Federalism in the European Union 9781472566164, 9781849462426

Citation preview

Preface We are in the middle of a severe recession and a continuing crisis with the single currency. Switch on the television or the radio and you will find pundits pontificating—the more Euro-sceptic amongst them with an almost palpable, savage glee. Will Greece leave the Euro? Will other ‘vulnerable’ Member States be picked off, one by one, as the ratings agencies (those mysterious and apparently all powerful gods) downgrade the sovereign debt of one Member State after another? Is Spain (Portugal? Italy? Ireland? whoever is next: pick your chosen target) too big to fail but too big to save? Can another (better) rescue plan be cobbled together and succeed in satisfying the markets? Can the Germans really be expected to go on showing ‘solidarity’ and writing cheque after cheque—and, if Bundeskanzlerin Merkel is willing to contemplate trying, will the Bundesverfassungsgericht be prepared to wear it? Is the whole idealistic post-war ‘European project’ doomed? Against that background, a compilation of essays that tries to examine where we have got to within the European Union on federalism and decentralisation, and where we may be going, may seem to some like the ultimate exercise in the futile or indeed the surreal. If the end of the Euro-world is nigh, does it really matter whether one conceptualises the European Union as a crypto-federal structure or maintains that it is merely a particular form of international organisation (or whether the truth is somewhere between the two)? The short answer is ‘yes, it matters a very great deal’. If we look back to 1992 and the Treaty of Maastricht, it is abundantly clear that, whilst there was a lot of enthusiasm for the single currency as an overt symbol of Euro unity, there was no parallel political will (as the necessary concomitant of creating such a single currency) to place fiscal policy under common control.1 Whatever the external trappings, the practical sovereignty that mattered—such as our freedom to determine our own market philosophy, reflecting the political choices made by our electorate—would remain vested in the central governments of the Member States. And so it continued for the next twenty years. In the meantime, within individual Member States, regions that were experiencing difficulties could hope to be helped out by wealthier regions through transfer mechanisms operated by central government. The better off may grumble, of course; but within a single State where there are wage differentials between richer and poorer regions, there is always quite a lot of central control over what the poorer regions are spending and (crucially) there is also a sufficient sense of solidarity, of ‘we’re 1 Hindsight is of course a wonderful thing. However, there were a few unpopular economists who pointed this out at the time: see, for example, M Friedman, ‘Why the euro is wrong for Europe’ (1997) 14 New Perspectives Quarterly 23–24.

vi Preface all in this together’ (even if it is a bit grudging at times) to keep the national show on the road. We are in the painful process of finding out that establishing a single currency without fiscal union in 1992 was a step that was both too far and not far enough. But the single currency is no longer a new-fangled toy that can be unmade by waving a magic wand over the experiment. Twenty years on, the single market is much more interlaced and interdependent than it was when the Treaty of Maastricht was signed; and the overall European project is much, much more than the mere single market. At the time of writing, the tide is running quite strongly in favour of the idea that—now that we really have our backs against the financial wall—what is needed is ‘more Europe’, not less. A closer degree of coordination of fiscal policy and more action by central banking is seen as crucial to rescue the Euro. The inevitable corollary, without touching a single Member State boundary on a map, is the creation of more European unity (even if the dreaded f-word is kept discretely in the closet). Member States who are not part of the Euro-zone are nevertheless still so deeply involved in the economics of the European project that—in their own interests—they cannot realistically remain aloof and uninterested in the precise details of how those at the centre of the storm set about weathering it. Thus, the exigencies of finding a technical solution to the present ‘crisis in the Euro-zone’ leads ineluctably to much bigger and deeper questions about the sort of European Union that our present half-way house may need to evolve into. What are the appropriate bases for the unity and solidarity that is so obviously required? In national terms, unity and solidarity link naturally to shared constitutional values. Citizens own their nation state through the democratic process. They expect their courts (even if the thought is seldom clearly articulated) to ensure that their governments respect fundamental rights and constitutional values and that the executive does not encroach upon the proper sphere of the legislator. Where the nation state has a federal structure, the courts patrol the proper separation of powers between the Bund and the Länder (or the central authority and the regions) and the citizen identifies both with his region and with his nation. The rule of law is woven deeply into the fabric of the civilised, democratic society of which each citizen is a part; and the jurisprudence of the constitutional court (or the supreme court, where a separate constitutional court does not exist) plays an important part in consolidating that society. Citizens do not usually, however, love their nation in the abstract. They grew up with it and identify with it. They know roughly where they stand and what they can expect and are suitably cynical about both, without ever seriously thinking of ‘leaving’ their nation and trying out a different one. They accept that they have duties as well as rights. They value what they identify as the good things about their nation, be that a tranquil life and good quality environment or sporting prowess and cool culture. They make wry faces about what is less good. But they do not (usually) regard those less desirable qualities as reasons for disowning the nation itself.

Preface vii Viewed in that light, the ‘European project’ (in the guise first of the ECSC, then the EEC and the EC, and now the European Union (leaving EURATOM to one side for a moment)) has been immensely successful in offering to those living within its remit decade after decade of undisturbed peace and relative prosperity—and the gift has been ‘banked’ invisibly. More publicity is given to the (mythical) directive on straight bananas than to twenty measures that improve consumer protection or facilitate free movement. The nationals of the Member States have become citizens of the Union almost by sleight of hand; and certainly without much open recognition of what that new citizenship might entail. Against that background, is this book not all about angels dancing on the head of a pin? Given the state the European Union is in, should one’s reaction not be that, ‘frankly we’ve rather more urgent problems at hand’? Quite the opposite. By addressing the two distinct but interwoven questions of whether the European Union itself is a federal system, and whether the European Union allows for federalization within the Member States, this book arguably points to the possible way out of both the EU’s crisis of legitimacy and its economic crisis. Bear with me in both respects. On the one hand, resolving the economic and budgetary crisis would (as scores of economists never cease to remind us) require setting straight the ‘design flaw’ committed at Maastricht and establishing coordinated regulation of banking and financial markets and strengthened controls over Member States’ budgets. The crisis of legitimacy, on the other hand, would require such reforms to be accompanied by a clear reinforcement of the democratic character of decision-making at the European Union level. In other words, both operations would tend to entail a further step in the construction of a ‘federal’ European Union. However, that endeavour will be entirely impossible unless the Union makes it perfectly clear that it continues to respect the constitutional autonomy and identity of its constituent Member States, including their federal and otherwise devolved constitutional structures, which in turn guarantee that decisions are taken as closely as possible to the citizens of the European Union. Federalism, then, could be the answer to the European Union’s quandaries. It is, of course, not an EU invention. As the comparative chapters in this collection of essays show, one of the more venerable examples can be found across the Atlantic. Naturally, the United States too continues to face fraught questions. According to the Supreme Court of the United States, which was recently confronted with the problem of how to ensure affordable health care for individual citizens without infringing on their personal liberty, ‘federalism secures to citizens the liberties that derive from the diffusion of sovereign power’, and by ‘denying any one government complete jurisdiction over all the concerns of public life, federalism protects the liberty of the individual from arbitrary power’.2 The Supreme Court also recalled

2

National Federation of Independent Business v Sebelius, 567 U.S. ___ (2012) (slip op, at 4).

viii

Preface

that the federal system rests on ‘what might at first seem a counterintuitive insight, that freedom is enhanced by the creation of two governments, not one’.3 For the European Union, Article 4 TEU would appear to provide an excellent programme for such a conception of federalism. It provides for the typical federal rule (compare, for example, the tenth amendment to the US Constitution) that competences not conferred upon the Union in the Treaties remain with the Member States.4 The Union is also under an obligation to respect the equality of Member States before the Treaties ‘as well as their national identities, inherent in their fundamental structures, political and constitutional, inclusive of regional and local self-government’. More specifically, the Union is to respect ‘their essential State functions, including ensuring the territorial integrity of the State, maintaining law and order and safeguarding national security. In particular, national security remains the sole responsibility of each Member State’.5 The Court of Justice has made it clear that it takes seriously respecting the national identities of the Member States.6 However, Article 4 TEU also provides the ‘glue’ to keep the federal construction together, in the form of the principle of sincere cooperation. Pursuant to that principle, the Union and the Member States are to assist each other, in full mutual respect, in carrying out tasks which flow from the Treaties. In particular, the Member States are to take any appropriate measure, general or particular, to ensure fulfilment of the obligations arising out of the Treaties or resulting from the acts of the institutions of the Union. The Member States are likewise required to facilitate the achievement of the Union’s tasks and refrain from any measure which could jeopardise the attainment of the Union’s objectives.7 Conceptually, the European Union’s particular brand of federalism is therefore an instrument both for recognising and accommodating national diversity and for transcending it, when and where required. The book’s two-part structure reflects ‘federalism’s Janus face’. It offers a wide-ranging discussion of federalism in the European Union, addressing issues that are pertinent not only to suggesting a way out of the current impasse but also to getting the citizenry of Europe engaged with, committed to and, perhaps, even fond of, the EU. It could not have come at a more appropriate time. Eleanor Sharpston Advocate General at the Court of Justice of the EU

3 4 5 6 7

Ibid 47. Art 4(1) TEU. Art 4(2) TEU. Case C-208/09 Sayn-Wittgenstein (ECJ, 22 December 2010), paras 92–94. Art 4(3) TEU.

