1. Introduction
The autonomy of the EU legal order has for years been a highly contested concept. In 2014, the Court of Justice of the European Union (ECJ) ended – at least for the time being – the EU’s accession to the European Convention on Human Rights (ECHR), taking the position that the draft Accession Agreement of 2013 would have threatened the autonomy of the EU legal order (Opinion 2/13).1 In 2019, the Court accepted that the investment chapter in the Comprehensive Economic and Trade Agreement between the EU and Canada (CETA) is compatible with EU law and does not undermine the autonomy of the EU legal order (Opinion 1/17).2 Autonomy remains an evergreen in the debates on the very nature of EU law.
This article argues that the autonomy of the EU legal order as construed by the ECJ is best understood in conceptual legal and absolute terms. It is meant to shield the ECJ’s conceptual legal claims from interference and in this way allow the ECJ to determine the ‘mental elements needed to think about things of the relevant sort’,3 that is to authoritatively determine the concepts (mental elements) necessary to think about all aspects of EU law (things of the relevant sort). Understood in this way, autonomy concerns the authority to determine the validity and interpretation of EU law as a self-contained and self-referential legal system distinguishable and independent from national and international law. Legal autonomy as construed by the ECJ is not relative as many authors have claimed.4 It has the purpose and effect of creating the jurisdictional element of sovereignty and that cannot come about in an incremental or relative manner. It cannot be based on arguments relating to the status of a self-contained regime of international law that gradually distances itself from the general rules of international law. By contrast, it is a conceptual claim giving birth to the assumption of apriority that can only be made in categorical terms. The way of coming about of the autonomy claim can be compared to a polity’s claim to statehood and sovereignty, which if it is accepted triggers a whole set of rights and consequences under international law. The question that the Court answered in Opinions 2/13 and 1/175 is not one of gradation (how autonomous is the EU legal order?). The Court answered the question of whether the EU’s participation in international legal regimes, such as the ECHR and CETA, may be capable of undermining the absolute conceptual legal autonomy of the EU legal order – no more and no less.
The ECJ held that submission to the jurisdiction of the European Court of Human Rights (ECtHR) is different from submission to the investor-state dispute settlement (ISDS) mechanism under CETA. This resulted in the questionable outcome that the EU could not submit itself to external human rights review but is able to offer an alternative route of investment protection. This article analyses the conceptual legal reasons that may justify the Court’s position on ECHR accession and CETA.
The article is structured as follows. Section 2 sets out how the autonomy of the EU legal order is best understood. It examines the ECJ’s case law in light of legal theoretical considerations and relates it to the separation thesis of Kelsen’s Pure Theory of Law. Section 3 explains why autonomy is of such relevance to the EU legal order. It indirectly explains why the Court so cautiously protects the autonomy of the EU legal order. Section 4 sets out that the autonomy of the EU legal system remains an inherently fragile construction because the Court cannot demonstrate the aprioristic character of EU law. Section 5 discusses how CETA aims to shield EU law from CETA. It examines the Court’s reasoning in Opinion 1/17 in light of the ECJ’s autonomy conception as developed in Section 2. Conclusions restate the main findings.
2. The ECJ’s autonomy conception in context
[T]he autonomy of EU law is not absolute but relative; it does not mean that EU law has ceased to depend, for its validity and effective application, on the national law of its Member States, nor that it has ceased to belong to international law.6
The quote presupposes a conception of the autonomy of EU law that is very different from the one of the ECJ. The Court’s own understanding of the autonomy of EU law is absolute, not relative. This section first explains the Court’s autonomy conception (Section 2.1) and its claim to be the only authority determining the validity and interpretation of EU law (Section 2.2). It then examines the ECJ’s autonomy conception in a broader theoretical (Section 2.3) and international context (Section 2.4), concluding that it is best understood in Kelsenian terms and that it serves in certain respects the same purpose as state sovereignty.
2.1. Relative factual autonomy and absolute conceptual autonomy
The autonomy of EU law was developed by the ECJ in a long line of case law.7 It construes EU law as autonomous, that is, not depending on national or international law for its validity.8 The Court’s position entails that the autonomy of the EU legal order requires the Court to be in the position to maintain from the authoritative perspective of EU law the claim that EU law ‘stems from an independent source of law’.9 The ECJ’s autonomy claim is a prime example of its meta-teleological approach to interpretation, pursuant to which the Court refers on a very high level of abstraction to systemic values.10
In its early case law, the Court worked with sovereigntist language. It did so explicitly (‘the Member States have limited their sovereign rights, albeit within limited fields, and have thus created a body of law which binds both their nationals and themselves’11). The Court understands EU law stemming from a separate independent source that originally derived from sovereign rights transferred from and by the Member States. However, the transfer is seen to have cut the link to the national legal order. The validity and interpretation of EU law is now autonomous and no longer rooted in and depending on the sovereignty of the Member States. This internal autonomy, that is autonomy from national law and the authority of national courts, gives the ECJ the tools to call on national courts to ensure the effectiveness and uniform application of EU law across the EU in the different national contexts.
With the EU becoming a powerful international actor, the ECJ’s focus shifted towards external autonomy and the relationship between EU law and international law. The Court held that EU law must be autonomous from international law in that obligations under international agreements can neither ‘affect the allocation of powers fixed by the Treaties’,12 nor ‘have the effect of prejudicing the constitutional principles of the [Treaties]’.13 This is in line with the codified Treaty objective of safeguarding the ‘values, fundamental interests, security, independence and integrity’14 of the Union.
As a minimum, the autonomy of the EU legal order, as construed by the ECJ, requires that national and international law and the interpretations offered by national courts and international courts and tribunals do not interfere with the power division or legal principles set out in the Treaties. Yet the Court goes further: it bases its reasoning on an understanding of the EU legal order as self-contained, self-referential and self-sufficient. By doing so, many critics have argued, the Court establishes an unjustified and illegitimate legal autarky in order to protect its own institutional powers to determine the validity and final interpretation of EU law.15 Advocate General Bot picked this point up in Opinion 1/17. He explained that autonomy is not autarchy, but left open what this precisely means.16 The criticism seems to start from a misconception. It seems to fail to distinguish between absolute conceptual (legal) autonomy and relative factual autonomy.
