Legal uncertainty is a universal feature of primary legal materials in all legal systems. The need for judicial interpretation arises from the fact of legal certainty in legal texts. In theory, judicial interpretation is a process which should be governed and constrained by legal methodology. The law must impose methodological constraint, because in the absence of methodological constraint the judicial process would be little more than arbitrary.
The judicial process in national or domestic legal systems is governed by broad criteria and traditions specific to each system, although in practice there is considerable convergence as to the various approaches judges may apply. Formally Treaty Interpretation in international law is governed by the rules of the Vienna Convention on the Law of Treaties.1 Unfortunately, however, just as legal uncertainty can never be wholly eliminated in primarily legal materials—whether in statutes or in treaties—there is no legal system where the accepted rules of interpretation and those governing their application are not likewise subject to legal uncertainty. Judicial discretion therefore is a fact of judicial life. The academic literature nonetheless distinguishes between courts whose interpretative approach is primarily text based and those which more liberally draw on other alternative criteria, especially teleological and policy criteria. Courts of the former type seek to minimize judicial discretion, typically ideal for cases where the text itself is ambiguous, whilst courts of the latter type expand judicial discretion beyond the sphere of textual uncertainty and seek to interpret legal instruments with reference both to the text and its underlying objects and purposes, which may be construed narrowly or more widely, and not necessarily with giving primacy to the former. Courts of the latter type are therefore also referred to as ‘activist’.
This article argues that the Court of Justice of the EU (CJEU) firmly belongs to the second of those categories of courts and, most notably in cases raising fundamental issues involving the distribution of legal competences between the European institutions and those of EU Member States, favours an activist purpose-based and gap-filling approach, which maximizes judicial discretion and, in case of conflict, often prioritizes the purposes of European integration over a more text-based interpretation, where the latter may suggest a less integrationist outcome. Universal rules of interpretation as enshrined in the Vienna Convention on the Law of Treaties are thus relied upon on a selective—if not instrumentalist—basis, as the next sections will show.
II. The sources of legal uncertainty
There are, broadly speaking, four types of legal uncertainty:
Linguistic vagueness which includes conceptual uncertainty and other linguistic ambiguities.
Value pluralism which, in the law, manifests itself in norm collision. Norm collision arises when the answer to a legal question may be governed by several applicable rules and/or principles which may conflict and suggest different interpretative outcomes, and it is unclear which rule or principle should take precedence because their normative status in the hierarchy of norms may be equal or indeterminate in relation to each other. Norm conflict commonly arises in cases where courts are asked to balance conflicting human rights, such as the conflicting demands of the right to privacy and freedom of expression.2 Other obvious examples of such norm collision are the conflictual relationship between any of the four freedoms of movement in EU law and fundamental rights in EU law3 or conflicts between any of the four freedoms of movement and the treaty- or case law-based public policy exceptions to these freedoms.4 Both in EU and ECHR [European Convention on Human Rights] law, such conflicts, either between conflicting rights or between rights and freedoms, are frequently settled with reference to the principle of proportionality. The vagueness of standards such as proportionality, however, means that an instance of norm collision is simply transformed into a balancing act involving the application of a conceptually uncertain standard which, by itself, imposes no, or little, constraint on judicial discretion.
Rule instability, which is specific to precedents5—because precedents, unlike written rules, have no stable and agreed wording. In English law, the ratio of a case is formally binding although it may be stated at different levels of abstraction and restated with different words; in other legal systems and EU law, precedents are treated as principles which are not formally binding and the precise wording of which is likewise not fixed and which also evolve, expand and contract, and are adapted over time. Rule instability must be distinguished from vagueness which, like norm collision, equally applies to written and judge-made laws. As stated above, vagueness is here used to refer to the linguistic uncertainty of particular words or phrases. By contrast, rule instability refers to uncertainty arising from changes in the specific judicial formulation of a rule or principle, and not as uncertainty arising from the words themselves. It may therefore also be described as wording instability, which only applies to rules or interpretations of rules established by case law, because written rules are fixed until they are amended.
Gaps in the law, where a court has to decide an issue which is governed by no obvious primary legal materials or precedents.
These four factors exist in all legal systems, and they explain why law is often uncertain as well as the need for judicial interpretation. Often, these types of legal uncertainty overlap. Norms of indeterminate status relative to each other conflict, additionally each norm may be vague in one or more respects, and previous judicial interpretations of the applicable norms may have removed some uncertainties of interpretation but given rise to new conceptual uncertainties and, moreover, it may not be clear whether their scope is confined to specific issues or general principles. Precedents may, in addition, also function as general authoritative interpretations, which complement or effectively replace the wording of a written rule with a judicial interpretation of it.
Not even the most skilled and careful draftsman can fully escape the effects of linguistic vagueness and the other forms of legal uncertainty. The sources of legal uncertainty explain why judicial interpretation cannot yield certainty. Primary legal uncertainty is reproduced at the secondary level of judicial interpretation for several reasons. First, there is no one generally accepted interpretative method or even methods. All there are, are multiple interpretative criteria: literal, purposive and systemic arguments, and in addition historical, comparative, and quasi-logical ones as well as maxims of general practical reasoning, such as a fortiori or ad absurdum arguments and, failing everything else, common sense and everyday reasoning. Secondly, the literal meaning of a legal provision is often not in doubt, yet if concepts are unclear so will be there literal meaning. And just as most law students, who go to university to obtain a law degree, do not see the degree as the end purpose, but as a purpose which is at the same time also a means to another, more long-term objective, such as making a living, practising law, or as a stepping stone to another career, legal rules rarely have just one purpose but several, some of which may be immediate and text-based, whilst others will be long-term and not obvious from the legislation at all. And where there are applicable precedents, these themselves need to be applied to new factual scenarios by reference to their literal meaning or purposes, and therefore pose the added problem of being capable of being stated and restated at different levels of abstraction, of being distinguished or applied by analogy, of being narrowed or widened and, generally, of evolving over time. Finally, the key problem is that for the application of the various interpretative criteria or topoi, there are no clear rules, or meta-rules which tell the judge how to apply which criterion in which circumstances or, more generally, in what order of priority the interpretative criteria are to be applied in each case.
Judges, in other words, have no method which tells them that in the case before them they have to apply the natural or technical meaning, whilst in another case a teleological approach is to be preferred and so forth: the process of legal interpretation leaves them free to choose and decide how to resolve the uncertainty, often on a case by case basis. The persistence of legal uncertainty at the secondary level of judicial interpretation raises two obvious issues:
First, legal reasoning cannot be scientific or methodological, in the sense that its outcomes can be verified by a shared or accepted method of argumentation. It is and can only be heuristic. By that I mean that the good lawyer, trying to advise or anticipate how judges will decide, need not only master the applicable legal materials, precedents, and interpretative criteria, but also has to take account of the institutional, political, and wider context in which judicial decisions are made, and of anything he knows about the judge’s or court’s personal motivation. For these will determine how the judicial axe will fall. Practical legal reasoning, in other words, requires both an appreciation of the acceptable justificatory arguments in a case, and an evaluation of the relevant political, social, institutional, and personal background relevant to the Court’s decision making in each case. It requires one to assess how the conventions of legal argumentation, if they exist at all, and the justificatory discourse constrain the Court, if indeed it does so at all, to say what it is inclined to say taking account of any extra-legal constraints on its motivation. It follows that one can reckon, but not predict, what a court will do, because there is no formula which guides the decision making of even a wholly unbiased court.