Acknowledgements This collection of essays is based on the proceedings of the workshop Federalism in the European Union, which was held at the Law Faculty of the Katholieke Universiteit Leuven on 12 and 13 May 2011. We should like to thank all of the participants for their insightful and inspiring comments and questions on the papers presented at the workshop. A special word of thanks is due to Jan Wouters, for opening the event, and to the panel chairs Steven Blockmans, Koen Lemmens, Geert Van Calster and Walter Van Gerven. We are also indebted to the generous financial and organisational support given by the Leuven Institutes for Constitutional Law and European Law, the Centre for Law and Cosmopolitan Values, and the Leuven Centre for Global Governance Studies. We wish to thank, in addition, Viviane Verbist and Marina Smeets for their much appreciated practical assistance during the conference. We are deeply grateful to everyone who made this edited collection possible. Special thanks go to Richard Hart, Rachel Turner, Mel Hamill, Tom Adams, and the whole team at Hart Publishing for their unwavering professionalism and their early and continued support of our project. We should also like to thank the authors of the individual chapters for their hard and excellent work, and for the outstanding commitment they have shown. Thomas Verellen deserves particular mention for his dedicated help during the editing process. Last, but not least, we should like to express our gratitude and appreciation to Georgios Pavlakos and the members of his Centre for Law and Cosmopolitan Values. They are a constant source of enthusiasm, encouragement and intellectual stimulation to us.

List of Contributors Joxerramon Bengoetxea is Professor of Jurisprudence and Sociology of Law at the University of the Basque Country. Monica Claes is Professor of European and Comparative Constitutional Law at Maastricht University. Elke Cloots is a PhD fellow of the Research Foundation—Flanders (FWO) at the Institute for Constitutional Law, Katholieke Universiteit Leuven. Geert De Baere is Assistant Professor of International Law and EU Law at the Katholieke Universiteit Leuven. He is also a senior member of the Leuven Centre for Global Governance Studies. Helder De Schutter is Assistant Professor of Social and Political Philosophy at the Institute of Philosophy, Katholieke Universiteit Leuven. Maartje de Visser is Assistant Professor at Maastricht University. Pavlos Eleftheriadis is a Fellow of Mansfield College and University Lecturer at the Faculty of Law, University of Oxford. Stef Feyen is a PhD fellow of the Research Foundation—Flanders (FWO) at the Institute for Constitutional Law, Katholieke Universiteit Leuven. Kathleen Gutman is Assistant Professor at the University of Antwerp and a post-doctoral research associate at the Institute for European Law, Katholieke Universiteit Leuven. Alexia Herwig is a post-doctoral research fellow of the Research Foundation— Flanders (FWO) at the Centre for Law and Cosmopolitan Values, University of Antwerp. Suzanne Kingston is Lecturer in Law at University College Dublin. Koen Lenaerts is Judge and President of Chamber at the European Court of Justice, and Professor of EU Law at the Katholieke Universiteit Leuven. Giuseppe Martinico is García Pelayo Fellow at the Centro de Estudios Politicos y Constitucionales, Madrid. He is also a researcher at the Centre for Studies on Federalism, Turin, and Adjunct Professor at the Scuola Superiore Sant’Anna, Pisa. Nikos Skoutaris is Assistant Professor at Maastricht University. Stefan Sottiaux is Professor of Constitutional Law and Discrimination Law at the Katholieke Universiteit Leuven.

xxiv List of Contributors Aida Torres Pérez is Constitutional Law Professor at Pompeu Fabra University, Barcelona. Piet Van Nuffel is Professor of EU Law at the Katholieke Universiteit Leuven. He is also a member of the Legal Service of the European Commission. Herwig Verschueren is Professor of International and European Social Law at the University of Antwerp and Visiting Professor at the University of Brussels (VUB). Maite Zelaia Garagarza is Professor of Constitutional Law at the University of the Basque Country.

Introduction: Federalism’s Janus Face ELKE CLOOTS, GEERT DE BAERE AND STEFAN SOTTIAUX

N

‘

ATIONALISM IS AN ideology of the past. Our goal is a federal and post-national Europe, a Europe of the citizens.’ By using the word ‘federal’ in its mission statement, the Spinelli Group, a recently-established network of European ‘Federalists’,1 quite obviously seeks to convey a message of more, not less, European integration. The Group’s avowed aim is to create a federal Europe, where all Member State nationals are equal as Union citizens, where national differences are transcended, and where important decisions in the fields of economics, social security and fiscal policy are adopted at the EU level.2 The concept of ‘federalism’, however, generally has a different—indeed the opposite—resonance in the domestic political realm. In formerly unitary States like Italy, Spain, Belgium and the United Kingdom (UK), the process of federalisation has implied granting political autonomy to sub-State groups, in order for minorities at the State-wide level to become self-governing majorities within their own territorial subdivision.3 Federalisation is thus seen as an instrument for recognising and accommodating, rather than transcending, national diversity. As a result of such domestic federalisation processes, the legal rights and duties of the nationals of the State in question are no longer identical but vary according to, for example, one’s region of residence. Significantly, with the exception of Belgium, the Member States examined in the present volume do not officially classify themselves as federal systems. They prefer, instead, to describe their governance structures in terms that they consider to be less evocative of political

1 The statement is extracted from the Spinelli Group’s manifesto, available at . 2 See especially Conclusions of the Shadow European Council of the Spinelli Group, 22 March 2011, available at ; Conclusions of the European Shadow Council of the Spinelli Group, October 2011, available at . 3 See more generally, J McGarry, B O’Leary and R Simeon, ‘Integration or Accommodation? The Enduring Debate in Conflict Regulation’ in S Choudhry (ed), Constitutional Design for Divided Societies (New York, Oxford University Press, 2008) 63–69.

2 Elke Cloots, Geert De Baere and Stefan Sottiaux disintegration, such as ‘devolution’ (UK), ‘Estado de las Autonomías’ (Spain) and ‘regionalismo’ (Italy). The epithet ‘federal’ thus seems to denote two opposite concepts, depending on the polity to which it is supposed to apply. When used in relation to the EU, federalisation appears to signify more unity, uniformity and (formal) equality throughout the Union, at the expense of the autonomous powers of its territorial subdivisions (ie, the Member States). In the context of multinational Member States, by contrast, federalism refers to increased self-governing powers for subState entities (eg, regions, communities) and, accordingly, to more legal diversity and less (formal) equality. What is more, ‘federal’ is a label most politicians, at EU and domestic levels alike, seem to shun, because it suggests a degree of political integration or disintegration, respectively, which they consider to be undesirable. What does the Janus-faced character of federalism in Europe teach us about the nature of the concept? Can both sides of federalism coexist? And how could ‘federalisation’ feasibly be an appropriate term for describing what has been going on in post-war Europe if it is rejected by so many political stakeholders? It is clear that ‘federalism in the European Union’ remains a matter of great controversy. This volume engages with the two major current debates on the phenomenon of federalism in the context of the EU. The first part of the book addresses the question whether the EU can be defined as a federal system, and whether it can learn from existing federations. In the second part, the attention shifts to federalisation processes within EU Member States, more particularly to the impact of these processes on EU law and vice versa.

PART I: FEDERALISM IN THE EU’S CONSTITUTIONAL STRUCTURE

Can the EU be described as a federal system? That question, which has variously aroused both antagonism and enthusiasm, is the focus of the first part of the present volume. It opens with a chapter by Koen Lenaerts, who understands federalism as ‘the balance of power between the search for “unity” and the protection of “diversity”’. Lenaerts explores that balance of power from three different angles. First, he notes that ‘diversity’ may feature not only among the Member States of the Union, but also within a single Member State. His study is therefore not limited to the division of powers between the Union and the Member States; it also covers the impact of EU law on domestic constitutional arrangements accommodating internal diversity. Secondly, Lenaerts examines the balance of power between unity and diversity as reflected in the Union’s political decisionmaking process. He argues that the question of the appropriate legal basis for an EU measure pertains not only to the balance of power between the distinct EU institutions, but also to the balance of power between the EU and the Member States. Thirdly and lastly, Lenaerts studies to what extent the balance of power between unity and diversity has been affected by the Treaty provisions on Union citizenship. He submits that the recent ECJ case law on the matter has sought to

Introduction: Federalism’s Janus Face 3 reconcile the protection of individual rights granted by EU law with the powers of the EU and national legislatures. A rather different conception of federalism is adopted by Pavlos Eleftheriadis. He argues that the federal model requires a single scheme of jurisdiction based on a coherent set of constitutional principles. Eleftheriadis’s account of federalism is premised on the thesis that a single scheme of jurisdiction is needed for institutions to be legitimate. Drawing on the work of Plato, Aristotle and Jeremy Waldron, Eleftheriadis submits that ‘[a] set of political institutions may be legitimate for a given group of people in a given territory at a given time only if it is the only one in force’. The plural character of the EU’s scheme of jurisdiction thus leads to the conclusion that federalism cannot be an appropriate constitutional idea for the EU. In Eleftheriadis’s view, the international model could offer a more appropriate framework for understanding the EU. Moving away from the conceptual debate about federalism and focusing directly on the allocation of competences over various tiers of government, Alexia Herwig’s chapter examines the role of the principle of subsidiarity in multilevel governance contexts. For Herwig, the subsidiarity principle essentially demands that a decision be made at the level of governance which is best placed to decide on the matter. While recognising the challenges that may come with the implementation of the principle in positive law, Herwig fully endorses the principle shown ‘in its best possible light’. The nub of her argument is that, for the law to be legitimate, the subsidiarity principle must be respected. More precisely, the law is legitimate only if it can be demonstrated that (collective) regulation is preferable over individual decision-making. From this general proposition, Herwig derives that the subsidiarity principle should inform the making of WTO law and EU law as well. Part I subsequently introduces three comparative reflections into the debate. Chapters four to six focus on specific issues with which multi-layered governance systems, including the EU, must grapple: how to resolve disputes regarding the allocation of powers, how to organise the protection of fundamental rights and how to conduct international relations. The respective contributions explore to what extent the solutions prevailing in certain federal states could serve as a model for the EU. In search of the common features shared by all constitutional courts in federaltype Member States, Monica Claes and Maartje de Visser undertake a comparative study of the role, organisation and operation of these courts. In a next step, the authors examine to what extent the commonalities found are equally exhibited by the ECJ. Claes and de Visser conclude that the ECJ resembles federal constitutional courts in many respects. For instance, one of the most important tasks of the ECJ and federal constitutional courts alike is to adjudicate disputes over the division of powers. Yet Claes and de Visser’s research also reveals that significant differences remain. The authority of the ECJ over issues of power allocation and fundamental rights protection, for example, is much more contested than that of the domestic constitutional courts examined.