Relative factual autonomy can be limited and is necessarily limited in an interconnected and interdependent world. Several reasonable points can be raised to demonstrate that institutionally and formally legally the autonomy of the EU legal order is not absolute, but that elements of interlocking dependence and possibilities for pressure exist in practice. The prime example is the joint power of all 28 Member States to amend the EU Treaties, including the powers of the ECJ. However, this, as such, does not make the conceptual legal autonomy of the EU legal order under the Treaties a matter of gradation. The Court’s autonomy conception is concerned with the latter.
Autonomy in this conceptual sense ensures the Court’s jurisdictional authority as the final authority within this complete epistemic system. It allows the Court to protect the EU legal order from normative interference that could be ‘liable to adversely affect the specific characteristics of EU law and its autonomy’.17 However, this should by no means be read as precluding international law from having effects within the EU legal order. It only establishes that the relationship between EU law and international law and the effects of the latter within the EU legal order can only be determined ‘by reference to [the EU’s] internal rules’.18 It leaves intact, as a matter of principle, the Union’s ability to bind itself under international law. Within the Court’s conception of autonomy the core question is whether any of these international obligations have the normative force to undermine the conceptual legal autonomy of the EU legal order.
2.2. Monopoly of final interpretation
The ECJ’s conception as explained in the previous section refers to conceptual legal (normative), rather than factual (empirical) autonomy. The latter refers to the factual ability to govern, adopt policies and implement them. The former requires that the interpretation of EU law must be established (and can only be challenged) by reference to EU law (within the conceptual framework established by EU law) and not by reference to other legal spheres, be they national or international. Autonomy allows the ECJ to insist on the monopoly to determine the validity of all EU law, independent from the legal concepts used by international law or the different national legal orders of the Member States.19
This self-contained nature of EU law and the monopoly of interpretation rest on the use of precedents by the ECJ.20 The internal validity of all legal interpretations is in a precedent system established in a self-referential manner, that is, by reference to earlier case law. Any of these interpretations can be challenged. Many are challenged. However, they can only be challenged by reference to EU law as interpreted by the Court. Hence, one could conclude that the autonomy of EU law as a self-contained and self-referential normative system is implicitly reaffirmed in every decision that claims its legality and legitimacy by resting on earlier interpretations of the law. In other words, the ECJ’s entire body of case law reaffirms the self-contained reading of EU law.
The ECJ’s evaluations of the EU’s attempts to submit to the jurisdiction of an international court and tribunal only bring the potential consequences of this self-contained nature to the fore. Ultimately, as we will see, it is the inherent weakness of construing a legal system that is created under international law as domestic that leads the ECJ to protect the conceptual autonomy of EU law so cautiously. The ECJ’s decisions on the EU’s ability to submit to the jurisdiction of international courts and tribunals are based on a stress test logic of whether external normative commitments may have the potential to undermine the Court-construed autonomy of the EU legal order.
2.3. EU law as the poster child of Kelsen’s Pure Theory of Law
The ECJ’s understanding of EU law seems to be an illustration of Kelsenian theory in judicial practice. Hans Kelsen’s work stands for positivism and monism.21 He construed all legal orders and all bodies of law as one single monist legal system. The positivist dimension of his work is that he assumed a ‘pure’ form of law based on a separation from both morality and fact.22 This understanding of law stands in opposition to Hart’s empirico-positivist theory that reads law and sociological fact as intimately connected. Kelsen strove to de-politicise and liberate law from ideology and construe it as a ‘borderless medium for creating and guiding society’.23
2.3.1. Positivism
Law, by definition, tends to keep some distance from empirical fact. It offers the normative framework through which the judge examines the facts and so determines what facts are legally relevant and irrelevant. For the EU in particular, a degree of separation is necessary for construing EU law as creating a ‘new legal order’. In comparison with national law, EU law is in many ways stripped of a clearly identified and identifiable cultural, societal and historical context. The best visual illustration is perhaps the Euro banknotes depicting stylised bridges. Furthermore, the social fact dimension of law traditionally draws on the state’s monopoly of force to compel compliance. EU law only benefits from a limited level of facticity in this sense.24 It does not possess a military or police force, and relies for its enforcement in principle on national authorities. High dependence on national courts and national administrations means that EU law does not possess the factual element of force to the same degree as it is usually possessed by sovereign States. This is another reason why the Kelsenian theory of law fits EU law particularly well. It allows it in the formal reading of the ECJ not to acknowledge its own weak facticity but to evoke respect and normative force beyond factual compliance. In other words, EU law exemplifies the separation between law and fact that underlies the two-way separation of pure law in Kelsen’s work. This separation is, as we will see below, the core argument that supports compatibility of CETA with EU law.25
EU law can also in other aspects be seen as taking Kelsen’s theory to the extreme. The neofunctionalist dynamics in EU integration make the European Union ‘the prime example of Kelsen’s doctrine of the evolutionary centralization of legal systems’.26 Kelsen predicted within his theoretical framework of pure law that increasing production of norms and their implementation would over time make international law resemble administrative law.27 This is precisely the situation in the EU today. While being of international origin, EU law directly regulates rights and obligations of individuals in a way largely comparable to national administrative law.
The ECJ’s powerful position and self-conception also reflects the court-centric understanding of the Pure Theory of Law. The great significance that Kelsen attached to courts and judicial review flows from his concept of the dynamic hierarchical structure of the legal system.28 Even if, arguably, Kelsen considered the use of vague norms by the judiciary as inappropriate and warned about it, in particular in the context of constitutional review, that is when a court decides on the legality of a law.29 He considered that the constellation of vague norms and constitutional review grants courts ‘an absoluteness of power which must be experienced as unbearable’.30
The ECJ’s emphasis on abstract teleological interpretation31 and the lack of attention paid to the adoption history goes well with Kelsen’s separation of law and fact. Relying on Habermas’s critical reading of Kelsen’s work, it can be seen as an elaborate architectonic construction of hierarchically arranged levels of norms, in which the persons who initially authored these norms are eclipsed soon after the act of authorship.32 This ‘uncoupling’ from ‘the moral and natural person’ (in the EU represented in the European Parliament and through the Council), opens the door for purposive or ‘purely functional’ interpretation of rights,33 of which the ECJ is regularly accused and which may come at the expense of historical considerations on the will of the authors.34
2.3.2. Internal monism and external dualism
Monism and dualism are conceptual tools to describe the relationship between different legal orders. Theoretically, these concepts developed from debates on ‘the nature of law, nature of international law, the overlap between international law and domestic law, constitutional law, and more importantly, on the authority of international law’.35 Dualism could be seen as ‘the easiest and most drastic form of norm conflict avoidance’.36 It allows domestic courts in essence to ignore ‘external’ legal orders. More recently the concept of pluralism has gained much traction. It focuses on the same conceptual issues as monism/dualism.37 Pluralism only places a different emphasis. It draws attention to the plurality of voices in the debate, the merit of this plurality, the system-inherent possibilities of contestation that flows from this plurality, and often argues in favour of the normative value of this plurality. Dualism, by contrast, focuses on the position of the individual interlocutor, that is, the individual legal orders.