Secondly, if there is no methodological certainty, or, at any rate, if there are no methodolodical boundaries for judicial reasoning and decision making, this would mean that legal interpretation is ultimately just a matter of discretion and convenience and that judges cannot be criticized for getting the law wrong. And if judges cannot be shown to get the law wrong, then the idea of the rule of law is at best a ‘noble lie’. Thus, if law is to be more than either politics or judicial convenience, it must impose some methodological constraint. Yet, how can judicial decisions based on the legitimate exercise of judicial discretion under conditions of legal uncertainty be distinguished from judicial choices ‘as the continuation of politics by other means’? There is no generally accepted answer but the Vienna Convention on the Laws of Treaties and the different judicial approaches that have evolved under it, which point to an obvious one based on the idea that the wording of a provision where clear should be conclusive.6 A distinction must thus be drawn between cases where the natural meaning of a legal rule is clear and those where it is not. In the second type of cases judges are not merely free but forced to exercise judicial discretion; in the first type of cases they choose to exercise judicial choice, even though they are not forced to. In the first type of case the exercise of judicial discretion is unnecessary and may be regarded as illegitimate on the grounds that the judges effectively usurp the role of legislator. In the second type of case the exercise of judicial discretion may be unavoidable, but it can be constrained by the requirement that the relevant legal context be confined to the legislation or treaty of which the legal provision in issue forms part and that the range of relevant objects and purposes be limited to those found in the text and that teleological considerations may not be invoked to support a contra legem interpretation.
Judges cease merely to make law and begin to exceed the judicial function of applying, rather than making law, when they override the natural meaning of a legal provision or its obvious purpose disclosed by the provision and its legal context, or, where there is no applicable legal provision and they start filling the gaps without declining jurisdiction, or depart from their own precedents where these have been formulated in accordance with appropriately defined interpretative rules or parameters. Inevitably, the methodological constraints hinted at can and must be refined further and cannot be exhaustive. Furthermore, grey areas will inevitably remain, for instance where the wording is unclear and there is no one clear purpose, or there are conflicting norms of equal normative status. Nevertheless, the distinction between the necessary and usurped exercise of judicial discretion, and that between decisions which respect the natural meaning of legal rules and methodological constraints and decisions which extend the exercise of judicial discretion beyond those constraints provides a useful framework for categorizing different judicial approaches even in view of the inescapable fact of legal uncertainty.
If the judicial function is above all to apply the law and only make law where the rules are unclear, then the distinction between the necessary and unnecessary exercise of judicial discretion may also be appropriately described as the difference between legitimate judicial law-making and judicial law-breaking. However, the purpose of this article is not to set out a comprehensive theory of the legitimate exercise of the judicial function, but rather to show that if, as we must assume, law should impose methodological constraints and not do violence unto language, the CJEU adopts a judicial approach which, in cases where the Court so desires, leaves it almost entirely free from methodological constraints, and free to disregard the wording of treaty provisions where it sees fit. In its near-complete freedom ‘from’ constraints and ‘to’ do what it likes, the Court of Justice, on a spectrum of plausible approaches to legal argumentation sanctioned by the Vienna Convention on the Law of Treaties (VCLT), should be placed at the very extreme end of judicial activism. Indeed, unless one takes the view that the VCLT imposes no or hardly any methodological constraints, the Court of Justice may be regarded as exceeding that scope of discretion afforded by it.
III. Treaty interpretation
The general rules of treaty interpretation are set out in Articles 31 and 32 of the VCLT. The Vienna Convention rules apply, in principle, to all international courts or tribunals, irrespective of their institutional set-up, subject matter, or geographical scope.
Articles 31 and 32 VCLT offer two main principles of interpretation. The first general rule (Article 31)7 is that treaties must be interpreted in ‘good faith’, in accordance with the ‘ordinary meaning’ of the ‘terms’ or text of the treaty, in their ‘context’, and in light of the treaty’s ‘object and purpose’. The VCLT’s second supplementary principle (Article 32)8 is that the ‘preparatory work of the treaty and the circumstances of its conclusion’ are only secondary sources of interpretation, to confirm meaning established under the first principle or in case the meaning of the treaty remains unclear or leads to an absurd result.
As general rules, these VCLT rules apply to all treaties, irrespective of the subject matter, goal, or number of parties to the treaty. Article 31 makes clear that the text must be the starting point of judicial interpretation and that the words in the text must be given their ordinary meaning, although it goes on to say that that meaning is to be established in context, with reference to other textual provisions and their objects and purpose. Thus, although according to the VCLT interpretation must start with the text of the treaty and Article 31 imposes a general framework of interpretation, the Convention rules are not without ambiguity and so a certain measure of discretion remains. In particular Article 31, whilst it states that a treaty must be interpreted ‘in accordance with the ordinary meaning’ of its terms, also refers to the context and the purposes of the words in question, and it does not expressly state which of these three criteria is to prevail in cases where their application may lead to different conclusions. Article 32 VCLT offers further guidance, in that it allows the tribunal to confirm the meaning resulting from the application of Article 31 VCLT, and thus, by implication, of choosing between different possible interpretations suggested by the various criteria, by taking account of the travaux préparatoires and the circumstances of the conclusion of a treaty. However, recourse to these supplementary criteria is optional, not mandatory. In addition, the academic literature favours the view that the VCLT rules, although generally applicable, are not necessarily exhaustive, and other additional principles or guidelines may exist or can develop as part of customary international law or within specific treaty regimes. This gives international tribunals additional flexibility.9
It has been suggested that the VCLT describes treaty interpretation ‘as a holistic, non-hierarchical exercise’ which involves the ‘summing up of text, context, and purpose’, ‘albeit one that starts with the text of the treaty’.10 This view is open to challenge on the grounds that, although Article 31 does not expressly state that the ordinary meaning of the terms of the text should always prevail over its purpose and object, Article 31 clearly states that the treaty text must be given its ‘ordinary meaning’, implying that it is only when that meaning is ambiguous or ‘manifestly absurd or unreasonable’ (Article 32) that contextual or teleological criteria may prevail over the text. It is thus only for a good reason, that is, textual ambiguity, vagueness, or absurdity, that the ordinary meaning of a treaty provision may be displaced by an interpretation based on its context or underlying purpose. Article 32 further suggests that in determining the ordinary meaning of the text, courts may adopt an historical approach by looking at the preparatory works and circumstances leading to the conclusion of the treaty.