4 Elke Cloots, Geert De Baere and Stefan Sottiaux The subsequent chapter, by Aida Torres Pérez, reflects on whether the EU can learn from the US system of power division as regards the protection of fundamental rights. Torres Pérez centres her research around two particular issues that multi-layered polities must confront in the field of rights protection: (i) to what extent are the polity’s territorial subdivisions bound by central rights; and (ii) how should conflicting interpretations of rights (ie, by the central and sub-polity level) be addressed? As far as the first question is concerned, Torres Pérez observes that in the EU, unlike in the US, the central (ie, EU/US) standard of rights protection is not in all circumstances binding on the States. Yet she also notes that the scope of application of EU rights to the Member States has expanded radically over the years. As regards the second question, Torres Pérez reminds us that, even in the US, there is still room for differing rights interpretations among the states, as the federal Constitution merely provides a floor, not a ceiling, of rights protection against state action. Nonetheless, the author does not recommend the blanket introduction of this floor-of-protection rule into the EU system of rights protection, given the challenges such a rule may pose for the efficacy and uniform application of EU law. In her view, the better approach would be to enhance ‘dialogue’ between the ECJ and the Member State courts with regard to the interpretation of rights. In the last chapter in Part I, Geert De Baere and Kathleen Gutman explore the approaches taken in the EU and the US for forging a coherent international relations policy for an internally differentiated polity. The primary focus is on analysing the constitutional framework governing international relations in the EU and the US, respectively, evaluating the allocation of competences between the central authority and the component entities in light of the respective ‘constitutional’ texts (the Treaties and the US Constitution) and relevant case law of the Court of Justice of the EU and the US Supreme Court. The chapter finds that although the EU and the US have different starting points when it comes to the vertical division of competences in international relations, both evidence a subtle shift or accommodation in the opposite direction. In both legal orders, the point of departure is therefore no longer the exclusive competence of either the central level or of the level of the component entities, but rather, how to reconcile the involvement of both levels in the international relations sphere.

PART II: EU LAW AND MEMBER STATE FEDERALISM

In the second part of the volume, the focus turns to the relationship between EU law and the devolution processes that have recently occurred in certain Member States. It may be argued that the EU and the Member States that have granted political autonomy to sub-State national groups essentially share the same quest. The constitutional order of each of these political communities reflects a search for an equilibrium between the unity of the polity as a whole and the autonomy of the polity’s territorial subdivisions. In that sense, the political entities in question

Introduction: Federalism’s Janus Face 5 may at least be said to pursue the aim of federalism, that is, to guarantee unity as well as diversity in a given polity.4 The pursuit of the purpose of federalism evidently takes a different form in the EU than in multinational Member States. Whereas the search for the optimal balance between political unity and autonomy has a centrifugal dynamic at the Member State level, it is an integrating force in the Union as a whole. The question as to how both tendencies can coexist is the common theme running through the contributions in Part II of the book. However, another question must be addressed first: should those two ‘federal’ enterprises be allowed to coexist? Why would both European integration and the devolution of powers to sub-State entities be worth protecting and enhancing? A first motive for maintaining both federal-type structures in the EU could be found in the EU Treaty. Article 4(2) TEU, in its post-Lisbon form, proclaims that the Union is to respect the Member States’ ‘national identities, inherent in their fundamental structures, political and constitutional, inclusive of regional and local self-government.’ Hence, EU law itself could be said to require that Member State federalism not be undermined by the European project.5 This textual argument may be strengthened by a more instrumental reason: federal-type arrangements in the EU may have beneficial effects. For instance, one may wonder whether the peace we currently enjoy in Europe is not, to a great extent, the result of the two ‘federalisation’ processes: European integration as well as internal devolution of powers. The peaceful coexistence of the national societies living in Europe seems inextricably linked with the continuous search for the appropriate equilibrium between political unity and autonomy, both at Member State and at EU level.6 Yet there are also deontological arguments that could be advanced in favour of federalism’s Janus-faced appearance in Europe. One is adduced by Helder De Schutter, whose thesis is predominantly justice-based. De Schutter puts forward a pluralistic understanding of national culture in the domestic as well as the supranational domain, or in other words, ‘national pluralism’. National pluralism simultaneously embraces the deepening of European unity and the political recognition of multiple national identities, whether they exist at the State or at the 4 See, eg AED Howard, ‘Does Federalism Secure or Undermine Rights?’ in E Katz and GA Tarr (eds), Federalism and Rights (Lanham, Rowman & Littlefield Publishers, 1996) 18. See also WS Livingston, ‘A Note on the Nature of Federalism’ (1952) 67 Political Science Quarterly 81, 88. 5 In the same way as Art 4(2) TEU requires the Union to respect a Member State’s status as a republic, which forms, according to the ECJ, part of a State’s national identity. See Case C-208/09 Sayn-Wittgenstein (ECJ, 22 December 2010), para 92. 6 A very forceful illustration of the link between federalism and peace in multinational States may be found in the Constitution of the Federation of Bosnia and Herzegovina. This Constitution was established as an annex to the Dayton Peace Agreement (1995), which put an end to the war in Bosnia and Herzegovina. It embodies complex power-sharing arrangements between the distinct ethnic groups. See in this regard also Sejdic' and Finci v Bosnia and Herzegovina Apps no 27996/06 and no 34836/06 (ECHR, 22 December 2009), para 45; Mijovic' J and Hajiyev J partly concurring and partly dissenting; Bonello J dissenting.

6 Elke Cloots, Geert De Baere and Stefan Sottiaux sub-State level. Federalism may offer an adequate structure for implementing the national pluralist programme. In sum, there may well be good reasons for preserving federalism’s Janus-faced character in the EU. The relationship between both federal phenomena, however, is not without problems. Various tensions may arise from their coexistence. If both are to be preserved in the EU, decision-makers would perhaps do well to resolve such tensions by ensuring that both ‘federalisation’ processes do not undermine one another. Translated into more specific terms, this means that European integration should not prevent the Member States from devolving powers to sub-State entities and, conversely, that devolution should not undercut the European integration process. The remaining contributions to the volume demonstrate that the institutions of both the EU and the federal Member States have recently taken significant steps in that direction. Whereas the Union has traditionally been portrayed as ‘blind’ to domestic federal structures, the second part of the book reveals that the Union institutions are willing to lift their blindfold in certain circumstances. Although this move has generally been welcomed by the contributors, a number of authors would like to see the Union institutions go further down this route. They put forward several proposals to make the Union legislature as well as the Union Courts more sensitive to Member State federalism than they (already) are. In addition, Part II reveals that not only the EU itself but also the federal Member States have made efforts to accommodate the other side’s federal aim. As the chapters concerning national legal systems show, the (constitutional) legislatures and the constitutional courts of the Member States in question have—each in their own way—sought to reconcile their States’ structural arrangements with the requirements of EU law. Broadly speaking, the tensions generated by federalism’s dual appearance come to the fore in two domains: the making of EU legislation, on the one hand, and (Union and domestic) court decisions, on the other hand. These two areas of tension form the backbone of Part II. A first series of essays assesses the role of autonomous sub-State entities in the making of Union legislation in the postLisbon era. The changes the Lisbon Treaty has brought about in this respect are discussed by Piet Van Nuffel. While acknowledging that the representation of regional interests in the Council, the European Parliament and the Committee of the Regions has not been significantly improved, Van Nuffel stresses the more prominent role domestic parliaments have achieved under the Lisbon Treaty. Most notably, the Treaty entitles national parliaments, first, to be informed of draft EU legislation and, secondly, to investigate whether those drafts respect the principle of subsidiarity. Although the Treaty itself does not extend these rights to regional legislative bodies, Member States may decide that the information must be passed on to the regions, and that the latter must be involved in the subsidiarity scrutiny process. Van Nuffel’s contribution clearly shows that, even after the Lisbon Treaty’s entry into force, the influence of autonomous sub-State entities on the making