Within the EU legal order, that is vis-à-vis national law, the ECJ adheres closely to Kelsen’s monism. A combination of primacy and direct effect results in a dynamic hierarchy as Kelsen constructed it. After his realist turn in the second edition of the Pure Theory of Law in the 1960s, his work should be read as requiring that one norm gives way and that the other takes priority, crucially, however, still requiring a formal act of derogation. In other words, that the lower-ranking incompatible act is no longer void but voidable. This fits very well with the ECJ’s doctrines of primacy and direct effect. Formally legally, the ECJ cannot (of course!) rule on the validity of national law, it can only find an incompatibility of national law with EU law and oblige national courts to disapply conflicting national law.
Many national constitutions presume that ‘states are logically prior to the law that they acknowledge’ (aprioristic character of the state).38 This presumption also lies at the structural basis of international law. Within the EU, the parallel to the ‘originality of states’ is the ECJ’s conception that Member States have transferred sovereign rights and that as a consequence the EU legal order ‘stems from an independent source of law’.39 The Court reasons that the EU legal order can only achieve its objectives and exist as it is construed under the Treaties if we assume this independence. Based on this therefore necessary precondition, the ECJ reads the EU Treaties to establish ‘a new legal order’40 of a domestic self-contained nature.
Kelsen’s Grundnorm doctrine is well known but very elusive. Two things are important to realise: first, the Grundnorm determines only validity not content; second, it is a necessary precondition for legal knowledge but cannot be the object of legal knowledge.41 It is an a priori in the Kantian sense. The Grundnorm could be imagined as what lies behind the historically first constitution. For the EU, this would be the transfer of sovereign rights from the Member States that vested the EU institutions with the power to issue valid norms in an autonomous legal system. It is there because it is logically necessary. It is a constitutive normative assumption that gives EU law its authority. As developed above, the very basis of the self-contained nature of EU law is the autonomy of the EU legal order. It is rooted in Kelsenian thought.
Kelsen tried to overturn the conception of international law as derived from sovereign will of States and construe law as a ‘social technique’ to change social reality.42 Kelsen wanted to make international law ‘the “universal” medium of the law available for novel political experiments’.43 This is precisely the step that the ECJ took with its ‘transfer of sovereign rights’ thesis, which disconnects EU law from direct influence of national law. The autonomy of the EU legal order is the conceptual means to allow an ‘objective’ construction of EU law, independent from the will/law of the Member States, individually or jointly, that allows it to rein in national (political) excesses.
The ECJ’s external approach differs from its internal monist approach. It takes a dualist perspective vis-à-vis international law.44 The fact that the ECJ approaches international law with a dualist logic, while expecting national courts to buy into its monist construction of the EU legal order, could be criticised for inconsistency.45 However, good reasons can be advanced in favour of this distinction: first, international law does not offer the same democracy, human rights and rule of law guarantees to which the EU is committed. Accepting a monist reading, with international law enjoying primacy, would undermine these guarantees under EU law. Second, and for that reason, national constitutional courts would not accept the primacy of EU law if the ECJ extended its monist reading to international law.46 Whether these reasons played a role in the motivation of the Court is difficult to say.
In Opinion 1/17, the Court formulated that the test of whether the autonomy of the EU legal order is affected hinges on the question of whether CETA ‘prevent[ed] the union from operating in accordance with [its] constitutional framework’.47 It states that the autonomy ‘resides in the fact that the Union possesses a constitutional framework that is unique to it’.48 The Court claims substantive autonomy (non-interference with substantive choices in the Treaties) by presupposing that the EU legal order possesses a unique constitutional framework based on the Charter and Article 2 TEU.
This may remind us of the campfire image used by Tom Flynn,49 pursuant to which each legal order is an epistemic system that operates in a way that only those within the reach of the campfire feel its normative heat, while those on the outside see the fire and acknowledge its existence without feeling the heat, that is, without being subject to its normative force. The campfire image is a pluralist image of many campfires that exist independent from each other and do not affect each other’s normative heat.
2.4. Creating a part of what constitutes state sovereignty
The Union is not conceived as sovereign. Under international law it does not possess the rights associated with sovereignty. States do. The Permanent Court of International Justice held in the Lotus Case50 that ‘[r]estrictions on the independence of States cannot ... be presumed’.51 This Lotus presumption is a general departure point for public international law.52 For States, the conceptual legal autonomy of their legal orders forms part of and is justified by their sovereignty.53 This is the jurisdictional element of sovereignty, which flows from the settled principle that the application of international law in domestic courts is for domestic legal systems to determine in ‘complete freedom’.54 In other words, the formal autonomy of a national legal order does not require an independent definition or justification. In its autonomy narrative, the ECJ makes for the EU a claim precisely to what forms the jurisdictional element of state sovereignty: to determine the law of the land and exercise self-referential jurisdiction.
