie there is no immediately apparent reason that the CJEU would enjoy exclusive or superior interpretive authority in relation to other ICTs, even over EU law.

Even these general international law rules, however, can be ‘contracted out’ of by states, through the operation of self-contained regimes.[136] The following section examines whether and to what extent the EU Treaties have ‘contracted out’ of the general international law rules sketched in this section, and therefore whether and to what extent they establish the exclusive jurisdiction of the CJEU (as outlined in Section 2 of this article) as a matter of public international law.

3.2.  The limits of CJEU exclusivity

3.2.1.   Limits of law: how self-contained is the EU legal order?

The EU has on occasion been referred to as a self-contained or special regime as described above.[137] It is often repeated that the EU Treaties have ‘created [their] own legal system’,[138] which ‘constitutes a new legal order of international law’.[139] EU law provides for highly-developed mechanisms for dispute settlement between Member States for disagreements related to the interpretation and application of EU law, most notably the CJEU itself.[140] It also provides for sanctions in the case of non-compliance,[141] for lawmaking procedures that are binding upon Member States,[142] and for remedies for those injured by a breach of EU law (extracontractual liability).[143]

Taken together, these elements amount to a highly complex and developed legal system, complete with secondary rules alternative to those in general international law. As a result, Lavranos has argued that EU ‘law must be regarded as a separate legal order that does not belong to the international or national legal order.’[144] As discussed above, however, this is not a conclusion that can be arrived at without further consideration. The existence of these rules and mechanisms does not in and of itself exclude the operation of general international law, unless and only to the extent that it may be said, on interpretation of the Treaties, that the state parties to the treaty intended for such exclusion.[145]

To a certain degree, the EU Treaties themselves thus remain part and parcel of international law, in that they must be interpreted through and in the context of international law. The EU Treaties are themselves subject to the general rule of treaty interpretation under Article 31 VCLT.[146] The CJEU has itself referred to the 1969 Vienna Convention on a number of occasions when interpreting external agreements, as well as the Vienna Convention on the Law of Treaties between States and International Organizations or between International Organizations (1986).[147] This is not to say that the CJEU accepts the applicability of the VCLTs as a matter of law – indeed, it has been reluctant to do so on several occasions, especially when it relates to the interpretation of the Treaties themselves.[148] Nonetheless, it remains true, from a public international law perspective, that the EU Treaties are treaties subject to the rules of the VCLTs.[149] It is important to recognise that the CJEU’s refusal to apply the VCLT when interpreting EU law is a feature of its internal constitutional reasoning, not a limitation on the Treaties’ status under public international law. For international lawyers, the Treaties remain subject to the ordinary rules of treaty interpretation in the VCLT, regardless of the interpretive autonomy claimed by the CJEU.

When it comes to state responsibility, the EU does establish many of its own rules with regard to sanctions, remedies, and liability. To the extent that ‘EU law operates a complete system of remedies, … there is no fallback’ to general rules on state responsibility.’[150] Despite the comprehensive nature of EU rules on state responsibility, there are some ‘exceptional’ situations in which the general international law on state responsibility would nonetheless apply.[151] These exceptional circumstances would likely include, for example, serious breaches of peremptory norms;[152] aspects of state responsibility not explicitly governed by EU law;[153] and in certain circumstances, and countermeasures.[154] As such, although there are elements of EU law that are ‘self-contained’, it is not such as to exclude all rules of general international law relating to state responsibility.

As such, it is clear that the EU Treaties continue to operate within the broader legal framework of international law, and that at least some general rules of public international law apply thereto. Returning to the claimed exclusive jurisdiction of the CJEU, the core question for present purposes is whether the EU Treaties definitively and completely exclude the jurisdiction of other ICTs over the interpretation and/or application of EU law; in other words, whether the exclusive jurisdiction of the CJEU has a legal basis from a public international law perspective. The following two subsections examine and assess the two avenues of CJEU exclusive jurisdiction as discussed in Section 2.

a) Limits of Art 344 TFEU

As discussed earlier, Article 344 TFEU is often referred to as (part of) the legal basis for the exclusive jurisdiction of the CJEU over the interpretation of EU law, even in cases which do not involve a dispute between Member States.[155] However, the provision itself does not appear to exclude the jurisdiction of other ICTs. This is for a number of reasons.