Introduction: Federalism’s Janus Face 7 of EU legislation remains highly dependent on the goodwill of the Member States. A comprehensive account of these entities’ actual impact can, therefore, be given only after consideration of the Member States’ legal systems. That study is meticulously undertaken by Nikos Skoutaris. His comparative research covers both nominally federal Member States (ie, Germany, Austria and Belgium) and what he refers to as ‘regionalised’ States (ie, Italy, Spain and the UK). Skoutaris’s chapter reveals a significant similarity between the States examined as to the participatory rights of their respective autonomous regions. For one thing, the central authorities are generally bound to inform their regional counterparts of European draft legislation. Furthermore, in all States considered, autonomous sub-State entities may be able to influence the position of the central parliament or government as regards EU law through their representation in the federal upper chamber and/or in special interregional and national-regional bodies. Lastly, Skoutaris points out that each Member State allows, under certain conditions, for regional officials to represent it in the Council. The main point of contrast relates to the participation of regions in the domestic subsidiarity scrutiny of European draft legislation. The existing approaches range from mere informal channels of regional influence, to the joint representation of all autonomous regions by the upper chamber, and even to a system where each regional legislative body is permitted to issue a reasoned opinion on the matter. Taking a more critical stance, Joxerramon Bengoetxea argues that the interests of a certain type of region are still insufficiently protected in the EU law-making process. His particular concern is with ‘autonomous constitutional regions’, defined as regions that (i) are constitutionally recognised in their Member State, (ii) possess an extensive degree of political autonomy, and (iii) have their own legislatures and executives. Bengoetxea recommends that these regions’ special status be recognised under EU law. Such recognition would entail profound institutional change. In this regard, Bengoetxea offers a variety of suggestions, including a greater weight for the opinions of autonomous constitutional regions in the Committee of the Regions, a split of their vote in the Council for those Member States containing such regions, and an extension to these regions of the privileged status enjoyed by the Member States in proceedings before the ECJ. Not only the EU decision-making process, but also the case law of the Union Courts in relation to Member State federalism has lately undergone major evolutions. In the present volume, particular attention is paid to the ECJ’s position on fiscal, social and linguistic policies prevailing in federal Member States. It is especially in those areas that tensions between EU law and domestic federal arrangements are likely to arise (and actually have arisen). This should hardly come as a surprise. Both social solidarity and linguistic interests are strongly intertwined with people’s national identity. Various multinational Member States have accommodated this reality by granting political autonomy to sub-State nations in these fields. The autonomous sub-State groups often employ their newly-won powers to establish financial solidarity mechanisms among their own members, and to encourage the use of the regional majority language in their own institutions and

8 Elke Cloots, Geert De Baere and Stefan Sottiaux in the region’s public life more generally. Union law, however, attempts, in many ways, to sever or overcome the link between national identity and people’s fiscal, social and linguistic rights and duties. As a consequence, EU law could severely constrain the discretion of autonomous sub-State entities to exercise their powers in the way they had envisaged. How the Court does and should adjudicate conflicts between EU law and regional fiscal, social and linguistic policies is therefore a question of eminent importance. It is addressed in chapters eleven to fourteen. Suzanne Kingston’s contribution analyses the new interpretation the ECJ has given to the EU norms on State aid in order to accommodate the devolution of taxation powers to regions. Kingston embraces the Court’s case law on the matter as a ‘mature’ approach to Member State federalism. She lauds the Court’s ruling that fiscal benefits granted by genuinely autonomous regions should not, by definition, be considered instances of ‘selective’ aid. In so deciding, the ECJ ensured that the prohibition of State aid would not preclude devolution of tax powers. Herwig Verschueren, however, is more sceptical of the ECJ’s interpretation of the fundamental freedoms in the face of regional social policies. He disagrees with the Court’s decision that, under certain conditions, the Treaty provisions on the freedom of movement also apply to situations involving the crossing of a border between two sub-State entities that have self-governing powers in the area of social policy. In Verschueren’s view, such intra-State cross-border situations should always be governed by domestic constitutional law, even if the person who crosses the interregional border is an ‘EU migrant’. The relationship between internal market law and language regulation in multinational federal Member States is explored in a joint contribution by Stefan Sottiaux and Elke Cloots. The authors propose a framework for adjudicating disputes in this field. They advance the thesis that the designation of a Member State’s official languages should be excluded from the scope of the EU norms on freedom of movement and nationality discrimination. Yet in Cloots and Sottiaux’s view, this mode of adjudication should be extended neither to issues arising in the periphery of the grant of official status to languages, nor to laws regulating the language to be used in communication between private persons, nor to language proficiency conditions attached to welfare benefits. In so far as they hinder cross-border mobility and/or treat persons differently on grounds of nationality, such language requirements are contrary to EU law, unless justified in light of communication concerns or a group’s fundamental interests in relation to its language. Furthermore, Cloots and Sottiaux argue that linguistic requirements imposed by private persons too should be subject to the EU norms on freedom of movement and nationality discrimination. Yet, contrary to public authorities, private persons cannot successfully invoke the fundamental linguistic interests of the language group to which they belong as a ground of justification. A final contribution on ECJ adjudication relating to Member State federalism is by Elke Cloots. She contrasts the Court’s case law on this matter with its approach to domestic constitutional rights. Cloots observes that, whereas the ECJ balances fundamental freedoms against constitutional rights, it resorts to a

Introduction: Federalism’s Janus Face 9 radically different mode of adjudication in cases involving a conflict between EU law and Member State federalism. Drawing from US constitutional theory, she describes this alternative decision-making method as ‘categorical’ and ‘rule-based’. Although Cloots is convinced that a categorical and rule-based style is the most promising avenue in this type of case, she highlights certain pitfalls the Court may encounter when opting for rules and categories rather than standards and balancing acts. Obviously, not all disputes between EU law and domestic federal-type arrangements reached the Union Courts. The great majority of them were fought out at Member State level, most notably in the domestic constitutional courts. Hence, the final piece needed to complete the puzzle of European federalism’s Janus face is an analysis of the relevant case law of the constitutional courts (if any) of ‘federal’ multinational Member States. The three ultimate chapters of the volume are therefore devoted to the constitutional courts of Italy, Spain and Belgium, respectively. More particularly, they examine to what extent these courts have allowed EU law to affect their States’ structural arrangements. To put the question the other way around, to what extent have the constitutional courts revised those arrangements to ensure the States’ compliance with their duties under EU law? As far as the Italian constitutional court is concerned, Giuseppe Martinico notes that the court has generally approved of EU law being used by the central level of government as a tool for re-centralising powers which had been devolved to the regions. More particularly, the court has permitted the central State to intrude in various ways upon the regions’ powers, in order to avoid the risk of State liability for non-compliance with EU law. The State legislature is allowed, for instance, to implement EU law in areas of shared competence not only through basic principles (as it normally should), but also through detailed provisions interfering with the regions’ powers. The latter provisions remain in force as long as they are not replaced by regional laws. Maite Zelaia Garagarza’s contribution shows that the Spanish constitutional court too has served as an arbiter of disputes between the central State and the territorial subdivisions regarding EU law. Similar to in Italy, the relevant cases before the constitutional court mainly revolved around the rights of autonomous sub-State entities to participate in the making and implementation of EU law. Yet the Spanish territorial subdivisions seem to have been more successful before the court than their Italian counterparts. Most notably, the Spanish court has sanctioned several attempts by Autonomous Communities to compensate their loss of autonomy due to European integration. The court approved, for example, of the establishment of a Basque Office in Brussels. It also cleared the way for Statutes of Autonomy claiming regional power to implement EU law and stipulating that the internal distribution of powers as established by Spanish constitutional law cannot be modified by EU law. Unlike what seems to be the case in Italy and Spain, the Belgian judgments that have recently sparked a fierce debate deal with power struggles between the sub-State entities themselves rather than between the federal and sub-State levels

10 Elke Cloots, Geert De Baere and Stefan Sottiaux of government. As Stef Feyen’s contribution makes clear, a major issue in Belgian constitutional law is whether the scope ratione personae of social and fiscal benefits conferred by sub-State entities (ie, in practice, the Flemish entity) should be defined along the same lines in inter-State and internal situations, that is, the lines set out by EU free movement law. The question, in other words, is whether the Belgian internal free movement standards should be interpreted in the same way as the corresponding EU norms governing inter-State situations. If this question were to be answered in the affirmative, the practical effect would be that the (relatively generous) Flemish benefits could in principle no longer be refused on grounds that are ‘suspicious’ under EU law, such as residence, not even as regards purely internal situations. Feyen demonstrates, however, that the constitutional court’s recent case law is not entirely clear on the matter, and recommends a more coherent approach. The three chapters on constitutional adjudication invariably confirm that EU law has influenced the governmental structure of the Member States under examination, often with the complicity or help—depending on the viewpoint—of the respective constitutional courts. Significantly, the paradigm case discussed in these contributions did not involve a Member State—as a whole—defending its constitutional order against intrusion from EU law. Rather, what was essentially at issue was a dispute between distinct domestic political entities, whereby at least one party resorted to EU law in an attempt to evade constitutional norms concerning the State’s ‘federal’ structure, or to have those norms modified by the court. In sum, in all three legal systems, EU law has been invoked—rightly or wrongly—as a weapon in internal power struggles. The extent to which constitutional courts have been receptive to such arguments, however, seems to vary across States, and even across time within the same State.

1 EU Federalism in 3-D KOEN LENAERTS*

I. INTRODUCTION

A

T THE OUTSET of the European integration project, ‘federalism’ was understood as a synonym of ‘centralisation’ which is often stigmatised as a threat to the nation-State. However, as Koopmans noted almost 20 years ago, that traditional understanding of ‘federalism’ was simply unable to explain the European integration project, as it clung excessively to the notion of ‘the State’.1 A discourse based on the search for the place where the sovereign power resided (the centre versus the periphery) gave rise to ‘the wrong debate’. Alternatively, he advocated a broad reading of the notion of ‘federalism’ according to which the latter provides a dynamic conceptual framework explaining how power is allocated between the different levels of governance. In the processes of integration, Koopmans argued, ‘federalism’ should thus be linked to the notion of ‘legal pluralism’.2 If federalism encompasses elements of ‘legal pluralism’, the question is then how one must order a plurality of sources of law. As Delmas-Marty observed, ‘ordering [legal] pluralism’ amounts to devising a method of analysis capable of reconciling ‘dispersion and fragmentation’ with a ‘unified structure’.3 For Pescatore, in terms of political and legal philosophy, federalism is grounded in two basic principles, namely, ‘the search for unity, combined with a genuine respect of the autonomy and legitimate interests of the participant entities’.4 I concur with those two authors: federalism should be understood as a method of analysis that looks at the balance of power between the search for ‘unity’ and the protection of ‘diversity’.5 * Judge and President of Chamber at the European Court of Justice, and Professor of European Union Law, Katholieke Universiteit Leuven. All opinions expressed herein are personal to the author. 1 T Koopmans, ‘Guest Editorial—Federalism: The Wrong Debate’ (1992) 29 CML Rev 1047. 2 Ibid 1051 ff. 3 M Delmas-Marty, ‘Ordering Pluralism’, EUI Max Weber Lecture no 2009/06, available at . 4 P Pescatore, ‘Preface’ in T Sandalow and E Stein (eds), Courts and Free Markets (Oxford, Clarendon Press, 1982) 3–4. 5 K Lenaerts, ‘Constitutionalism and the Many Faces of Federalism’ (1990) 38 American Journal of Comparative Law 205.