The ECJ’s claim may or may not be rooted in the Treaties and the Union’s objectives, and may or may not be accepted in practice by national courts. It certainly does not flow as easily from the status of the Union as – at best – a regional economic integration organisation (REIO), which, while it enjoys additional rights in certain contexts, is still an international organisation. International law’s categorising influences the legal thinking about the EU. This is different from direct legal consequences. It offers ‘a pre-existing language and a pre-existing system of interpreting the world’ that we need to ‘move within it if we wish to be heard and understood’ that creates a system of judicial socialisation.55 In other words, it creates as a system of judicial socialisation. ‘It conveys to us a certain interpretation of the social reality to which it is addressed under the veil of objectivity, or naturalness.’56
The ECJ does not and also cannot draw on these interpretations of social reality but can only justify its sovereigntist self-conception of the Union with a reasoning valid under the logic of EU law. Yet, as we will see, this conception of the Union is not shared by national or international courts – perhaps it is even most openly rejected in the context of investment law.57
At the same time, it does not straightforwardly go against international law. In its advisory Opinion on the Legality of the Use by a State of Nuclear Weapons in Armed Conflict, the International Court of Justice (ICJ) accepted a form of autonomy of international organisations.58 It held that constituent instruments of international organisations are ‘conventional and at the same time institutional’ and that ‘their object is to create new subjects of law endowed with a certain autonomy, to which the parties entrust the task of realizing common goals.59 The ICJ spoke of ‘a “system” designed to organize international cooperation in a coherent fashion by bringing [international organisations], invested with powers of general scope, into relationship with various autonomous and complementary organizations, invested with sectorial powers’.60 In other words, the ECJ’s position that the autonomy of the EU legal order flows from its Treaties and objectives, while it may not be endorsed in all its consequences, does not apparently contradict the ICJ’s position of linking the autonomy of international organisations to their powers as conferred on them by the parties.
Sovereignty and autonomy, as construed by the ECJ, share the assumption of an ability to exercise jurisdiction over a territory in a self-referential and formally independent manner. Autonomy, hence, while not entailing all rights associated with sovereignty, serves in certain contexts a similar function as sovereignty. Kelsen developed a doctrine of the identity of law and state61 as a counterpoint to the position of Carl Schmitt, who advocated for the priority of the political concerns of the state. Again, Kelsen’s understanding chimes well with the ECJ’s interpretation of the EU as a legal construction of a polity62 in which sovereignty is decoupled from the will of the personified state. If we accept the identity of law and polity for the Union, protecting the self-contained nature of the EU legal order is about protecting the existence of the Union. The next section supports this point by developing the particular relevance of autonomy for the Union.
3. The relevance of autonomy
This section explains why the autonomy of the EU legal order is a necessary precondition for the particular nature of EU law (Section 3.1) and why it is the basis of the ECJ’s claim that EU law is domestic law and can therefore take a dualist approach to international law (Section 3.2).
3.1. Autonomy as an essential characteristic of EU Law
In Opinion 1/17 the Court explains that ‘autonomy … stems from the essential characteristics of the European Union and its law’.63 This position presupposes a particular nature of the EU and its law that leads to the condition of autonomy. In terms of judicial argumentation, the order of basing autonomy on the essential characteristics – the effectiveness of EU law based on primacy, direct effect and the preliminary ruling procedure – may not be surprising as the former are less controversial than the latter. However, it raises questions of circularity and mutual dependence.
Autonomy is also an essential feature of the EU legal order, arguably the presupposed condition that allowed the Court to develop the essential characteristics of primacy and direct effect. The term ‘essential’ is consciously used in this context. The point is that in the absence of an essential element, here EU law’s claim to autonomy, EU law would no longer function in the way EU law functions. If EU law were directly dependent in its validity and interpretation on national and international law (without the ECJ being in the position to filter and determine the precise effects of national and international law within the EU legal order) it would no longer enjoy the same effectiveness. By definition, this is a thesis that excludes the opposite claim, that is that the very nature of EU law would not change in the absence of an in-and-of-itself-valid claim of autonomy. Arguing that the formal autonomy claim of the ECJ is essential also entails that it cannot be compromised in face of other considerations.
Balancing or compromising the absolute conceptual legal autonomy is not possible without changing the nature of EU law and threatening the equilibrium of powers within the EU legal order. A high degree of acceptance of the essential characteristics, primacy and direct effect exists among all involved actors, including national courts and governments. They are considered ‘essential’ in the sense that EU law would not be the same if they structurally did not work or were structurally not accepted. What is most contested is in fact the necessary precondition, namely the autonomy of the EU legal order, which also removes its further development from the direct influence of national (or international) courts.
This also explains the ECJ’s cautious position aimed at ensuring under all circumstances that the autonomy of the EU legal order remains protected, most apparent in Opinion 2/13. In this Opinion, the Court approaches the draft Accession Agreement as if it were a prenuptial agreement: it must rule out all probability of a challenge to that autonomy rather than assessing the balance of probability as to whether such a challenge may occur.
3.2. Autonomy as a precondition for the court’s dualist position
The main feature of dualism appears to be that international law and [domestic] laws are viewed as separate legal systems, which may be defined as self-contained, because within each system the only existing rules are those that are part of the system. Rules which are not created within the system may nevertheless be relevant for the system if they are referred to by a rule included in the system.64
This quotation nicely demonstrates why the autonomy of the EU legal order is also a conceptual precondition for the ECJ’s dualist approach to international law. Only if the EU legal order is a self-contained normatively complete system can it decide which norms of international law have internal effect.
As with most national constitutions, the EU Treaties leave many questions open. The autonomy claim allows the ECJ to depart in all its case law from the standpoint that EU law is domestic law. This becomes particularly apparent in its case law concerning the autonomy vis-à-vis international law. The assumed self-contained nature of EU law allows thinking of EU law as domestic law and introducing dualist elements into the reception of international law within the EU legal order.
Opinion 2/13, in which the ECJ rejected the draft agreement of EU accession to the ECHR, was characterised by some as a new era of closing off and isolating the EU legal order.65 Six years earlier, however, the ECJ held in Kadi (2008) that the ‘autonomous’ nature of the EU legal order ‘cannot be prejudiced by an international agreement’66 and was strongly criticised for taking the European legal order as the one and only frame of reference, changing its international law-friendly attitude and moving to a more ‘dualist’ position than before.67 I would argue that the Court has taken the same standpoint all along, namely, that the EU legal order is autonomous. Only the specific consequences of this position differ greatly in the circumstances of the individual case.
At the same time, the strictly construed independence from national and international law raises many questions; similar to those with which the pluralism literature is grappling: Is it possible to conceive of the jurisdictions in Europe (national, international, EU) as fundamentally different, genuinely alternative epistemic systems, without accepting ‘facts by virtue of which one of these systems is more correct than any of the others’?68 Can the other legal systems then be more than facts (lacking the normativity of law)? Does constitutionalism require monism?69
Significant authoritative voices deny the EU’s domestic nature and the autonomy of its legal order. The pressure from international law to dissolve EU law into international law increases with the intensified interaction between EU and international law. The next section lays out both the actual and the potential internal and external contestations.