In the academic literature on the subject and in judicial practice, broadly three main schools of thought of treaty interpretation are commonly distinguished.11 These may conveniently be called the ‘intention of the parties’ school, which is also sometimes referred to as the ‘founding fathers’ school;12 the ‘textual’ or ‘ordinary meaning of the words’ school;13 and the ‘teleological’ or ‘aims and objectives’ school.14 The ideas of these schools are not necessarily exclusive of one another, and theories of treaty interpretation can be constructed (and indeed are normally held) compounded of all three.15 The VCLT and the academic literature on the subject refer to all three elements underlying these approaches. Moreover, as so often, the academic literature, much less than the text of the VCLT, does not settle which is to prevail in cases of conflicting outcomes.
The VCLT generally applies to all treaties and tribunals. However, the textual ambiguities and the lack of doctrinal guidance in the literature and case law governing the application of Articles 31 and 32 gives tribunals considerable scope for manoeuvre and it is often an open question whether, on any particular interpretative issue, a tribunal will rely primarily on: (a) the text of the treaty; (b) the intent of the parties to the treaty; or (c) the underlying objective that the treaty seeks to attain. Divergence between international tribunals in the practical application of the VCLT rules of treaty interpretation has not gone unnoticed in the academic literature.16 When applying VCLT rules, some tribunals are more likely to be guided by the text, whilst others give greater weight to other factors (especially the objective) and favour a more teleological, creative interpretative approach.
International tribunals that are associated with the textual approach are the WTO’s Appellate Body,17 or the International Court of Justice (ICJ). By contrast, examples of international tribunals that tend to favour a teleological approach are the CJEU18 and the European Court of Human Rights.19 Both these courts take an evolutionary approach to treaty interpretation in that they take legal instruments as ‘living instruments’ whose meaning is not tied to the original and historical understanding of textual terms at the time of their conclusion nor to the subjective intention of the parties as may be deduced from the preparatory works. Tribunals favouring the textual school may also be described as deferential to the parties of the treaty: they are more likely to regard themselves as agents giving effect to the intention of the parties, in particular as revealed by the text or the historical documents surrounding its conclusion. Teleological courts, by contrast, are activist and gap-filling beyond the rules provided for in the treaty. They are more likely to regard themselves as ‘self-confident agents’ operating largely independently of the parties that made the treaty and established the tribunal.20
IV. The Court of Justice’s approach
The Court of Justice itself has summarized its interpretative approach in Merck v Hauptzollamt Hamburg-Jonas as follows:
. . . in interpreting a provision of [Union] law it is necessary to consider not only its wording, but also the context in which it occurs and the objects of the rules of which it is part.21
In general terms, the CJEU at first sight appears to echo the text of Article 31 VCLT.22 Its approach to legal interpretation also appears conventional in other respects. In many respects its legal argumentation does not, at first sight, differ fundamentally from that of other courts, especially that of most national constitutional and international courts. The CJEU appears to apply many of the same techniques and arguments, although, in such broad terms, this applies to most courts in practically all modern legal systems.23 Second, the CJEU relies on the literal meaning of provisions more frequently than on any other consideration, except its own previous decisions. In 2011, the Court referred to the wording or particular provisions in nearly 97 per cent of its decisions, compared to 74 per cent for purposive arguments.24 The Court thus does not ignore but places greater emphasis on the wording than other criteria. And third, the CJEU liberally cites its own previous decisions and does so at least as frequently as other higher courts.25
Does this mean that the Court of Justice is not very different from other courts, and that uncertainty in its decision making is no greater than in other legal systems or, to the extent that the Court’s decisions may be more uncertain, that uncertainty above all reflects the greater conceptual uncertainty and pervasive norm collisions in the primary legal materials of EU law? Convenient as this conclusion may seem to be to apologists of EU integration, it is simply not true. Why? There are some features of the Court’s approach which are not apparent in its reformulation of Article 31 VCLT, nor spelled out expressly elsewhere in the Court’s extensive body of case law.
First, although the Court frequently refers to the wording, this, in itself, establishes little. The Court does so in perfunctory manner and without extensive discussion. Crucially, the Court’s respect for the wording of provisions is subject to a critical proviso: compared to many other courts, the Court of Justice is relatively more willing to give priority to teleological criteria over linguistic criteria in cases where both types of arguments conflict and a teleological argument favours an integrationist outcome or, more rarely, supports an otherwise politically convenient solution.26
Secondly, the Court of Justice rarely, if ever, uses historical arguments.27
Thirdly, the Court often implicitly and sometimes explicitly takes account of meta-teleological criteria and not merely, as it suggests in Merck, teleological considerations,28 which refer to the explicit ‘objects of the rules of which [a legal provision] forms part’.29 This applies despite the fact that the Court rarely expressly refers to the ‘ever closer union’ objective and only somewhat more frequently to the ‘spirit of the Treaties’.30 In important cases, however, the idea of ever further integration is almost always implicit, as it is inseparable from the principles of the uniform application of Union law as well as the effectiveness of Union law.31
Fourthly, as soon as the Court decides with implicit reference to ‘ever closer union’ or the principles that favour an integrationist solution, that decision effectively becomes a precedent. Because many important cases were decided on meta-teleological considerations, understood to include those principles which favour an expansive interpretation of the scope of Union law and of the competences of the EU institutions, the body of precedents itself acquires a communautaire or pro-Union flavour. The importance of de facto precedents in the Court’s argumentation is borne out by the fact that there is hardly any case in which the Court does not refer to at least one previous decision. In both 1999 and 2011 there were more cases in which it cited case law than any classical interpretative argument.32 In referring back to its case law, the Court thus implicitly also relies on meta-teleological considerations. Precedents thus solidify and reinforce the Court’s communautaire leaning. Moreover, the appeal to previous decisions enhances judicial credibility and lends later decisions the aura of legal objectivity, simply because, in analysing a case, not every relevant previous case is excavated and subjected to legal analysis.33
Fifthly, the Court of Justice operates in an extremely permissive political environment. In domestic legal systems, it is open to the legislator to override court decisions by passing appropriate legislation. Judgments by the Court of Justice, by contrast, can be reversed only by the Court itself or by unanimous treaty amendment by the Member States which, in all cases which affect or, as the Court would say, ‘clarify’ the balance of powers between the EU and its Member States, requires unanimity amongst all Member States. In other words, it is as difficult to reverse a Court of Justice judgment as it is to amend the existing Treaties. Moreover, the meaning of the Treaties, although not their wording, invariably reflects, and thus evolves, with ongoing Court judgments. The CJEU’s decision making, in turn, then builds not on the treaty text itself but on its own body of interpretative decisions. The Court of Justice interprets the Treaties as living, not historical instruments.34
Sixthly, in contrast with several other courts applying the VCLT, the Court of Justice in considering literal, purposive, and other criteria, does not accept a hierarchy amongst these criteria.35 The CJEU does not attach a clear consistent weight to specific criteria. It presents its conclusion as the cumulative result of the variable application of all criteria. The Court’s approach to legal reasoning may therefore be described either as a cumulative or variable approach.36
Finally, the Court’s variable or cumulative approach combined with its meta-teleological dimension gives the Court’s decision making its distinctive pro-Union of communautaire tendency; its predisposition to resolve legal uncertainty in favour of further integration of the Union’s interests.