First, the provision is directed towards Member States, not towards other states, nor towards ICTs. This aligns with the rule in Article 34 VCLT, namely that ‘[a] treaty does not create either obligations or rights for a third State without its consent.’[156] Article 344 TFEU cannot, therefore prohibit non-Member States from bringing a claim before an ICT that may involve the interpretation of EU law. Of course, non-Member States are not bound by EU law and would therefore not have standing to bring a case before an ICT alleging a breach of EU law, but EU law could nonetheless form part of a dispute between a Member State and a non-Member State, for example if the EU member state raised EU law obligations as some form of justification for conduct that would otherwise be unlawful,[157] or alleged a normative conflict between EU law and international law, or if EU law were integrated in the interpretation of the state’s international law obligations under Article 31(3)(c) VCLT.[158] Thus, provided that the dispute at hand is not brought by a Member State, but rather by a non-Member State, Article 344 TFEU would arguably present no direct bar to jurisdiction, regardless of whether dispute involved the interpretation of EU law.

There is nothing in Article 344 TFEU that refers to a Member State participating in such a case once it is brought, e.g. by submitting arguments or even making counterclaims; only submitting a dispute is prohibited. Thus, a reading that focused on the ‘ordinary meaning to be given to the terms of the treaty’[159] would not prohibit such participation. A more teleological, object-and-purpose-focused reading[160] might equate counterclaims with the submission of a dispute, but would surely not go so far as to prohibit participation as such, and thus would allow arguments to be raised by the Member State in the context of a case brought by a non-Member State before an ICT that was not the CJEU (although the duty of loyalty would at the same time require the Member State to raise and defend the relevance of EU law in such proceedings).[161]

Turning to the role of the ICT itself, Article 34 of the 1986 Vienna Convention on the Law of Treaties similarly provides that a treaty does not create obligations for third organizations (which would include other ICTs).[162] Article 344 TFEU therefore does not bind ICTs as such. Thus, even if a Member State did violate Article 344 TFEU by bringing a dispute concerning EU law before an ICT other than the CJEU, it does not necessarily follow that that ICT would be prohibited from exercising its jurisdiction. Although the CJEU’s case law often frames autonomy in terms that suggest an exclusive interpretive authority over EU law, this understanding operates only within the EU legal order and cannot, as a matter of international law, bind other ICTs. The Court is not purporting to impose obligations on third states or on international tribunals, and its rulings under Article 344 TFEU regulate only the conduct of the Member States. Nonetheless, the Court’s articulation of autonomy inevitably produces external effects, as it asserts that only the CJEU may definitively interpret EU law. Other ICTs, however, remain free to interpret EU law where it is relevant to the disputes before them, and nothing in international law requires them to defer to the CJEU’s position. This divergence between the Court’s internal constitutional reasoning and the international legal framework underscores the broader tension at the heart of the exclusivity claim.

One must also bear in mind that jurisdiction is granted to other ICTs by way of treaty, either the statute of the ICT itself,[163] a compromissory clause,[164] or some other international agreement. Even if Article 344 TFEU were indeed interpreted as excluding the jurisdiction of other ICTs over the interpretation of EU law, that could place it in normative conflict with the legal instruments granting jurisdiction to the particular ICT. As noted above, EU law is not inherently superior to other rules of international law in such cases of conflict, and it would be a question of interpretation in particular cases of conflict as to which rule were applicable.

A final point to be made in relation to Article 344 TFEU is that it only refers to the submission of disputes, and does not prohibit the exercise of advisory functions of an ICT, such as the capacity of the ICJ and the ITLOS to give advisory opinions.[165] Such advisory proceedings could overlap with and involve the interpretation of EU law in any context in which EU law provides for substantive legal obligations, from free trade to the protection of the environment, and would fall outside of the scope of Article 344 TFEU as they are neither a dispute, nor are they submitted by Member States.

The CJEU’s interpretation of Article 344 TFEU is thus not binding on other ICTs.[166] Other adjudicative bodies are therefore free to interpret Article 344 TFEU as they see fit. From a public international law perspective, as just demonstrated, Article 344 TFEU does not appear to establish the exclusive jurisdiction of the CJEU. It may prohibit Member States from bringing disputes related to the application of EU law before another ICT, but it does not prevent advisory proceedings nor disputes brought by non-Member States that involve the interpretation of EU law.

b) Limits of CJEU broader exclusive jurisdiction over the interpretation of EU law

The other core argument underlying the exclusive jurisdiction of the CJEU is that it is necessitated by the nature of EU law and the EU legal order more broadly as discussed in Section 2. This, again, is a question of treaty interpretation. At its heart, the argument posits that it is essential that EU law (both the Treaties and secondary legislation) be interpreted in such a way that preserves the unity and consistency of EU law, and that the CJEU is uniquely positioned to be able to carry out that task given the specific features of the EU judicial system, including the preliminary ruling procedure.