14 Koen Lenaerts The purpose of the present contribution is thus to explore that balance of power by looking at three different federal questions. Both the subject-matter of and the problem introduced by those three questions are very diverse. This is done on purpose as I attempt to show that debates on federalism should not be confined to examining whether the European Court of Justice (ECJ) abides by the principle of conferral6 and that federalism has ‘many faces’. Since each question addresses federal issues from a different dimension (vertical, horizontal and diagonal), I attempt to examine federalism from a global perspective. The present contribution is divided into three sections. First, ‘diversity’ may take place not only among the Member States of the Union but also within a single Member State. As Article 4 TEU clearly states, the Union is committed to respecting the ‘national identity [of the Member States], inherent in their fundamental structures, political and constitutional, inclusive of regional and local self-government’. In light of that Treaty provision, one may argue that the application of EU law may not deprive regional and local governments of their regulatory autonomy. For example, diversity within a single Member State would support the contention that, in order to pursue a legitimate objective, a region should be able to adopt measures the beneficiaries of which are residents of that region alone. Yet, at the same time, it is settled case law that the internal constitutional arrangements of a Member State do not provide valid grounds justifying a breach of EU law.7 This means that Member States, such as Germany, Belgium or Spain, may not rely on their federal structure with a view to infringing EU law. Stated differently, diversity ‘in one and the same Member State’ may not lead to ‘supranational fragmentation’. By looking at the relevant case law of the ECJ on State aids, section II. is devoted to examining how EU law accommodates national federal systems within the multi-layered structure of the EU. Secondly, achieving the balance of power between ‘unity’ and ‘diversity’ is not a task for the judiciary alone, it is also a task with which the political process may be entrusted. This is what Wechsler, a famous American scholar, called ‘the political safeguards of federalism’.8 In essence, he posited that even if the United States Supreme Court renounced judicial enforcement of the principle of enumerated powers, the US Constitution would provide sufficient political ‘checks and balances’ safeguarding the autonomy of the States. Still, this does not mean that the political safeguards of federalism and the judicial process are mutually exclusive, or that one takes over where the other fails. On the contrary, they

6 K Lenaerts, ‘Federalism and the Rule of Law: Perspectives from the European Court of Justice’ (2010) 33 Fordham International Law Journal 1338. 7 See Case 69/81 Commission v Belgium [1982] ECR 153, para 5. See, more recently, Case C-247/98 Commission v Spain [2000] ECR I-2823, para 20; Case C-383/00 Commission v Germany [2002] ECR I-4219, para 18; Case C-388/01 Commission v Italy [2003] ECR I-721, paras 25–26; Case C-357/03 Commission v Austria, not reported, para 10; Case C-358/03 Commission v Austria [2004] ECR I-12055, para 13. 8 H Wechsler, ‘The Political Safeguards of Federalism: The Role of the States in the Composition and Selection of the National Government’ (1954) 54 Columbia Law Review 543.

EU Federalism in 3-D 15 mutually reinforce one another. For example, the ECJ may contribute to enhancing the political safeguards of federalism, encapsulated in the right of veto, when adjudicating over the choice of the appropriate legal basis. In order to illustrate this point, section III. looks at the ruling of the ECJ in Parliament v Council (the IFI Case).9 Last, but not least, federalism is examined in light of the Treaty provisions on EU citizenship. Section IV. thus looks at federalism from the standpoint of individual rights. As such rights operate erga omnes, they have a double dimension in federal systems. Horizontally, they are to be understood, in Dworkinian terms, ‘as trumps’ against the political majority, be that at national or EU level. Vertically, when individual rights are granted by the supranational entity, they operate as a ‘unifying device’ that brings together the citizens of the different component States, forging a common identity. Hence, supranational rights limit diversity. As to the EU, until the decision in Ruiz Zambrano was delivered,10 the application of the Treaty provisions on EU citizenship appeared to be conditioned upon the existence of a ‘cross-border’ link. Unlike rights that the national constitutions bestow upon individuals, the rights attaching to the status of EU citizen were not ‘free-standing’. However, Ruiz Zambrano has emancipated EU citizenship from the constraints inherent in its free movement origins, so that even in the absence of actual physical movement, the Treaty provisions on EU citizenship may apply. It is thus to the contextual study of that case that section IV. is devoted.

II. THE VERTICAL DIMENSION: ARTICULATING DIFFERENT FEDERAL STRUCTURES

A. General Observations Most Member States of the EU have one form or another of geographical decentralisation, as a result of which decision-making powers are vested in regional or local authorities. Various models of competence-sharing coexist in which sub-national entities take different forms and names.11 For the sake of simplicity, I shall refer to them as ‘regions’ throughout this text. In accordance with the constitutional arrangements of numerous Member States, powers which were previously exercised by the central authorities have thus been devolved to regional authorities. This ‘centrifugal trend’ may imply that infra-State bodies are granted tax powers. In Belgium, for instance, the regions are inter alia competent to determine the rate and the applicable exemptions

9

Case C-166/07 Parliament v Council (International Fund for Ireland Case) [2009] ECR I-7135. Case C-34/09 Ruiz Zambrano (ECJ, 8 March 2011). 11 See K Lenaerts and P Van Nuffel, European Union Law, 3rd edn (London, Sweet & Maxwell, 2011) para 14-014. See also K Lenaerts and N Cambien, ‘Regions and the European Courts: Giving Shape to the Regional Dimension of Member States’ (2010) 35 EL Rev 609. 10

16 Koen Lenaerts with respect to inheritance tax. They are also empowered to determine the taxes applicable on the transfer of immovable property. Other regions, such as the Basque Historical Territories in Spain, have broader tax powers and are also competent in the field of corporation tax. As mentioned in the introduction, the Treaties now impose a general obligation on the Union to respect the fundamental political and constitutional structures of the Member States as regards ‘regional and local self-government’, which are an inherent part of their national identity.12 Therefore, EU law does not oppose as such the devolution of powers—in particular, tax powers—to regional authorities within the Member States. Whether tax powers are devolved to regional authorities (and if so, how) is a constitutional question for each Member State alone to decide.13 However, each Member State must see to it that EU law is correctly implemented across the entire national territory, regardless of how powers are internally allocated. A region is indeed part and parcel of a Member State. As such, regions are bound by the Treaties. This means that EU law may impose obligations on regional authorities. Within their sphere of competence, regions are required to implement directives. It is thus settled case law that each Member State is free to allocate powers internally and to implement directives by means of measures adopted at regional level, provided that that allocation of powers enables the directive in question to be implemented correctly.14 Where a region fails to implement an EU measure, the Member State concerned will be held liable under EU law for such failure.15 For example, in a case concerning the failure of the Austrian Land of Carinthia to implement an EU directive, the ECJ recalled that though each Member State may freely allocate areas of internal legal competence as it sees fit, the fact remains that it alone is responsible to the EU under Article 258 TFEU for compliance with obligations arising under EU law.16 A Member State is thus not entitled to hide behind the domestic division of powers or federal structure in order to avoid the ECJ making a finding of an

12

Art 4(2) TEU. See, eg, Joined Cases 51/71 to 54/71 International Fruit Company and Others [1971] ECR 1107, para 4; Case C-428/07 Horvath [2009] ECR I-6355, para 49. 14 See Case C-156/91 Hansa Fleisch Ernst Mundt [1992] ECR I-5567, para 23; Horvath, above n 13, para 50. 15 Each Member State must ensure that individuals obtain reparation for damage caused to them by non-compliance with EU law, whichever public authority is responsible for the breach and whichever public authority is in principle, under the law of the Member State concerned, responsible for making reparation. See, eg, Case C-302/97 Konle [1999] ECR I-3099, para 62 (a territorial decentralised body may be held liable); Case C-424/97 Haim [2000] ECR I-5123, paras 61–62 (a functionally decentralised body may be held liable); Case C-224/01 Köbler [2003] ECR I-10239, paras 44–47 and 50 (the Member State must designate the court competent to determine disputes concerning the reparation of damage resulting from judicial decisions). See G Anagnostaras, ‘The Allocation of Responsibility in State Liability Actions for Breach of Community Law: A Modern Gordian Knot?’ (2001) 26 EL Rev 139. 16 Case C-358/03 Commission v Austria [2004] ECR I-12055, para 13; Case C-87/02 Commission v Italy [2004] ECR I-5975, para 38. 13

EU Federalism in 3-D 17 infringement,17 or to escape its obligation to bring such infringement to an end.18 Consequently, a Member State may not rely on the defence that, under national constitutional law, the federal executive has no authority to give instructions to a devolved legislative authority which is in breach of its Treaty obligations.19 It follows that whilst EU law does not interfere with the internal division of powers between national and regional authorities within a Member State, regions exercising their own constitutional powers must nevertheless do so in a manner consistent with EU law.