4. Contesting the Union’s original rights construction
The ECJ’s construction of the EU as a polity with original rights is contested and contestable. Its autonomy narrative is vulnerable first and foremost because it cannot demonstrate but must claim the aprioristic character of EU law. Sovereign states, by contrast, automatically benefit from the presumption that their constitution is of aprioristic character and hence does not require another original source justifying its validity. Their exclusive right to decide what acts shall have effect on their territory is virtually undisputed under international law. It functions as an independent, overriding justification.70 For the EU this is very different. It needs to justify not only competence to regulate, but also its claim to autonomy. The inherent contestability of its nature is what distinguishes the EU from other international actors, be they (federal) States, whose sovereignty is accepted, or international organisations, whose institutions or organs do not make the same essential claim to autonomy.
The next subsection explains the subsisting internal contestations by national courts (Section 4.1); how the different claims about the nature of EU law remain incommensurable but keep each other in check (Section 4.2); and how a challenge within the logic of EU law differs from the existing contestations by national courts from the standpoint of national law (Section 4.3). Finally, the last subsection discusses the exceptional conceptual authority of the ECtHR and concludes that EU accession to the ECHR could potentially have undermined the autonomy of the EU legal order (Section 4.4).
4.1. Internal pushback in a vulnerable equilibrium of constitutional claims
National constitutions vest national constitutional courts with a mandate to ensure the continuous existence of the polity, as construed by the constitution. This includes protecting the sovereignty (and autonomy) of the national polity.71 This also flows from the very nature of state sovereignty. It is inherent in sovereignty that the sovereign fights against being made irrelevant, against usurpation by external forces.72
National constitutional courts adhere to a conception of sovereignty that is necessarily and inextricably connected to statehood. The German Federal Constitutional Court (BVerG) is traditionally the most vocal national constitutional court in explicating limits to integration as it sees them to flow from the German Constitution.73 Other national courts, however, are even more adamant than the BVerG. The Danish Supreme Court considered in its Maastricht Treaty decision that if too many competences were transferred to the EU, Denmark’s statehood would be threatened.74 Similarly, the Spanish Constitutional Tribunal stated that the limits imposed by the Spanish Constitution to what competences can be transferred to the EU are intended to protect Spanish sovereignty.75 The case law of a large majority of national constitutional courts is based on the understanding that national law imposes limits on the development of the EU legal order. This is very different from the ‘original rights’ position of the ECJ. In exceptional circumstances, individual highest national courts have even flatly refused to give effect to a preliminary ruling that they considered as going against those limits. An example is the case of Ajos (2016).76 In this case, the Danish Supreme Court refused to disapply the incompatible national provision on the basis of the general principle of EU law prohibiting discrimination on grounds of age. It considered that the ECJ had created the general principle in a manner that could not be justified on the basis of methods of interpretation. Ajos may be classified as the most direct and confrontational judicial challenge so far.77
What these contestations demonstrate is that national courts are in practice willing to challenge the ECJ’s authority and how it construes the EU legal order. National courts can do so effectively because they do not form part of a hierarchical legal and judicial structure. The validity of their decisions does not depend on compliance with Union law. The combination of actual contestation and the inherent fragility of a system dependent on a willingness to cooperate explains the ECJ’s cautious approach. At the same time, the challenges of national courts have also demonstrated that the system can outlast incidental challenges that are put forward within the logic of national constitutional systems rather than EU law.
4.2. Equilibrium of irreconcilable claims
The irreconcilable claims of the ECJ on the one hand and national constitutional courts on the other are central to the balance of powers between the EU and its Member States. Their authoritative force depends on the ability of each of the judicial actors to make a convincing claim within the logic of their epistemic system, that is their own legal order. Each claim is internally consistent and rests on an entire body of constitutional norms and case law that also enjoys factual support (compliance). At the same time, the claims are irreconcilable and unable to relate to an ‘objective’ external mutual point of reference. Each claim tells how the domestic law of the court making the claim relates to and affects the law of the others. The national positions necessarily deny (albeit implicitly) the autonomy of EU law.
The EU autonomy claim and the national counterclaims are necessary to preclude a normative hierarchy between EU and national law.78 The existence of incommensurable claims is necessary, if not constitutive, for the EU to rely on pluralist sources of legitimacy.79 It is what maintains the irreconcilable pluralist legal situation, in which, in the words of Neil MacCormick,
[n]one of the Member States is indebted to the union for the terms or the provisions of its constitution. Each has a constitution whose roots and whose basic legitimacy are independent of any grant by the Union. ... Conversely, the validity of the Union’s constitution and legal order is not derivative from the validating power of any state’s constitutional order.80
The point is precisely that none of the claims trumps the other or can be shown to be ‘more valid’. They cannot be reconciled because they argue within different epistemic systems and do not have a common point of reference. The point is that the EU is built on an equilibrium of irreconcilable claims about the nature of EU law. It requires the pluralism of competing claims to autonomy to continue to exist in the way that it functions at present.
Each side has the formal legal possibility to hijack or challenge the status quo, in which different claims coexist. Member States possess the political power and national courts possess the institutional power to rein in the EU and the ECJ. The ECJ possesses the legal power to determine the interpretation of EU law. Substantive convergence is possible and is happening. However, the absence of normative hierarchy excludes unity of law and makes real conflict between the norms of both legal orders possible.
4.3. Challenges within the logic of EU Law
While internally unranked and irreconcilable claims of autonomy and authority exist side by side, externally the stakes and constraints are different. First, national and EU actors interlock and overlap in the sense that Member States are institutionally represented within the EU. Second, national and EU law interlock in a tight embrace, where national law gives effect to EU law and EU law has recourse to national constitutional traditions. The close interlocking of actors and law is probably best illustrated by the formalised judicial cooperation between the Court of Justice and national courts in the preliminary ruling procedure. Third, all actors involved have a multifaceted stake in a well-functioning and effective Union. Indeed, within the EU legal order, any structural non-compliance or fundamental challenge comes at great cost.