The Court’s communautaire predisposition tends to be irrelevant in most run-of-the-mill cases, which concern the application of more or less clear, detailed, and technical provisions. Examples of these are mostly agriculture, VAT, Customs Union, and Tariff cases. Unsurprisingly, in these areas the Court generally pays close regard to the wording of the applicable provisions, and it is exceptional for the Court to reach a conclusion based on teleological criteria, which are opposed to a literal reading. However, the Court’s pro-Union default position becomes crucial and often the decisive factor in cases involving major issues of principle, such as constitutional issues or the division of competences between the EU and Member States. In constitutional cases the communautaire tendency inclines the Court to resolve legal uncertainty in favour of meta-teleological objectives, especially the ‘ever closer union’ objective which, implicitly, is present in many of the Court’s most influential decisions. The Court in such cases may disregard the lex specialis principle and override a more or less specific rule in a particular policy area, which may suggest a less integrationist outcome, in favour of a meta-teleological reading, based not on the ‘objects of the rules of which [the rule] forms part’37 but the general aspirations of the Union Treaties.
At the same time, it should be added, the Court’s communautaire predisposition is just that, a tendency, not an inevitable or foregone conclusion. The Court of Justice is, in fact, a politically most astute court. Where Member States’ political or budgetary sensitivities are engaged, the Court frequently adopts a compromise solution which may involve deferring to the Member States concerned, either on the facts or in law, or most commonly with reference to the flexible proportionality principle, which involves minimal constraints for future decisions and leaves the Court’s future discretion largely untouched.38
The Court of Justice’s pro-Union interpretative stance is evident in most of its decisions determining the scope of Union law and the division of competences between national and Union law, in a number of key cases widening and entrenching the scope of EU citizenship law and in its judicial response to the euro crisis.
V. The early constitutional cases
In its early grand constitutional cases from van Gend,39Costa,40 the Internationale Handelsgesellschaft,41Nold,42 through to Francovich,43 the Court was presented with a legal question to which the Treaty provided no compelling answer: it reached an integrationist outcome without textual support, in defiance of established public international law, yet not quite openly in opposition to the text.44 In these cases, the Court filled the ‘gaps’ left by the signatory Member States. ‘Gap-filling’ together with a readiness to adopt a more teleological, rather than textual approach is one of the key features of an activist court. Where, in addition, such a court does not approach each case de novo, but in the light of both the text, its purposes, and its own previous decisions, judicial decisions, especially where they are relevant to the legal order as a whole, reinforce the existing direction and, in general terms, tend to move the judicial decision-making process further down the same road. This phenomenon characterizes the evolution of EU law: the principles established in the CJEU’s principal ‘constitutional’ cases quickly became part of the foundation of EU law and have been cited or, commonly simply applied without express reference in many subsequent key cases where they have had the effect of tilting the balance generally in favour of an integrationist outcome. With each subsequent communautaire decision, the body of precedents became a degree more integrationist—the goalposts were shifted.45 The CJEU’s case law on the scope and content of EU citizenship rights provides a good example of the cumulative integrationist effect of a few cardinal initial judicial choices.
VI. EU citizenship cases
In many EU citizenship cases, however, the Court also goes one step further: it expands the right to financial and other benefits of moving EU citizens in the absence of Treaty support and, in addition, in defiance of the principle of conferral and in opposition to detailed provisions of the applicable secondary legislation. This is evident from the following examples.
Economically active EU migrant citizens, notably EU migrant workers, have long enjoyed an equal right to in-work and contributions-based benefits on par with domestic workers, which derives from the free movement of persons provisions of the EU Treaties in connection with the prohibition of discrimination on grounds of nationality.46 EU legislation adopted under the relevant treaty provisions47 provides for a more specific framework in this area, which operates as an exception, governing work-related economic and social benefits for economically active EU citizens and their family members, to the general social policy prerogative of Member States.
In contrast, neither the EU Treaties nor the relevant EU legislation provide a textual basis for such a general exception in favour of economically inactive citizens and their family members.48 Article 18 TFEU moreover limits the operation of the non-discrimination principle to those areas of law which fall within the scope of the EU’s competences under the Treaties. The Treaties do not empower the EU to adopt legislation governing benefits entitlements to EU migrants who are not workers or ex-workers. Nor does Title X TFEU49 include rules governing benefits entitlements for economically inactive EU citizens as falling within the list of social policy fields where the Union has a supporting or supplementary competence to support the policies of the Member States.50 Where the Treaties do not confer legislative competence, legislative powers remain with the Member States. It follows that, in contrast to in-work benefits, benefits entitlements for economically inactive EU migrants fall outside the scope of the non-discrimination principle. Where EU citizens leave their home country and enter another EU Member State without employment or adequate financial resources, the EU Treaties, according to the ordinary meaning of the words used, do not confer a right to social assistance in the host state. The adoption of Directive 2004/3851 has not changed the pre-2004 distribution of competences, nor could it in the absence of a treaty change.
The entry into force of the Lisbon Treaty in 2009, likewise did not extend the Union’s powers in this area. Instead, and in contradistinction to the treaty rights of workers and ex-workers and their family members, Directive 2004/38 states that EU citizens who are not ‘workers or self-employed persons in the host Member States’ only have a right to residence in another EU Member State for more than three months if they ‘have sufficient resources for themselves and their family members not to become a burden on the social assistance system of the host Member State during their period of residence and have comprehensive sickness insurance cover in the host Member State’.52
The ‘sufficient resources’ requirement does not apply to EU migrants who have acquired the right to permanent residence, who entered as workers, or who after a minimum period of employment become unemployed. It is likewise qualified with regard to retired workers who wish to remain in the host Member State.53 Article 24(1) of the Directive further states that Union citizens ‘shall enjoy equal treatment with the nationals of that Member State [only] within the scope of the Treaty’. This provision, which echoes similar provisions in the EU Charter of Fundamental Rights,54 means that any rights conferred under the Directive must not be construed as extending the boundaries of EU law or redefining the division of competences between Member States and the Union. Article 24(2) specifically provides that ‘the host Member State shall not be obliged to confer entitlement to social assistance . . . , nor shall it be obliged, prior to acquisition of the right of permanent residence, to grant maintenance aid for studies, including vocational training, consisting in student grants or student loans to persons other than workers, self-employed persons, persons who retain such status and members of their families’.