B. State Aids Nonetheless, the ECJ does take into account the domestic status of those regions under the constitutional law of the Member States concerned. This became clear in a number of decisions on State aid, starting with Portugal v Commission (the so-called Azores judgment).20 That case involved measures enacted by the Azores, an autonomous region of the Portuguese Republic with independent legislative powers, inter alia in the field of taxation. In accordance with its powers under Portuguese constitutional law, the legislative assembly of the Azores had significantly reduced the rates of national income and corporation tax for economic operators taxed in the Azores. The Commission considered these reductions to be State aid.21 It was undisputed that tax reductions could indeed constitute ‘advantages’ granted ‘through State resources’.22 The remaining issue to be determined,

17 See Case 69/81 Commission v Belgium [1982] ECR 153, para 5. See, more recently, Case C-247/98 Commission v Spain [2000] ECR I-2823, para 20; Case C-383/00 Commission v Germany [2002] ECR I-4219, para 18; Case C-388/01 Commission v Italy [2003] ECR I-721, paras 25–26; Case C-357/03 Commission v Austria, not reported, para 10; Commission v Austria, above n 16, para 13. 18 Case 96/81 Commission v Netherlands [1982] ECR 1791, para 12; Joined Cases 227–230/85 Commission v Belgium [1988] ECR 1, paras 9–10. 19 Case C-323/96 Commission v Belgium [1998] ECR I-5063, paras 40–42 (in which judgment was given against Belgium because the Flemish Parliament had infringed Community rules on the award of public contracts). A defence based on the internal problems in Belgium fails likewise. In that regard, the ECJ merely states that the Belgian difficulties are of a purely domestic nature, since they result from its political and administrative organisation and, accordingly, do not constitute a case of force majeure (Case C-236/99 Commission v Belgium [2000] ECR I-5657, para 22). 20 Case C-88/03 Portugal v Commission [2006] ECR I-7115. See also R Greaves, ‘Autonomous Regions, Taxation and EC State-Aid Rules’ (2009) 34 EL Rev 779. 21 See Commission Decision 2003/442/EC [2003] OJ L150/52. However, the Commission also considered that the scheme was justified under the derogation of Art 107(3)(a) TFEU, taking into account the specific situation of the Azores as an outermost region within the meaning of Art 349 TFEU except insofar as it applied to firms that carry on financial activities or activities of the ‘intragroup services’ type (ibid). 22 See, eg Case 156/98 Germany v Commission [2000] ECR I-6857, paras 26–28. For further discussion, see, inter alia, W Schön, ‘Taxation and State Aid Law in the European Union’ (1999) 36 CML Rev 911; C Quigley, ‘General Taxation and State Aid’ in A Biondi, P Eeckhout and J Flynn (eds), The Law of State Aid in the European Union (Oxford, Oxford University Press, 2004) 207; J Winter, ‘Re(de)fining the Notion of State Aid in Article 87.1 of the EC Treaty’ (2004) 41 CML Rev 505; P Nicolaides, ‘Fiscal State Aid in the EU: The Limits of Tax Autonomy’ (2004) 27 World Competition 365.

18 Koen Lenaerts however, was whether the tax reductions in question were ‘selective’ so as to amount to State aid.23 In other words, could the tax reductions be said to be a general measure, in that they applied to all economic operators taxed in the Azores, or were they a selective measure, conferring an advantage only on those operators as compared to economic operators active in other areas of Portugal? The issue boiled down to the question of the appropriate framework of reference for the selectivity assessment: the Member State, being Portugal, or the region in question, being the Azores. The Commission argued, as it had consistently done in earlier cases and in its guidelines, that the only permissible framework was the Member State as a whole.24 It inferred this, inter alia, from the general scheme of the Treaties and from the fundamental role Member States play in defining the political and economic environment in which economic operators conduct their activity. It followed, in the Commission’s view, that the tax reductions at issue were a selective measure. Portugal, on the other hand, considered that, given that the tax reductions at issue resulted from the exercise of the Azores’ regional competences, the right framework of reference was the autonomous region of the Azores and not the country as a whole.25 This would mean in turn that the tax reductions at issue were not a selective measure since they applied to all economic operators with tax residence in the Azores. The ECJ agreed with the Portuguese view. It pointed out that in regions with sufficiently autonomous powers, those regions, and not the central Member State authorities, play a fundamental role in the definition of the political and economic environment in which undertakings operate. Where such an autonomous region had adopted a tax rate lower than the national rate and applicable only to undertakings present in the territory within its competence, the relevant context of assessment was to be the geographical area in which that region exercises its powers and not the country as a whole.26 The ECJ then put forward three criteria for assessing whether a measure could be regarded as having been adopted in the exercise of sufficiently autonomous powers. First, the decision had to be taken by a regional or local authority with a political and administrative status separate from that of the central government. Next, it should have been adopted without the central government being able to intervene directly as regards its content. Lastly, the financial consequences of a reduction of the national tax rate for undertakings in the region should not have been offset by aid or subsidies from other regions or the central government.27

23 For the requirement of selectivity, see Case C-66/02 Italy v Commission [2005] ECR I-10901, para 94. 24 The Commission referred to its Notice on the application of the State aid rules to measures relating to direct business taxation [1998] OJ C384/3. Paragraph 17 of the Notice states: ‘The Commission’s decision-making practice so far shows that only measures whose scope extends to the entire territory of the State escape the specificity criterion laid down in Article [107(1) TFEU].’ 25 See Portugal v Commission, above n 20, paras 38–40. 26 Ibid paras 54–66. 27 Ibid paras 67–68.

EU Federalism in 3-D 19 Or, in the words of Advocate General Geelhoed, the region in question had to be institutionally, procedurally and economically autonomous.28 In the case before it, the ECJ found that the third of these criteria was not met. It found that the Azores lacked economic autonomy because the financial consequences of the tax reductions were partly offset by budgetary transfers from the central government. These transfers were moreover directly linked with the decision of the Azores to set a lower tax rate.29 The bottom-line was that the Azores did not bear the financial consequences of such a reduction in full. Accordingly, the lower tax rate set by the Azores was to be seen as a selective aid measure.30 These criteria were developed further by the ECJ in the Unión General de Trabajadores de la Rioja case.31 It concerned tax measures adopted by the three Territorios Históricos belonging to the Basque Autonomous Community, which have considerable autonomous powers under Spanish constitutional law.32 The measures provided for a rate of corporation tax that was significantly lower than the basic rate set in national Spanish corporation tax law, and for important tax deductions that had no equivalent under the latter law. As in the foregoing case, the question to be answered in the main proceedings was whether these measures constituted State aid, which again boiled down to the determination of what was to be considered the right framework of reference for the selectivity test. The ECJ repeated its finding that the legal framework appropriate for determining the selectivity of a tax measure may be limited to the geographical area concerned where the region occupies a fundamental role in the definition of the political and economic environment in which the undertakings present on the territory within its competence operate.33 It further clarified the three criteria it had first put forward in Commission v Portugal, without however making a final appraisal, given that it concerned a request for a preliminary ruling under Article 267 TFEU. The ECJ clarified, first of all, that a finding that the three criteria are fulfilled leads to the conclusion that the region in question plays a fundamental role in the definition of the political and economic environment in which undertakings operate. It rejected the Commission’s submission that playing such a role is 28 Portugal v Commission, above n 20, Opinion of AG Geelhoed, para 54. The ECJ itself embraced these expressions in Joined Cases C-428/06, C-429/06, C-430/06, C-431/06, C-432/06, C-433/06 and C-434/06 Unión General de Trabajadores de la Rioja and Others [2008] ECR I-6747, although it referred to the third condition as the criterion of ‘economic and financial autonomy’ (para 51). 29 See Portugal v Commission, above n 20, paras 69–79. 30 The ECJ held that the tax reductions were not justified by the nature and overall structure of the Portuguese tax system (ibid paras 80–85). 31 Unión General de Trabajadores de la Rioja and Others, above n 28. 32 Tax measures of the Historical Territories had already been the subject of a significant number of cases before the ECJ and CFI. AG Kokott in her Opinion states that more than 40 cases had already been brought before the European Courts in this connection (Unión General de Trabajadores de la Rioja and Others, above n 28, Opinion of AG Kokott, para 5). Among the most recent ones, see Joined Cases T-30/01 to T-32/01 and T-86/02 to T-88/02 Diputación Foral de Álava and Others v Commission [2009] ECR II-2919, confirmed on appeal by Joined Cases C-465 to 470/09 P Diputación Foral de Vizcaya and Others v Commission (ECJ, 9 June 2011). 33 Portugal v Commission, above n 20, para 50.