Externally, by contrast, the stakes are very different. Non-EU actors, including international courts and tribunals, are game changers for the cooperative relations of the Union and its Member States. They are not part of the interlocking embrace. They do not share the same interest in the functioning of the European project. They are not subject to the primacy of EU law nor the jurisdiction of the ECJ. Primacy requires national actors to disapply a rule of national law that infringes EU law. This allows national courts to hold national administrations and governments in check. Non-EU actors are, by contrast, under no general obligation to disapply an international rule that goes against EU law. On the contrary, they may legally be on safe and sturdy ground to ignore EU law. In Opinion 1/17, the Court confirmed the institutional outsider status of the ISDS mechanism envisaged in CETA. The Court concluded that it ‘stands outside the EU judicial system’.81 It cannot request preliminary rulings from the ECJ. This is fully in line with is previous case law.82
When the EU concludes an international agreement, this agreement becomes part of EU law and is as such binding on the EU institutions and the Member States.83 In principle, this binding force extends to the decisions of international courts and tribunals established under the agreement.84 This makes it possible for national courts to rely on these decisions, including their interpretation of EU law, as if they were EU law.85 In other words, when an international court interprets EU law (as collapsed into) international law and denies its domestic nature, this gives national courts a valid conceptual argument to challenge the ECJ’s autonomy claim by reference to a joint framework of reference. It allows the national court to rely on the international tribunal’s interpretation, binding on the ECJ, to argue that (certain obligations of) international law prevail(s) over EU law. This is the crux of the threat of decisions of international courts and tribunals for the autonomy of the EU legal order. How problematic this may be in the individual case depends on the formal legal authority and the hierarchical status of the interpretation of the international court or tribunal in question.
4.4. The exceptional position of the European Court of Human Rights
International courts and tribunals approach the Union through the lens of international law. International law is steeped in the concept of state sovereignty. It sees the Union as an international organisation, or at best as an REIO, which, while it enjoys additional rights in certain contexts, is still an international organisation. While it can legally bind itself, in principle on eye-to-eye with States, the Union is ‘neither sovereign nor equal’.86 It remains from the perspective of international law at least partially penetrable, in that behind the organisation there are still the Member States. The Member States retain the legal capacity to conclude binding agreements with regard to their territory irrespective of EU law, including international agreements concluded by the EU.
The ECtHR sees the EU as an ‘international organization’ to which States ‘have transferred part of their sovereignty’.87 It refers to EU law as ‘international legal obligations’ of the contracting parties.88 Indeed, the ECtHR’s deference to the ECJ in its Bosphorus doctrine is based on its understanding that EU law is of an international nature and that the ECJ is an international court.89
The ECtHR enjoys an exceptional authoritative status within the EU legal order. EU fundamental rights, which are interpreted in line with the ECHR and the case law of the ECtHR, are part of the ‘foundations’ of EU law. This is confirmed by Article 6(3) TEU and Articles 52(3) and 53 Charter of Fundamental Rights. In the appeal case of Kadi I, the ECJ held that all derogations under the EU Treaties must comply with the foundations of EU law. In other words, the Member States’ right of derogation is limited by these foundations.90 In other cases, the Court evaluated primary law in light of the ECHR.91 The case law allows the conclusion that authoritative claims under the ECHR rank above ‘ordinary’ primary law.92
Because of its authoritative status, the ECtHR is the most illustrative example of how international courts and tribunals may threaten the autonomy conception of the ECJ. A ruling of the ECtHR engaging with the relationship between EU and international law or between EU and national law would enjoy exceptional authoritative status within the EU legal order. The interpretation of EU law in a ruling of the ECtHR may be formally legally limited because the ECtHR considers EU law as fact93 and because, strictly speaking, ECtHR rulings are binding only inter partes. Yet, because of the constitutional status of the ECHR and the ECtHR’s case law in primary law and in the ECJ’s own case law, the ECtHR’s rulings should be seen as conceptually exceptionally authoritative. Their persuasive force within the EU legal order can hardly be overestimated. Furthermore, the ECtHR’s analysis of human rights and the attribution of responsibility may not only be based on a particular understanding of how the EU legal order works and interrelates to other legal spheres, but these considerations may in fact be inseparable from the human rights analysis.
National courts could be placed in a position where they are confronted with interpretations of the ECJ and of an international court or tribunal, both binding under EU law, that take a very different view of the nature of EU law and the ensuing relationship with international law. They would have to reconcile different normative positions if they are relevant to the outcome of the case. Pluralism in the individual case cannot logically exist.94 If national courts gave priority to the interpretation of the international court or tribunal over the interpretation of the ECJ this would challenge the authority of the ECJ. It might also mean that the national courts would follow the view that EU law forms part of ordinary international law and that it therefore cannot claim primacy over international obligations within the national legal orders. A situation in which national courts were placed between a rock and a hard place (almost) arose in the context of individual sanctions.95 National courts were confronted with UN obligations that breached fundamental rights. The ECtHR held that the UN obligations prevailed over the Convention (because the Convention is international law and cannot claim dualism vis-à-vis other international law). In Kadi (2008) the ECJ took a dualist position and gave priority to EU fundamental rights over the Member States’ UN obligations. If the ECtHR had also considered EU law and held that the UN obligation prevailed also over EU law (because the ECtHR considers it international law) national courts could potentially have ended up in a situation where they had to take a position on which interpretation to follow.96 When asked to rule on the matter the UK Supreme Court avoided any consideration of diverting interpretations by ignoring EU law.97
The potential force of a challenge to the authority of the ECJ resulting from a ruling of the ECtHR that is picked up by national courts would be of a very different quality than any potential challenge presented by national courts within the logic of their own national legal system. The ECHR and the ECtHR’s case law would work as the missing common conceptual point of reference that allows the different legal orders to be integrated logically within one hierarchy. This would undermine the ECJ’s ability to maintain the internal validity of its autonomy claim.
CETA aimed to avoid the same trap as the draft agreement on EU accession to the ECHR. In the words of AG Bot, CETA intended to create two ‘co-existing legal systems’.98
5. CETA’s separatism and Opinion 1/17
CETA aims to protect the independence of EU law, which allows the ECJ to reason within a dualist logic. The separation provision of Article 8.31.2 takes a four-pronged approach. It states, first, the ISDS mechanism ‘shall not have jurisdiction to determine the legality of a measure, alleged to constitute a breach of [CETA]’; second, it ‘may consider, as appropriate, the domestic law of a Party as a matter of fact’; third, it ‘shall follow the prevailing interpretation given to the domestic law by the courts or authorities of that Party’; and, fourth, ‘any meaning given to domestic law by the Tribunal shall not be binding upon the courts or the authorities of that Party’.