Notwithstanding the restrictive provisions of Directive 2004/38 and the general social policy competence of the Member State, the CJEU, in a series of seminal cases, nevertheless extended the right to social assistance to various groups of economically inactive Union citizens. The Court also adopted an expansive construction of the concept of Union citizenship and its rights in other less high-profile areas of law, which do not fall within the competences of the Union institutions, such as national laws governing the family names.
In Grzelczyk,55 a pre-2004 case, the Court was asked to decide if the applicant, a French student, was entitled to a ‘minimex’ maintenance grant while studying in Belgium. National law laid down that only Belgian nationals were entitled to the minimex. In consequence, the Court stated that ‘a student of Belgian nationality . . . who found himself in exactly the same circumstances as Mr Grzelczyk would satisfy the conditions for obtaining the minimex. The fact that Mr Grzelczyk is not of Belgian nationality is the only bar to its being granted to him’. However, the Court went on to opine that ‘Union citizenship is destined to become the fundamental status of nationals of the Member States enabling those who find themselves in the same situation to enjoy the same treatment in law irrespective of their nationality, subject to such exceptions as are expressly provided for’.56 Furthermore, the Court held that Union law ‘accepts a certain degree of financial solidarity between nationals of a host Member State and nationals of other Member States’.57 On those grounds, the CJEU concluded, the minimex could not be denied.
The Court’s decision may be morally attractive. However, Articles 18 and 21 TFEU refer to the ‘scope of application of EU law’ and thus restrict the application of the equal treatment principles to those areas where the Union has legislative competence as provided for in Articles 3, 4, and 6 TFEU including derogations from EU law.58 The Union enjoys such shared legislative competence in social policy matters only in so far as it falls within the scope of the Treaty. Non-work related social assistance and financial support for students do not fall into this area. Accordingly, neither the equal treatment nor the proportionality principle should have been invoked to impose constraints on Member State social policy choices.
In Grzelczyk, the CJEU implicitly extended the competences of the Union legislator. The Court affirmed the decision in Grzelczyk in the later cases of D’Hoop59 and in Ioannidis60 where, following the adoption of Directive 2004/38, the CJEU also confirmed that foreign-schooled job-seekers had the right to the Belgian allowance, notwithstanding domestic legislation that Belgian nationals had to meet certain residence requirements before qualifying for the minimex allowance. The CJEU applied the same reasoning in Bidar,61 where it held that, notwithstanding a residence requirement under national law, a migrant student who does not meet that requirement, might qualify for a subsidised student loan in the host state if, in view of other factors, the resultant exclusion is disproportionate. Bidar and Ioannidis were decided after adoption of Directive 2004/38, which limits the right to free movement of economically inactive EU citizens to workers and those with sufficient resources to maintain themselves. The CJEU disregarded this provision and instead justified its conclusions by analogical reference to the underlying rationale established in its previous decisions, including Grzelczyk and D’Hoop. The latter cases had been under the pre-Directive 2004/38 Treaty framework, which confers Union competence only in the field of in-work benefits and, by implication, leaves legislation governing social assistance to economically inactive or non self-sufficient EU migrants to the Member States.62
The CJEU has followed a similarly expansive, integrationist approach to the application of the equal treatment principle in the field of differential tax treatment for national and non-national pensioners. The decision in Rüffler63 is morally appealing, but, although it does not directly conflict with Directive 2004/38, it is without treaty basis. In accordance with the principle of conferral, legislative powers which are not transferred to the Union remain with the Member States. The Court of Justice’s decision in Rüffler thus impliedly extends the competences of the Union. In this line, the Court of Justice’s application of the equal treatment principle to different rules for nationals and EU migrants in relation to post-bankcruptcy debt relief in Radziejewski64 encroaches on the powers of the Member States for exactly the same reason.
Further examples of judicial integrationism can be found in areas other than social or welfare policy, such as national rules governing the registration of surnames.65 Again, in the circumstances the decisions appear humane, yet the EU Treaties do not confer any competence on the EU institutions in the field of national rules governing the registration of surnames. The equal treatment legislation consequently does not apply. National legislation, outside the legislative competences of the EU, falls outside the ratione materiae of Union law and the CJEU does not have a general human rights jurisdiction in matters of national law. Equally, the protection against expulsion under EU law for family members of EU citizens is confined to cross-border situations, that is, to EU citizens and their family members who have moved from one Member State to another. It does govern wholly internal situations in one Member State where the EU citizen, from whom the rights of his family members derive, has not exercised his freedom of movement. In Zambrano, however, the Court held that Article 20 TFEU precluded the expulsion of family members of an EU citizen who had never exercised his freedom of movement.66 In all but name, the CJEU again arrogated a general human rights jurisdiction to itself, for which the Treaties provide no basis.
VII. The CJEU’s euro crisis decisions
Nowhere, however, is the Court’s pro-Union bias more evident than in the litigation which has arisen out of the euro crisis. Here the Court takes the crucial step from law-making to law-breaking: it decides one way, when the Treaty says otherwise. The term ‘law-breaking’ is, of course, rarely used in academic literature and least of all with regard to courts or judges. The more common term would be mis-application of the law by judges. By law-breaking here are meant situations where a court adopts an argumentation that imposes minimal methodological constraint and maximizes judicial discretion over when and where to rely on which type of interpretative criteria in preference to others without. It equally applies to situations where putative literal arguments do not reflect the ordinary meaning of the terms of the relevant provision, or when a court adopts a teleological interpretation at variance with the ordinary meaning of the treaty, which is supported by historical evidence about the intention of the parties and does not lead to absurd, but simply a politically or economically inconvenient result, or one that threatens important and influential business and financial interests.
The first patently integrationist and political decision is Pringle.67 Before turning to the judgment, a few words must be said about the nature of the EU’s monetary union. The European single currency, known as the euro, was established with the Treaty of Maastricht, which was adopted in 1993.68 The Maastricht Treaty,69 however, bases monetary union on the basis of the principle of individual national budgetary responsibility. This means that, although Member States agreed to a common monetary policy to be conducted by the European Central Bank (ECB), they would remain responsible for the management of their own public debt levels which, inter alia, are influenced by the interest rate and other policies of the ECB. To ensure responsible financial management by Member States, a no bailout clause was inserted in the Treaties. This is Article 125 TFEU, which states that
[t]he Union shall not be liable for or assume the commitments of central governments, regional, local or other public authorities, other bodies governed by public law, or public undertakings of any Member State, without prejudice to mutual financial guarantees for the joint execution of a specific project. A Member State shall not be liable for or assume the commitments of central governments, regional, local or other public authorities, other bodies governed by public law, or public undertakings of another Member State, without prejudice to mutual financial guarantees for the joint execution of a specific project. (emphasis added)
To those who can read, Article 125 TFEU is as clear as legal provisions can be: there is to be no mutual financial assistance between Eurozone governments, except for very specific projects limited in scope. However, in the wake of the euro crisis, Euro governments very quickly began to ignore the no-bail clause, and in 2012 instituted a permanent bailout fund, the so-called European Stability Mechanism (ESM), with a total volume of €500bn and, including other funds, of €700bn.