20 Koen Lenaerts an additional criterion that needs to be fulfilled before the said three criteria are assessed.34 With regard to the second criterion, the ECJ held that a region’s ‘obligation to take into consideration the State interest in order to respect the limits of the areas of competence which are accorded to it’, does not, generally, constitute an element calling into question the procedural autonomy of that region where it adopts a decision within those limits.35 The essential element in this regard is whether or not that body adopts decisions independently, ie without the central government being able to intervene directly as regards their content.36 With regard to the last criterion, the ECJ held that the mere existence of financial transfers between the central State and its regions does not suffice to demonstrate that these regions do not assume the financial consequences of the tax measures which they adopt. For the economic autonomy of a region to be called into question, it is necessary that there is a causal relationship between the tax measure adopted and the compensation paid by the central State, or another region, to the region in question.37 It follows from the case law just discussed that regional tax measures will no longer automatically be considered to be selective aid measures. Tax measures adopted by regions with sufficient institutional, procedural, and economic and financial autonomy will be considered in their own right and not by reference to tax measures adopted by the central government of the Member State of which they form a part.38 Accordingly, the Court of First Instance (now the General Court) annulled a Commission decision which declared Gibraltar’s proposed company tax legislation to be regional aid incompatible with the Treaties.39 It considered, inter alia, that the Commission had wrongly taken the United Kingdom as the framework for its analysis rather than just Gibraltar, since Gibraltar fulfilled all of the three criteria established by the ECJ in its Azores judgment.40 Obviously, although such regional tax measures will not be considered to be regionally selective, they may still be materially selective and qualify as State aid.41 34

Ibid paras 53–60. Ibid para 108. 36 Ibid para 107. The ECJ held in this connection that the mere fact that decisions by a region are subject to judicial review by the courts of the Member States to which they belong does not as such call into question their procedural autonomy, where these courts review compliance with the preestablished limits on the areas of competence of the different State authorities, organs or bodies, but do not determine those limits (ibid paras 79–83). 37 Ibid paras 123–35. 38 See the detailed discussion in J Bousin and J Piernas, ‘Developments in the Notion of Selectivity’ (2008) European State Aid Law Quarterly 643. 39 See Cases T-211/04 and T-215/04 Gibraltar v Commission [2008] ECR II-3745. An appeal is currently pending before the ECJ (see Joined Cases C-106/09 P and C-107/09 P Commission and Spain v Gibraltar and United Kingdom). For a discussion, see P Rossi-Maccanico, ‘Gibraltar and the Unsettled Limits of Selectivity in Fiscal Aids’ (2009) European State Aid Law Quarterly 63. 40 Portugal v Commission, above n 20, paras 89–116. 41 See Case C-169/08 Presidente del Consiglio dei Ministri v Regione Sardegna [2009] ECR I-10821, in which the ECJ considered that a regional tax on stopovers for tourist purposes, adopted by the Region of Sardinia, was selective since it was applicable only to operators of aircraft or pleasure boats whose tax domicile was outside Sardinia (see paras 56–63). For other examples, see Joined Cases 35

EU Federalism in 3-D 21 The case law discussed is relevant first and foremost for Member States with an asymmetrical system of devolution of powers in the field of taxation, in which certain regions enjoy more autonomy than others in determining the tax applicable in the territory within their competence. If one of those regions sets a tax rate lower than the national rate, it will be considered to have been granted a selective advantage if the reference framework is held to be the Member State as a whole. By contrast, if all regions enjoy the power to set tax rates, there is no national reference rate that can serve to determine selectivity and, hence, regional tax measures cannot be qualified as State aid.42 Not surprisingly, the cases discussed involve Member States with asymmetrical devolution of powers, namely Portugal, Spain and the United Kingdom.43 But in our view the case law is also of great importance to other Member States in which important powers are conferred upon bodies at regional or local levels, not least federal Member States like Austria, Belgium or Germany, regardless of whether the devolution of powers is symmetrical or asymmetrical.44

C. Other Fields The cases on State aid just discussed show that the European Courts have proved sensitive to the constitutional structure of Member States, and to the competences attributed to regional and local governments. However, this sensitivity of the Courts is not limited to the field of State aid. It is long-established case law that Member States are free to allocate powers internally as they see fit and to leave implementation of some aspects of EU law to infra-State bodies.45 The law of the EU does not intervene in the internal allocation of powers as long as such allocation enables EU law to be implemented correctly.46

T-127/99, T-129/99 and T-148/99 Territorio Histórico de Álava and Others v Commission [2002] ECR II-1275; Joined Cases T-92/00 and T-103/00 Territorio Histórico de Álava and Others v Commission [2002] ECR II-1385, in which regional tax measures were held to be materially selective rather than regionally selective (upheld on appeal: Cases C-183/02 P and C-187/02 P Demesa and Territorio Histórico de Álava v Commission [2004] ECR I-10609 and Joined Cases C-186/02 P and C-188/02 P Ramondín and Others v Commission [2004] ECR I-10653, respectively). 42 See the three scenarios distinguished by the ECJ in Portugal v Commission, above n 20, paras 63–65 (referring to the Opinion of AG Geelhoed, paras 50–54). See also Commission, ‘Report on the implementation of the Commission notice on the application of the state aid rules to measures relating to direct business taxation’ C(2004)434, para 33. 43 In addition, Spain and the UK intervened in the Azores case in favour of the Portuguese tax system (see Portugal v Commission, above n 20, paras 45 and 49). 44 In Belgium, the devolution of tax powers to regions and municipalities should probably be classified as ‘symmetrical’, but it remains uncertain to what extent symmetry is required in order to escape the test of the three criteria set out by the ECJ. See E Mattioli, M Vervoort and T Bruyninckx, ‘Fiscale voordelen vanwege regionale en lokale overheden: onder staatssteunbedreiging?’ (2009) Tijdschrift voor Fiscaal Recht 387. 45 See already Case 96/81 Commission v Netherlands [1982] ECR 1791, para 12. 46 In this connection, see above n 7.

22 Koen Lenaerts This is amply demonstrated by the Horvath case.47 It concerned the implementation in the UK of a regulation laying down rules on income support for farmers in the framework of the Common Agricultural Policy.48 In the UK, the Common Agricultural Policy is among the matters which fall within the responsibility of each devolved administration. Accordingly, each of the devolved administrations adopted rules to implement the EU Regulation concerned. The rules adopted in respect of England laid down stricter minimum requirements for good agricultural and environmental condition than those adopted by the devolved administrations of Scotland, Wales and Northern Ireland. The applicant in the case, a farmer entitled to income support under the Union rules, argued that this situation amounted to unlawful discrimination under EU law. The ECJ did not accept this view. It merely referred to its case law holding that ‘the prohibition on discrimination is not concerned with any disparities in treatment which may result, between the Member States, from divergences existing between the legislation of the various Member States so long as that legislation affects equally all persons subject to it’,49 and applied this to the relationship between regional authorities within a single Member State. Accordingly, in the case at hand, mere divergences between the measures adopted by the different devolved administrations in order to comply with Regulation No 1782/2003 did not constitute discrimination. The fact that the ECJ now applies this case law, mutatis mutandis, to differences between the rules adopted by regional authorities of a single Member State demonstrates, again, that, in areas in which such authorities are autonomously competent to adopt rules, they will from the viewpoint of EU law be equated with Member States. Accordingly, the appropriate reference framework to assess discrimination in Horvath was not the UK as a whole but rather the territories of the different devolved administrations.

D. Concluding Remarks The case law discussed amply demonstrates the openness of the ECJ towards the constitutional structure of the Member States and the internal allocation of powers to regional and local governments. Where Member States have endowed regions with sufficiently autonomous powers, the ECJ will assess the implementation of EU law by reference to the territory in which each of these regions is competent rather than by reference to that of the Member State to which they 47

Horvath, above n 13. Council Regulation (EC) 1782/2003 establishing common rules for direct support schemes under the common agricultural policy and establishing certain support schemes for farmers and amending Regulations (EEC) 2019/93, (EC) 1452/2001, (EC) 1453/2001, (EC) 1454/2001, (EC) 1868/94, (EC) 1251/1999, (EC) 1254/1999, (EC) 1673/2000, (EEC) 2358/71 and (EC) 2529/2001 [2003] OJ L270/1. 49 Ibid para 55. The ECJ referred to Joined Cases 185/78 to 204/78 Van Dam and Others [1979] ECR 2345, para 10; Case C-177/94 Perfilli [1996] ECR I-161, para 17; and Case C-403/03 Schempp [2005] ECR I-6421, para 34. 48

EU Federalism in 3-D 23 belong. In the specific field of State aid, the ECJ has, moreover, announced clear criteria for determining whether a region may be considered to be sufficiently autonomous. There seems to be no good reason why the same criteria could not be used in other contexts as well to assess compliance with EU rules.50 As Advocate General Kokott observes in her Opinion in the Trabajadores de la Rioja case,51 the ECJ’s case law successfully strikes a delicate balance between two important principles. On the one hand, it respects the autonomy which Member States confer on their regional authorities. By showing itself prepared to consider measures in the constitutional context of the Member State concerned, the ECJ takes a more realistic view than the Commission’s traditional assumption that central Member State authorities always play the key role in a Member State’s economy. In a sense the ECJ’s reasoning is a reference avant la lettre to the first sentence of current Article 4(2) TEU, as modified by the Treaty of Lisbon. On the other hand, the ECJ is clearly sensitive to the argument that Member States should not be allowed to hide behind their constitutional structure and circumvent EU rules through a purely formal internal division of powers. Precisely for this reason it puts forward rather stringent criteria, which will be fulfilled only in the case of regions that are truly autonomous. At the end of the day, the balance struck by the ECJ perfectly guarantees effective enforcement of EU law while respecting each Member State’s constitutional choices as to the conferral of regional autonomy.

III. THE HORIZONTAL DIMENSION: THE APPROPRIATE LEGAL BASIS

A. General Observations on the Choice of the Appropriate Legal Basis In accordance with the principle of conferral, all EU measures must be grounded in a legal basis provided for by the Treaties.52 As the ECJ stated in its Opinion 2/00, ‘the choice of the appropriate legal basis has constitutional significance, since, having only conferred powers, the [EU] must tie the contested decision to a Treaty provision which empowers it to approve such a measure’.53 The requirement for all EU measures to have the appropriate legal basis (or, as the case may be, the appropriate legal bases) pursues two different, albeit intertwined, objectives.