Both the Advocate General and the Court found the safeguards in CETA sufficient to ensure that a clear-cut distinction could be made between domestic courts applying domestic law and the ISDS mechanism applying CETA.99 They were convinced that the ISDS mechanism does not rule on EU law. The question of whether this position is justifiable hinges on whether CETA succeeds in formally protecting the conceptual legal autonomy of EU law and allowing the ECJ to continue to reason within a dualist logic vis-à-vis international law.
The following five subsections examine the different safeguard mechanisms. The final subsection concludes that the ECJ’s position in Opinion 1/17 was coherent but can still be criticised (Section 5.5).
5.1. No decision on legality and law as fact
The stipulation that the ISDS mechanism cannot determine the legality of EU law and considers EU law ‘as a matter of fact’100 is central to protecting EU law’s conceptual legal autonomy. An interpretation as a matter of fact cannot conceptually interfere with the interpretation of, or determine the validity/legality of, law.
Considering domestic law as fact is a widespread practice of international courts and tribunals. The ICJ considers national law as fact. The ECtHR considers the domestic law of the contracting parties as fact. This may also be the reason why the draft Accession Agreement did not address this point. The fact that CETA stipulates that its tribunals consider domestic law, including EU law, presupposes and hence confirms the domestic nature of EU law. It buys into the logical framework that EU law is domestic rather than international law.
Formally legally, treating EU law as fact protects the position of the ECJ as the only authority to determine the validity/legality and interpretation of EU law. This does not preclude the ISDS mechanism from offering an opinion on the interpretation or even validity of EU law. The ECJ’s position is fully in line with its reasoning in the case of Kadi (2008).101 In this case, the ECJ took the position that reviewing and annulling an EU legal instrument that gave effect to a Security Council resolution did not call into question the UN measure but was in all its formal legal effects limited to the EU legal order. The ECJ’s conception is based on the above developed separation of law and fact. Formally legally speaking, the ECJ did not rule on the UN measure; yet, when the ECJ declares an EU law measure that blindly gives effect to UN measures contrary to human rights to which both the EU and the UN are committed this cannot be seen but as entailing a negative judgement on the legitimacy of the UN measure. The ECJ’s conception rejects a lifeworld control of the system, to borrow Habermas’s terminology.102 Because of its formal conception of law the ECJ does not have to consider that the awards of the ISDS mechanism may very well have a factual influence on the interpretation of EU law.103
The separation of law and fact and the separation of two legal spheres should be distinguished. The former builds on positivism. It aims to protect the normativity of law from lifeworld control, that is power, politics and factual coercion. The latter is a dualist conception, in which the effects of one legal sphere within the other are determined pursuant to the internal rules of that legal sphere. The determination that the ISDS mechanism under CETA considers EU law as a matter of fact builds on the former to achieve the latter. Formally stipulating that the ISDS mechanism under CETA cannot make normative claims on the interpretation of EU law explicates that no valid normative claim under EU law can be based on the interpretation of the ISDS mechanism. The ECJ determines the effect and status of international law within the EU legal order to the extent that they are not stipulated either in the EU Treaties or in the international agreement. However, because of the lack of a hierarchical relationship with national courts, the ECJ would not directly be able to ensure that national courts follow its position. Hence, the formal stipulation that the interpretations of EU law by the ISDS mechanism do not enjoy normative authority is an attempt to ensure that national courts cannot attribute such authority to these interpretations.
5.2. Interpretation of EU law
The ISDS mechanism must and will at times interpret EU law, that is give meaning to its wording in order to determine its effects and so assess whether EU law breaches CETA.104 However, it ‘does not give judgments but rather issues awards’105 and is bound to the ‘prevailing interpretation’ by the ECJ, and even more importantly the interpretation of the tribunal in this context does not have any conceptual legal meaning. It cannot be raised before the ECJ as authoritative in any way.
This is what Opinion 1/17 must mean when it states that the ISDS mechanism’s examination ‘cannot be classified as equivalent to an interpretation … of domestic law’.106 The Opinion distinguishes between a substantive legal dimension and an institutional dimension of autonomy. The former is protected so long as the ISDS mechanism cannot offer a conceptual legal interpretation of EU law and the latter is protected so long as it cannot prevent ‘the EU institutions from operating in accordance with the EU constitutional framework’.107 In particular, the formulation of the second (institutional) leg of autonomy begs the question at what point looming factual consequences (damages) could constitute a limitation of the operative discretion of the EU institutions under the EU constitutional framework. In turn, this would raise the question whether this could ever qualify as (factually) interfering with the EU institutions’ autonomy.
5.3. Exclusion of bindingness and direct effect
Excluding the binding force of any meaning given to domestic law is a reconfirmation that any necessarily occurring interpretation of EU law by the ISDS mechanism cannot have normative force within the EU legal order and on the ECJ. It protects separatism from a different angle, leading to the same result.
All of the EU’s international investment agreements explicitly exclude direct effect, subject to a stipulation that this shall not affect the enforcement of awards. CETA is no exception.108 Ironically, it should be added the lack of direct effect of CETA was also put forward as a justification for the need for investment protection.109
The lack of direct effect excludes annulling secondary EU law because it breaches CETA but it does not exclude a general requirement of consistent interpretation as a form of indirect effect. However, this is limited to interpretation consistent with CETA and the ISDS mechanism’s interpretation of CETA. The exclusion of normative force of the EU law interpretations of the ISDS mechanism excludes consistent interpretation with interpretations of EU law.
5.4. Protection of internal power division
Article 8.21 CETA allows the Union to determine whether the investor who is bringing a claim to the ISDS mechanism seeks to challenge a measure of the Union or a Member State. This protects the exclusive jurisdiction of the Court over the internal power division. As the Court points out,110 this is an important distinction between CETA and the draft Accession Agreement. In the latter, the ECtHR retained the power to exceptionally determine the correct co-respondent.111
5.5. Opinion 1/17: Coherent but not beyond criticism
The intended shielding of EU law from any conceptual effects within the EU legal order is meant to allow the ECJ to continue to act as a domestic court and take a dualist approach. The separatism in CETA is in particular aimed to protect the ECJ from any potential claims of national actors based on normative effects flowing from the decisions of the ISDS mechanism.