That the CJEU is uniquely well positioned to be able to interpret EU law is relatively uncontroversial. Its competence to interpret and apply EU law and its central role within the EU legal order is clear from the terms of Article 19(1) TEU referred to above. That does not, however, make it the only body that is able to interpret EU law. The core question is whether, by agreeing to the Treaties and the legal order that they set up, Member States agreed to exclude the jurisdiction of other ICTs by necessary implication, on the basis that EU law must be consistently and uniformly interpreted for the sake of legal certainty.

In this respect, it is difficult to see how the interest of legal certainty more generally is sufficient to necessarily imply the exclusive jurisdiction of the CJEU. Legal certainty is a value intrinsic to all legal systems, including international law, and other subsystems of international law (especially international criminal law, where the criminal sanction creates an even stronger role for the principles of legality and legal certainty).[167] Nonetheless, this interest in legal certainty throughout international law does not have the legal effect of overcoming the general rules related to the jurisdiction of ICTs. In other words, it does not necessitate the exclusive jurisdiction of any particular adjudicative body. International law does not, for example, provide for exclusive jurisdiction of the WTO Dispute Settlement Mechanism over international trade law purely based on the need for legal certainty in inter-state trade relations; nor does the International Tribunal for the Law of the Sea have exclusive jurisdiction over the interpretation and application of the UNCLOS, nor does, as discussed above, the CERD Committee have exclusive jurisdiction over the interpretation of the CERD, despite the obvious desirability of legal certainty in the application of international human rights law.[168] This can, as discussed above, lead to divergence in interpretation and conflicting judicial decisions between different ICTs; this is nonetheless accepted as an inevitable part of an international legal order in which multiple ICTs, some of which have general jurisdiction, exist. Uncertainty and inconsistency may indeed be an undesirable outcome, but undesirable outcomes are not generally prohibited in international law.

The question then is whether the specific institutional and judicial set-up of the EU, including the preliminary reference procedure, necessarily implies a uniformity of interpretation and therefore the exclusive jurisdiction of the CJEU. It is submitted here that it does not. In the words of Cannizzaro, the CJEU’s position is ‘based on a misconception of the task assigned’ to ICTs established under international agreements.[169] They are tasked primarily with the dispute resolution within their specific mandate for which reason they necessarily interpret relevant national and international law: without the ability to do so ‘the jurisdiction of international courts and tribunals would be disrupted at its roots’.[170] Indeed, the steady proliferation of ICTs and their decisions means that may be ‘inevitable’ that their mandates increasingly overlap in substance with EU law.[171] Referring back to de Witte’s point above, the CJEU would not be bound by those interpretations. As such, and bearing in mind the familiarity of international law with conflicting interpretations and an absence of institutional hierarchy,[172] there is insufficient legal basis here for the exclusive jurisdiction of the CJEU over all questions of EU law. Once again, therefore, we return to relying on the level of deference given by other ICTs as discussed above.

In sum, there are no bulletproof arguments for the CJEU’s exclusive jurisdiction to be found in the ‘essential characteristics’ of EU law, nor in Article 344 TFEU; certainly not when viewed through the lens of an international legal order that is familiar with a certain degree of legal uncertainty and normative conflict. In short, from a public international law perspective, the exclusivity of the CJEU’s jurisdiction is but a paper crown. The extent to which this crown will be respected is thus entirely dependent upon other ICTs and the degree of deference that they grant to the CJEU and its claims to exclusivity. The following section therefore examines the approach taken thus far by other ICTs, and uses this as the basis upon which to assess the likelihood of future respect of the CJEU’s paper crown.

3.2.2.   Limits of deference: approaches of other ICTs to the exclusivity of the CJEU’s jurisdiction

Where there are no hard and fast legal limits to the exercise of jurisdiction by an ICT over the interpretation of EU law, then the question becomes one of deference.[173] In other words, to what extent are ICTs willing to respect the boundaries that have been said by the CJEU?; to what extent do they accept the authority of the Court’s ‘paper crown’ namely the exclusivity of its jurisdiction? Three key cases raised thus far in which such questions were considered by an ICT other than the CJEU illustrate a range of different approaches.

The MOX Plant arbitrations both concerned the same set of facts, namely complaints by Ireland against the United Kingdom regarding alleged radioactive discharges from a Mox plant in Sellafield (a nuclear site in Cumbria, England). Ireland initiated separate arbitration proceedings arguing a violation of the OSPAR Convention and the UK’s information-sharing obligations thereunder,[174] and a violation of UNCLOS due to alleged marine pollution from the plant.[175] Various aspects of EU law were potentially relevant to the proceedings, including environmental legislation and certain Euratom provisions. The UK argued that neither arbitral tribunal had jurisdiction as a result of then-Article 292 EC Treaty.[176]