In the Pringle case, the CJEU was asked to assess the compatibility of the ESM with Article 125 TFEU, the so-called no-bail out fund. The Court upheld the legality of the permanent rescue fund, essentially on two grounds. First, the CJEU resorted to a disingenuous, indeed deceptive literal argument: mutual financial assistance between two or more Eurozone countries, via the establishment of a rescue fund, and the assumption of existing debts of one such country by one or more other Eurozone members are two entirely different things, because ‘assistance (via a rescue fund) amounts to the creation of a new debt, owed to the ESM by that recipient Member State, which remains responsible for its commitments to its creditors in respect of its existing debts’.70 In other words, the ‘no bailout’ clause does not forbid assistance given through an intermediary.
In essence, the Court here confines Article 125 TFEU to cases where the existing debt of one country is assigned to another Member State, so that that state steps into the shoes of the original debtor and assumes legal liability for its pre-existing debt. Any other form of mutual financial assistance involving a transfer of the default risk from the original debtor to the assisting state(s), but without a formal assignation of one and the same debt, the Court of Justice in effect maintains, is outside the scope of the ‘no-bail out’ clause. If, as the Court evidently concludes, Article 125 TFEU was never intended to prevent the transfer of financial risk between Eurozone governments, then the so-called ‘“no-bail out” clause’ does little or nothing to restrict the mutualization of debt within the Eurozone.
Not only is it unprecedented for the CJEU to rely on an extreme and, as has been shown, specious literal interpretation which, at the same time, effectively reduces a central Treaty provision to absurdity/redundancy. The Court’s supposed literal interpretation even fails on its own terms. Article 125 TFEU, on a literal interpretation is not restricted to direct bailouts between states. The words ‘shall not be liable or assume’ cover situations where the default of one country triggers a legally binding promise of support by another, though the resulting obligation may be legally distinct. Moreover, under the ESM Treaty,71 Eurozone members guarantee loans and guarantees given by the ESM, according to a contribution formula equivalent to their shares in the capital of the European Central Bank. However, if one Eurozone member is unable to honour its commitments, then, according to Article 25 ESM Treaty, it falls to the remaining ESM members to assume the shortfall.72 They thus assume the commitments of the failing Member State—precisely the ‘assumption of liability’ which, even on the Court of Justice’s view, is prohibited by Article 125 TFEU. The supposedly literal interpretation of Article 125 TFEU patently conflicts with an equally literal interpretation of the ESM Treaty.
Secondly, the difficulties associated with the Court’s only apparently literal interpretation may explain why the Court also refers to ‘the preparatory work relating to the Treaty of Maastricht’ which, it states, discloses ‘that the aim of Article 125 TFEU is to ensure that the Member States follow a sound budgetary policy?’.73 The ESM, the Court continues, takes account of this objective in that it links the award of financial assistance to ‘strict conditionality . . . [designed to] ensure that the Member States pursue a sound budgetary policy’.74 The ESM, the Court thus concludes, complies with Article 125 TFEU, as it ‘ensures that the Member States remain subject to the logic of the market when they enter into debt, since that ought to prompt them to maintain budgetary discipline. Compliance with such discipline contributes at Union level to the attainment of a higher objective, namely maintaining the financial stability of the monetary union’.75
It should be noted that Article 125 TFEU has one obvious aim—to prevent mutual financial assistance—and that aim is clearly set out. That aim admittedly, in turn, is designed to ensure a sound monetary policy, just as any immediate aim usually has another more long-term objective. Yet, in invoking that further goal, what the Court ignores is that the EU Treaties make a very clear choice as to how ‘sound budgetary policy’ is to be achieved, which is apparent from the wording of Articles 123, 125, and 127 TFEU—that budgetary discipline was to be achieved not through the mutualization of debt, but via individual national responsibility for public debt.
The outcome in Pringle came as no surprise either to judicial commentators, politicians or the financial markets. The Court essentially upheld a done political deal.76 To give this convenient conclusion at least the semblance of legal credibility, the Court relies on two types of arguments it usually employs only extremely rarely: a putative literal argument, which is both at variance with the ordinary meaning of Article 125 TFEU and Article 25 ESM Treaty, and thus is no literal argument at all. Second, it refers to the preparatory works as support for its teleological reading—something it practically never does, as the Court commonly chooses to present the Treaties as evolving and not as historical documents to be interpreted with the subjective intention of the signatories77 As regards the Court’s teleological argument, that argument both conflicts with the textual choice that sound budgetary policy is to be achieved through self-reliance, not mutual assistance, and it runs counter to basic economic theory and empirical psychological evidence that the mutualization of debt reduces, rather than enhances incentives for budgetary discipline.
The more convincing construction of the rationale of Article 125 TFEU, it is submitted, would have been an ordinary language reading, according to which ‘the assumption of the commitments’ of one Member State by another would have been taken to refer to, and strictly prohibit, any de facto transfer of the financial risk of public debt between Member States, save where expressly provided for in the Treaties and strictly limited to the purposes of those exceptional provisions.