Generally, international agreements do not regulate the issue of dualism or monism. They only occasionally determine whether the agreement can have direct effect or be self-executing.112 Domestic courts based on the conceptual framework of domestic law develop their position vis-à-vis international law. The ECJ ruled that the strategies of normatively disconnecting EU law and the ECJ from the decisions of the ISDS mechanism (EU law as fact; ECJ not bound by decision) sufficiently addressed the conceptual autonomy concern. In addition, making the ECJ’s view authoritative (ISDS mechanism must follow prevailing interpretation of EU law) aimed at lowering the likelihood of substantive differences in interpretation. Excluding the bindingness of the interpretation of domestic law and direct effect are aimed at excluding challenges based on decisions of the ISDS mechanism. However, neither of these safeguard strategies addresses the removal of disputes from the institutional backbone of the interlocking embrace of EU law and courts on the one hand, and national law and courts on the other: the preliminary ruling procedure. The ECJ decided that this risk can be taken and that it does not necessarily amount to a threat to the autonomy of EU law. The Court’s position can be very well reconciled with its previous rulings on autonomy and is in line with its overall Kelsenian understanding of law as at least separable from fact. At the same time, the fact that the Court accepted the safeguard clauses as sufficient indicates that it did not take the same cautious prenuptial agreement approach it took in Opinion 2/13.
Kelsen sought to liberate law from sociological, ethical, political and psychological insights. He rejected the possibility that a unification of the insights could be possible. This separation thesis has been challenged by many, including most prominently Habermas and also Alexy.113 In Between Facts and Norms 114 Habermas explores how a purely normative theory of society, state and law, which capriciously locks itself on the inner perspective of the communicating actors of a lifeworld, remains unreflected and weak. It does not capture but ignores the cold, disenchanting external views of sociological observers, which are necessary to grasp the social fact dimension of law. This makes it ultimately normatively empty as it cannot respond to the moral concerns of persons. Habermas, however, also rejects the radical one-sidedness of the mere external observer’s gaze as revealing the full truth. It disregards the context-transcending claim of law to express valid moral norms. In Habermas’s own work, both facticity and (normative) validity are interconnected in a reciprocal need for one another. Law is the medium closes the gap between the two.
Habermas’s criticism is also applicable to the ECJ’s separatism. It ignores the fact that social and economic forces that lie beyond the control of the EU may create imperatives that, despite the fact that they have no formal legal authoritative effects within the EU legal order, influence the interpretation and understanding of law. A particular multinational corporation may exercise undue factual influence on the normative capacity of a State or the EU to regulate its domestic affairs, for example by threatening to withdraw its business activity and hence its jobs and taxes. The ambition of the ISDS mechanism and possibly the Multinational Investment Court (MIC) goes further. It aims to increase publicness and coherence by creating a system whose decisions are meant to have precedent-like effects stretching across the globe and carrying even into disputes under different international agreements. The individual award, potentially with major financial implications in the individual case, is meant in this way to throw a shadow far beyond the individual dispute. In addition, the outcome that the EU cannot expose itself to the human rights jurisdiction of the ECtHR but can submit itself to an ISDS mechanism may appear normatively questionable.
6. Conclusions
This article offered a conceptual reading of the ECJ’s use of autonomy. This reading should be understood as a lens that allows us to see more and different things when analysing the Court’s case law. The article’s absolute conceptual understanding of autonomy explains the Court’s positions in Opinions 1/17 and 2/13, as well as earlier case law on autonomy. It may also be a helpful tool to consider in future treaty making intended to set up international courts or tribunals.
At the same time, the article, perhaps more indirectly, justified the ECJ’s conception of autonomy. The Court’s concern with autonomy is plausible if the autonomous a priori nature of the EU’s legal order is an essential characteristic of EU law. Assuming that autonomy is a necessary – albeit not sufficient – condition for EU law to exercise self-referential jurisdiction over a territory and its citizens in the way it does, it becomes clear why the Court takes a cautious approach to legal constructions that may undermine this autonomy. It allows the Court to claim a domestic nature for EU law that is similar to what national legal orders automatically enjoy as part of state sovereignty.
The self-contained nature of a legal order is formally complete, that is, absolute. It is protected by an independent judiciary. It protects the effects and relevance of the decisions taken within the legal and procedural framework of that legal order. While the legal order can vest any external norm with exceptional normative force, this force depends on the acceptance by that legal order. This continues to be the case even if the exceptional force of the external norm sets aside internal norms. The ECJ constructs in this legal conceptual sense the autonomy of the EU legal order as absolute. EU law refers to national and international law, for example, to the common constitutional traditions of the Member States and the ECHR, and draws from their concepts. Yet this does not as such undermine the self-referential character of EU law as the ECJ constructs it.
The ECJ’s conception of the autonomy of EU law resides in an essentially contested presumption, namely that EU law stems from an independent origin and does not depend for its validity on either national or international law. The ECJ’s autonomy claim and national courts’ perspectives denying the autonomy of the EU legal order and construing it as framed and dependent on international and national law are equally valid. All courts reason within their own epistemic systems. Maintaining the ECJ’s autonomy claim as equally valid requires that it cannot be epistemically challenged in a system-inherent manner. The Court’s conception of autonomy focuses on the need to protect the integrity of the EU legal order as an epistemic system. If we accept that the EU is a legal construction and can only exist through law, protecting the conceptual autonomy of EU law is about protecting the existence of the EU as such.
This article shed some light on the underlying theoretical considerations of the Court’s autonomy claim and the essential relevance of being able to maintain that claim as internally coherent. It explained why Opinion 1/17 is coherent with the Court’s previous case law on autonomy. A dualist understanding of the ECJ vis-à-vis international law is in principle accepted.115 What has led to stronger criticism are the consequences of the Court’s conception of autonomy, for example for the EU’s accession to the ECHR. However, this article argued that dualism without conceptual autonomy is not possible. CETA met the Court’s requirements of protecting the autonomy of the EU legal order because it introduced a formal legal separatism in order to protect the conceptual autonomy of EU law.
The Court’s conception of autonomy in Opinion 1/17, however, is reasonably subject to all the Habermasian criticism of Kelsen’s positivism. In its formal positivist conception of law, it focuses on law’s abstract conceptual quality, while being able to ignore the factual pressures that also have an effect on the interpretation of law. To make it more concrete, it seems unconvincing that the threat of partially very costly awards has no bearing on the considerations of judges and the interpretation of EU law. At the same time, it is not perhaps for the Court to act upon this realisation.