The UNCLOS arbitration tribunal’s approach demonstrated a high degree of deference towards the CJEU. It noted that there were a number of ‘European Community law issues’ at stake.[177] The EC was at the relevant time also a party to UNCLOS, meaning that there was a possibility that the CJEU could itself exercise jurisdiction not only over EC law proper but also over some or all of the relevant provisions of UNCLOS.[178] The UNCLOS Tribunal viewed the EU as a ‘separate legal order’,[179] and found that the question ‘as to the scope and effects of European Community law applicable in the present dispute …[was] to be decided within the institutions of the European Communities, and particularly by the European Court of Justice.’[180] This was not for the parties to decide themselves, and ‘the Tribunal is in no better position.’[181] It held that ‘it would be inappropriate for [the Tribunal] to proceed further’ without resolution of the jurisdiction question by the CJEU.[182] To do otherwise could lead to conflicting interpretations which ‘would not be helpful to the resolution of the dispute.’[183] It thus suspended proceedings prior to the CJEU’s decision on jurisdiction.

The OSPAR arbitration tribunal, by contrast, did not consider that any source of EU law would be applicable in the dispute at hand. Instead, they focused exclusively on the interpretation of the OSPAR Convention,[184] and ‘ignored’ the potential implications of Article 292 EC Treaty.[185] Lavranos has criticized the Tribunal’s Award on the basis that, in deciding the case in the absence of EU law, it did not consider relevant EU jurisprudence on the interpretation of informational duties of states,[186] but instead made their decision in the absence of EU law and EU jurisprudence.

The Ijzeren Rijn (or ‘Iron Rhine’) case related to a Dutch-Belgian plan to reinstate an old cross-border railway line which happened to cross through a designated protected habitat area on the Dutch side. Arbitration proceedings were brought to settle the share of costs involved, with both parties acknowledging that Article 292 TEC could pose a possible barrier but with neither objecting to the Tribunal’s jurisdiction over the matter. The Ijzeren Rijn Tribunal carried out a rather detailed interpretation of EU law so as to finally find that the application of EU law was in fact not necessary in the present case.[187] Thus, even though it clearly accepted the EU’s exclusive jurisdiction over the application of EU law, it did not shy away from interpreting its scope.

From these three examples we can already see a diversity of approaches and of levels of deference to the CJEU, from the more deferential (in the case of the UNCLOS tribunal), to less (in the case of the OSPAR arbitration tribunal), and somewhere in-between (Ijzeren Rijn).[188] This demonstrates that, even where an ICT does respect the CJEU’s ‘paper crown’, as it is presented here, the approach varies between ICTs.

It must also be noted that all the cases addressed above, and indeed all cases so far in which the CJEU’s allegedly exclusive jurisdiction has been at issue thus far, have been before arbitral tribunals, and not before permanent international courts. In comparison with a body like the ICJ, such arbitral tribunals would seem much more likely to adopt a deferential approach. The ICJ, as demonstrated in the examples above in relation to the CERD Committee and the ICTY, does not shy away from adopting its own views on matters that fall within the specific competence of another body – views that can and do diverge from those of that body. Thus, despite the relatively deferential approach taken so far, it is doubtful that a permanent body like the ICJ would take as deferential a stance if confronted with an equivalent request to resolve a dispute or give an advisory opinion.

4.   Conclusion

The principle of autonomy, as articulated by the CJEU, has become a cornerstone of the EU legal order. Yet, when placed in the broader context of public international law, its foundations appear far less secure. This is particularly true in relation to the CJEU’s purported exclusive jurisdiction over all matters relating to the application and interpretation of EU law.

The CJEU’s exclusive jurisdiction has been presented here not as an established tenet of international law, but as a normatively and legally contested assertion. While the CJEU frames this exclusivity as essential to preserving the coherence and uniformity of EU law, this argumentation amounts to an insufficient legal basis when viewed through a public international law lens. From the perspective of international law, the EU remains a treaty-based entity, whose legal system does not inherently command inherent precedence over other rules of international law. Similarly, no hierarchy exists among international courts and tribunals that would oblige them to defer to the CJEU’s authority.

The CJEU’s exclusive interpretive authority functions more as a claim than a universally accepted norm – a paper crown whose force and effect depends less on legal necessity than on the choices of other actors in the international legal system. Its authority is not self-executing, but is rather contingent upon recognition by others.

That recognition has thus far been more-or-less forthcoming from other ICTs, albeit to differing degrees. It remains, however, far from certain that such deference will continue to be given, particularly by more permanent ICTs. In a fragmented legal landscape, the future of the CJEU’s exclusive jurisdiction will turn not on the robustness of its doctrinal claims (which, it has been established, are lacking), but on the degree to which they are accepted – or resisted – by the wider community of international adjudicative bodies.