An equally overtly political and integrationist decision by the CJEU in defiance of the Treaty is the 2015 Gauweiler judgment.78 In Gauweiler, the Court had to consider whether the European Central Bank’s first unlimited government bond buy programme, the so-called OMT, fell within the ECB’s mandate. Under the Treaties, the ECB may only conduct monetary policy, whilst economic policy is the prerogative of the Member States. The EU institutions may support the economic policies of the Member States, but the ECB must not stray from its Treaty objective, which is to combat inflation and ensure price stability. The key issues in Gauweiler were: first, are unlimited government bond buys a lawful monetary policy instrument under the EU Treaties? And, secondly, do the EU Treaties authorize the ECB to conduct monetary policy aimed at ‘safeguarding an appropriate monetary policy transmission and the singleness of the monetary policy’. The Court held, first, that bond buys were monetary policy and thus part of the ECB’s mandate and that, for those reasons, ‘safeguarding an appropriate monetary policy transmission and the singleness of the monetary policy’ was a legitimate monetary policy objective. The Court here not only effectively amended the Treaties, which confine the ECB’s monetary policy to the pursuit of price stability and, subject to that overriding objective, to specific foreign exchange and payment systems operations, but it also took the ECB’s declared aims at their face value, when it held that the nature of a policy measure is to be assessed primarily by reference to its objectives and not its substance nor its effects. In Pringle, the Court had held that bond buys by the ESM ‘to preserve the stability of the euro’ were economic, not monetary policy. In Gauweiler, it decided that, when the ECB buys bonds for an allegedly different purpose, bond buys become monetary policy just because, and for no other reason that, the ECB buys the bonds and states that it does so for a different objective. The Court’s argumentation not only ignores the Treaty text, but contradicts its own previous decisions according to which the aim of EU acts is to be determined objectively, not subjectively.79
On the first issue, the CJEU’s conclusion that, depending on their objective, bond buys may be a legitimate monetary policy measure, does not only sit uneasily with the view reached in Pringle, that purchases of government debt securities represent an economic policy instrument aimed at ensuring the survival of the euro, It also ignores the fact that Article 123 TFEU does not only prohibit monetary financing of euro governments by the ECB by means of direct bond buys from the issuing public bodies,80 but that recital (7) of Council Regulation 3603/9381 extends the prohibition of monetary financing of government debt to ‘purchases made on the secondary market’, which may have the effect ‘to circumvent the objective of that Article’. Recital (8) of the Regulation lends additional support to a broad and very strict construction of the prohibition of monetary financing and states that neither the ECB nor the European System of Central Banks may engage in purchases ‘of marketable debt instruments’ issued by a Eurozone Member State which may, in any way, ‘help to shield the public sector from the discipline of market mechanisms’. In Gauweiler, the CJEU not only simply ignored Council Regulation 3603/93 in its entirely, it also conveniently ignored the fact that, in Pringle, it had accepted that limited bond buys by the ESM are an appropriate and legitimate economic policy instrument to allow Eurozone Member States to roll over and issue government debt instruments at a cost below the market rate.
Economic and monetary policy, however, are conventionally distinguished by reference, inter alia, to the instruments they respectively employ. Moreover, unlimited bond buys by the Central Bank have inevitably the same effect, by increasing demand and distorting the private investor market, as bond buys by another body, and that effect is inevitably greater if the ECB’s bond buys are in principle unlimited—in contrast to the ESM, which operates under the constraint of overall credit limits. Gauweiler was decided nearly five months after the ECB had launched a quantitative easing bond-buying programme and nearly three years after the announcement of the OMT. At the time of the Gauweiler judgment, the economic effects of the ECB’s bond-buying programme had become obvious. Both the announcement of the programmes and the commencement of the purchases had resulted in substantial declines in risk premiums and the interest on newly issued Eurozone bonds. This is precisely the type of monetary financing of public debt which Article 123 TFEU and Regulation 3603/93 prohibit. The CJEU, however, chose to uphold the ECB’s ‘unconventional’ monetary policy in defiance of the economic evidence, as readily as the ordinary meaning of recitals (7) and (8) of Council Regulation 3603/93.
The CJEU’s conclusion on the second principal issue, that is, of whether the improvement of the transmission mechanism for the ECB’s monetary policy is a legitimate monetary policy, likewise plainly disregards the Treaty. The EU Treaties and the Statute of the ECB are clear on this point. The overriding objective of monetary policy in the Eurozone is the maintenance of price stability,82 whilst the tasks of the ECB also include the carrying out of foreign exchange operations, the management of foreign currency reserves, and the maintenance of the payment systems within the Eurozone. The ECB’s declared aim, in both the OMT and its subsequent Quantitative Easing programme of improving the transmission mechanism for its monetary policy, has no treaty basis. Moreover, market economies are not planned economies. In market economies the central bank sets base interest rates and minimum reserve requirements for commercial banks and carries out short-term open market foreign exchange and securities operations, but the transmission of these central bank monetary ‘impulses’ is deliberately left to the market. Commercial banks then determine interest rates and lending volumes to the corporate sector and private households, and the interest rate and credit limits they set for loans to individuals incorporate a variable risk premium, which reflects the default risks of individual borrowers. The ECB’s attempt to interfere with the transmission mechanism directly conflicts with the principle of an ‘open market economy’ according to Article 120 TFEU.
Importantly, the CJEU’s conclusion that the improvement of the transmission mechanism represents a legitimate monetary policy objective not only defies the Treaties, the reasoning behind it also runs counter to the Court’s settled case law that the objective of an EU act is to be determined objectively, not subjectively.83 In Gauweiler, the Court held that the nature of a measure—that is, whether it is to be regarded as monetary or economic policy—is to be assessed principally with reference to the declared or putative aims stated the ECB, even if the measure ‘may have indirect effects’ which have nothing to do with monetary policy.84 Previously, the CJEU had consistently held that to determine the nature of an EU measure, primary regard must be had to its objective effects, and that the subjective intention or declared objective of the legislator or policy maker is not determinative, but only the starting point. In particular, it had always insisted that ‘the Court must verify whether the measure whose validity is at issue in fact pursues the objectives stated by the Community legislature’.85 In Gauweiler, the CJEU engaged in no such assessement. It simply conveniently ignored settled case law that, in determining the aim of a Union measure, its effects cannot be ignored.
And, as for the declared aim of the OMT programme, the Treaty states that the objective of the ECB is to secure price stability. The OMT, by contrast, is not even aimed at price stability, but the transmission of monetary policy, the so-called ‘singleness of monetary policy’ and, in all but name, the preservation of the Euro.86 Neither aim has any basis in the Treaty. The measure is turned into monetary policy just because the ECB says so. The ECB, according to the CJEU, does not even have to justify its measures by reference to Treaty provisions. Monetary policy, quite simply, is what the ECB says it is. It can thus do anything, provided the Member States accept it.
In Pringle, the Court purported to rely on, but in fact dismissed, the wording of Article 125 TFEU to justify mutual financial assistance by reference to an implied purpose which is not even in the Treaties, that is, the stability of the euro, thus discarding the more immediate purpose of preventing mutualization of debt by Member States. And in Gauweiler, the ECJ took the view that the nature of a measure may be determined by its subjective purpose and not objectively by due reference to its effects. Moreover, that subjective purpose may be accepted, even if it is not contained in the Treaty, thus effectively freeing the ECB from any judicial oversight.
Pringle and Gauweiler are two decisions which illustrate how far, in promoting the goal of further EU integration, the CJEU is prepared to stretch the methodological constraints on judicial decision making. In Pringle, the CJEU overrides the natural meaning of Article 125 TFEU as well as its wider legislative context, especially Article 21 ESM Treaty, and defies the teleo-systemic purpose of Article 125 TFEU in the treaty context, provided by Articles 120 to 127 TFEU. It opts instead for a reading based on a meta-teleological objective, that is, the principle of solidarity to preserve the euro, in defiance of the express legislative choice made by Article 125 TFEU, namely that a sound budgetary policy is to be achieved through national budgetary responsibility, as well as the basic economic axiom that mutualization of debt reduces, and does not enhance, incentives to pursue sound budgetary policies. In Gauweiler, the CJEU chooses to disregard implementing legislation, especially Council Regulation 3603/03, relevant to the interpretation of Articles 123 and 127 TFEU. Furthermore, the Court, in defiance of its own settled case law, effectively granted the ECB the right to create new monetary policy. This, in spite of the TFEU manifestly restricting the ECB to the pursuit of price stability and the management of foreign exchange operations, and the maintenance of the payments system within the Eurozone.
The foregoing discussion has shown that in the areas of the division of competences, EU citizenship, and fundamental rights and in the Eurozone litigation the CJEU has adopted a strongly integrationist and meta-teleological approach which goes well beyond the cumulative approach suggested in Merck v Hauptzollamt Hamburg-Jonas.87 The most extensive body of examples of the Court’s integrationist and meta-teleological bias is provided by its numerous decisions on the scope of and individual rights conferred by EU citizenship. In this area the Court has used the ‘equal treatment’ principle of the EU Treaties progressively to override more restrictive secondary legislation, which limits the entitlement of economically inactive EU migrants, by reference to the requirement that they must have sufficient resources to maintain themselves. The adoption of Directive 2004/38 explicitly spells out that requirement, which logically follows from the fact that the EU Treaties restrict the right to in-work benefits of current and former EU migrant workers, but not of economically inactive EU migrants. Once in Grzelczyk and D’Hoop the CJEU had relaxed that requirement, and thus, in effect, judicially amended the Treaties, the Court, in subsequent cases, could then justify further extensions of benefits entitlement by reference to its earlier case law, which in all but name replaced the more restrictive legislation with a more permissive judicial interpretation.
In Gauweiler, the Court disregarded settled case law, which ran counter to its integrationist objective. In its citizenship judgments, to reinforce a previous departure from the wording of the Treaties and the Directive, the decisions of the Court abound with references to previous decisions, in preference to an analysis of the precise wording of the underlying legislative provisions. The Court’s approach to its own previous decisions is thus highly selective. It will liberally rely on previous decisions when this lends an air of the familiar and legal objectivity to justify an integrationist decision, whilst it will readily ignore a previous decision if it runs counter to the Court’s pro-Union, integrationist objectives.
The Court’s meta-teleological or integrationist leanings tend to be far less evident and often irrelevant in cases which deal with specific issues in EU legislation in the fields of agriculture and fisheries, the environment, competition, or transport, which raise no fundamental issues involving the interests of the Union or the division of powers between the EU and the Member States, and where the various interpretative criteria rarely point to very different conclusions. In these circumstances the Court rarely does violence to the language of the legislation.88
However, where the outcome matters, in the sense that the case itself affects the interests of the Union or raises significant issues for the EU legal order as a whole, the Court of Justice’s cumulative approach gives the Court almost complete freedom of interpretation, because it maximizes methodological flexibility and uncertainty, as illustrated by decisions adopted in the area of EU Citizenship or the Eurozone. The Court of Justice regards itself free to consider literal, systemic, and teleological considerations, without any rule of priority or hierarchy between them and with no fixed weight to be given to each consideration. In ‘constitutional’, human rights, or competence cases, it will usually follow the ordinary meaning where it supports an integrationist answer to the legal question raised, but where the literal meaning does not, it will commonly prefer to rely on other arguments where those are more favourable to an expansive reading of the competences of the EU institutions. Nor does the Court regard itself as bound by its precedents. It readily cites its own case law in support of its integrationist conclusions, but disregards previous decisions when they run counter to judicial preference.
The cumulative approach leaves the Court free to give the greatest weight to whatever criterion supports its preferred conclusion, and if it chooses purpose, it may pick and choose whatever purposes best suits the Court’s intentions. The purposes the Court may then invoke are not necessarily confined to those written into the Treaty, nor to the most immediate objects evident from the legislative context—purpose, in the CJEU’s view, may be presumed as well as treaty-based, and the Court may refer to meta-teleological considerations just as readily as to either the immediate or indirect purposes, each of which may be either subjectively or objectively construed, depending on which most readily support the Court’s preferred solution. Moreover, by purpose the Court may also refer to effects, means, functional criteria, or general consequences. In general terms, whether explicitly or impliedly, when the Court invokes purposive interpretative criteria, it almost always falls back on what best suits EU integration, even where this contradicts specific Treaty provisions and thus conflicts with the lex specialis principle in international law.
Paradoxically, the decision making of the CJEU is not subject to unusually high legal uncertainty. The CJEU’s decisions are probably more certain than those of other courts. They are certain, however, not because its approach is governed by a high degree of methodological rigour, but because the Court’s pro-Union predisposition is so settled. Judicial predisposition, not method or logic, is the best guide to the Court of Justice’s legal argumentation. The CJEU may thus be regarded as a tribunal which combines a relatively high degree of legal certainty in its decisions with extreme methodological freedom in its judicial reasoning—a degree of methodological flexibility which, in its flexible approach to the use of the various interpretative criteria, previous judicial decisions, and the Court’s ready reliance on non-textual and meta-teleological considerations, must be placed at the extreme ‘teleological’ end of the interpretative options provided by the Vienna Convention on the Law of Treaties. If, as all good law should be, the Vienna Convention is construed as imposing genuine methodological constraints, then the CJEU, in so far as its flexible interpretative approach frees it from almost any methodological constraints, frequently operates outside the interpretative freedom afforded by Articles 31 and 32 VCLT. It upholds instead its own communautaire version of a ‘good faith’ reading of legal provisions, relying upon or disregarding at leisure settled canons of interpretation as well as the ordinary meaning of EU Treaty and legislation provisions.
A treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose.
The context for the purpose of the interpretation of a treaty shall comprise, in addition to the text, including its preamble and annexes:
any agreement relating to the treaty which was made between all the parties in connexion with the conclusion of the treaty;
any instrument which was made by one or more parties in connexion with the conclusion of the treaty and accepted by the other parties as an instrument related to the treaty.
There shall be taken into account, together with the context:
any subsequent agreement between the parties regarding the interpretation of the treaty or the application of its provisions;
any subsequent practice in the application of the treaty which establishes the agreement of the parties regarding its interpretation;
any relevant rules of international law applicable in the relations between the parties.
A special meaning shall be given to a term if it is established that the parties so intended.
leaves the meaning ambiguous or obscure; or
leads to a result which is manifestly absurd or unreasonable.
Overdraft facilities or any other type of credit facility with the European Central Bank or with the central banks of the Member States (hereinafter referred to as ‘national central banks’) in favour of Union institutions, bodies, offices or agencies, central governments, regional, local or other public authorities, other bodies governed by public law, or public undertakings of Member States shall be prohibited, as shall the purchase directly from them by the European Central Bank or national central banks of debt instruments.
Paragraph 1 shall not apply to publicly owned credit institutions which, in the context of the supply of reserves by central banks, shall be given the same treatment by national central banks and the European Central Bank as private credit institutions.