50 Interesting to note in this regard is that AG Trstenjak in her Opinion in the Horvath case explicitly referred to the criteria developed by the ECJ for assessing regional selectivity in State aid cases. The Advocate General considered that the same approach was to be applied mutatis mutandis to the situation of that case ‘because the criteria employed by the Court are an aid in assessing the existence of a situation which, irrespective of the sector-specific terminology—“selectivity” in the case of the law on aid—ultimately constitutes no more than a difference in treatment between economic operators’ (Horvath, above n 13, Opinion of AG Trstenjak). 51 Unión General de Trabajadores de la Rioja and Others, above n 28, Opinion of AG Kokott, paras 54–57. 52 See also Lenaerts and Van Nuffel, above n 11, para 7-009. 53 See Opinion 2/00 [2001] ECR I-9713, para 5.

24 Koen Lenaerts On the one hand, the choice of the appropriate legal basis ensures compliance with the principle of institutional balance, as it prevents an EU institution from wrongfully choosing a legal basis which increases its own decision-making powers to the detriment of the powers of another EU institution.54 This point is illustrated by cases where the ECJ was asked to determine whether the Council was wrong to choose a legal basis providing for the consultation procedure (or consent procedure) instead of a legal basis providing for the co-decision procedure.55 On the other hand, in the absence of a legal basis, the EU is not empowered to act: such action comes within the residuary competence of the Member States.56 This is now made clear by Article 4(1) TEU, which states that, in accordance with Article 5 TEU, ‘competences not conferred upon the Union in the Treaties remain with the Member States’.57 In addition, the choice of the appropriate legal basis aims to ensure the good functioning of ‘the political safeguards of federalism’.58 Depending on the type of voting rules within the Council that are applicable to the EU measure in question, a Member State will enjoy more or less power. The more power it enjoys within the Council, the more capacity it has to block EU measures which oppose its own national interests.59 A Member State is at the peak of its powers where the Council’s decision-making procedure requires there to be a unanimous vote, ie each Member State has a right of veto. For example, Article 352(1) TFEU, which may be defined as the ‘suppletive legal basis’, is designed to fill the gap where no specific provisions of the Treaty confer on the [EU] institutions express or implied powers to act, if such powers appear none the less to be necessary to enable the [EU] to carry out its functions with a view to attaining one of the objectives laid down by the Treaty.60

Article 352(1) TFEU operates as a ‘residual clause’, since that provision ‘may be used as the legal basis for a measure only where no other provision of the [TFEU] gives the [EU] institutions the necessary power to adopt it’.61 Under that Treaty provision, the Council must act unanimously, after obtaining the consent of the European Parliament. Since that Treaty provision expands the powers of the EU,

54

Lenaerts and Van Nuffel, above n 11, para 7-010. See T Tridimas and G Gari, ‘Winners and Losers in Luxembourg: A Statistical Analysis of Judicial Review before the European Court of Justice and the Court of First Instance (2001–2005)’ (2010) 35 EL Rev 131. 56 Lenaerts and Van Nuffel, above n 11, para 7-011. 57 Art 4(1) TEU. 58 This expression is borrowed from H Wechsler, above n 8. For an excellent application of that doctrine to the EU context, see E Young, ‘Protecting Member State Autonomy in the European Union: Some Cautionary Tales from American Federalism’ (2002) 77 New York Law Review 1612. 59 See also Lenaerts and Van Nuffel, above n 11, para 7-011. 60 Opinion 2/94 [1996] ECR I-1759, para 29; Joined Cases C-402/05 P and C-415/05 P Kadi and Al Barakaat International Foundation v Council and Commission [2008] ECR I-6351, para 211. 61 See, eg, Case 45/86 Commission v Council [1987] ECR 1493, para 13; Case C-350/92 Spain v Council [1995] ECR I-1985, para 26; Case C-84/94 United Kingdom v Council [1996] ECR I-5755, para 48; Case C-22/96 Parliament v Council [1998] ECR I-3231, para 22; and Case C-436/03 Parliament v Council [2006] ECR I-3733, para 36. 55

EU Federalism in 3-D 25 without reforming the general framework set out in the Treaties,62 the unanimity voting rule within the Council operates as a political check: it ensures that the interests of the Member States are duly protected. The same applies when the EU legislator chooses a legal basis which gives the EU exclusive competence over a legal basis which provides only for shared competence. If the EU legislator decides to adopt an EU measure under a Treaty provision in relation to which the EU enjoys exclusive competence,63 it will not be possible for the national parliaments to check whether the EU legislator has complied with the principle of subsidiarity.64 It follows from the foregoing that the choice of the appropriate legal basis preserves the vertical and horizontal constitutional balance sought by the authors of the Treaties. In order to determine the choice of the appropriate legal basis, the EU institution’s subjective view as to the objective pursued is irrelevant, given that ‘the choice of legal basis for [an EU] measure must rest on objective factors which are amenable to judicial review, including in particular the aim and the content of the measure’.65 If an EU measure pursues two aims or components, one of those aims or components being identifiable as the main one, the other being merely incidental, ‘the measure must be founded on a single legal basis, namely that required by the main or predominant aim or component’.66 Conversely, where an EU measure ‘simultaneously pursues a number of objectives or … has several components, without one being incidental to the other, … such a measure will have to be founded, exceptionally, on the various corresponding legal bases’.67 However, the use of a dual legal basis is not possible ‘where the procedures laid down for each legal basis are incompatible with each other or where the use of two legal bases is liable to undermine the rights of the [European] Parliament’.68 In such a case, the ECJ must decide which of the two legal bases applies. Moreover, if

62 Opinion 2/94, above n 60, para 30, where the ECJ stated that ‘Article [352 TFEU], being an integral part of an institutional system based on the principle of conferred powers, cannot serve as a basis for widening the scope of [EU] powers beyond the general framework created by the provisions of the [Treaties] as a whole and, in particular, by those defining the tasks and the activities of the [EU]’. See also Kadi, above n 60, para 203. 63 The TFEU lists the areas where the EU enjoys exclusive competence (Art 3 TFEU) as well as the areas where the EU and the Member States share competences (Art 4 TFEU). 64 See Art 12 TEU. See also Protocol No 1 on the role of National Parliaments in the European Union and Protocol No 2 on the application of the principles of subsidiarity and proportionality. 65 See, eg, Case C-300/89 Commission v Council (Titanium dioxide) [1991] ECR I-2867, para 10; Case C-336/00 Huber [2002] ECR I-7699, para 30; Case C-176/03 Commission v Council [2005] ECR I-7879, para 45; and Case C-440/05 Commission v Council [2007] ECR I-9097, para 61. See also M Klamert, ‘Conflicts of Legal Basis: No Legality and No Basis but a Bright Future under the Lisbon Treaty?’ (2010) 35 EL Rev 497. 66 See Case C-155/91 Commission v Council [1993] ECR I-939, paras 19 and 21; Case C-36/98 Spain v Council [2001] ECR I-779, para 59; Case C-338/01 Commission v Council, para 55; and Case C-91/05 Commission v Council [2008] ECR I-3651, para 73. 67 Case C-211/01 Commission v Council [2003] ECR I-8913, para 40; and Case C-91/05 Commission v Council, above n 66, para 75. 68 Commission v Council (Titanium dioxide), above n 65, paras 17–21; Case C-94/03 Commission v Council (Rotterdam Convention) [2006] ECR I-1, para 52.

26 Koen Lenaerts the EU legislator was wrong to rely on a dual legal basis, the appropriate legal basis being a single Treaty provision, then the EU measure in question will be annulled in so far as that error has vitiated the procedure for adopting it. Conversely, if the essential point of the legislative procedure has not been in any way compromised by the simultaneous reference to the two legal bases, the contested EU measure will not be annulled.69

B. Parliament v Council (the IFI Case) In Parliament v Council (the IFI Case),70 the ECJ opted for a compromise solution, accommodating the political safeguards of the Member States (their right of veto) with the prerogatives of the European Parliament (the co-decision procedure). The facts of the case may be summarised as follows. In order to consolidate peace and reconciliation between nationalists and unionists throughout Ireland, the UK and Ireland concluded the 1985 Anglo-Irish Agreement.71 Article 10(a) of that agreement provides that [t]he two Governments shall co-operate to promote the economic and social development of those areas of both parts of Ireland which have suffered most severely from the consequences of the instability of recent years, and shall consider the possibility of securing international support for this work.

To that end, the UK and Ireland concluded another international Agreement creating the International Fund for Ireland (‘the IFI’),72 which is an international organisation the objectives of which are to promote economic and social advance, and to encourage contact, dialogue and reconciliation between nationalists and unionists throughout Ireland.73 Along with the United States, Canada, New Zealand and Australia, the EU is a donor to the IFI. Regulation No 1968/2006 (the ‘contested regulation’) provided the legislative framework for the payment to the IFI of the EU’s financial contributions for the period 2007 to 2010. It was adopted on the basis of ex Article 308 EC (now Article 352(1) TFEU).74 However, the European Parliament brought an action for annulment against that regulation, arguing that the Council had chosen the wrong legal basis. In its view, the EU enjoyed the necessary powers under ex Article 159 EC (now Article 175 TFEU) to adopt Regulation No 1968/2006, since that regulation sought to promote social and economic cohesion in Ireland. As a matter of fact, the EU had 69 Case C-491/01 British American Tobacco (Investments) and Imperial Tobacco [2002] ECR I-11453, paras 108–11. 70 Parliament v Council (International Fund for Ireland Case), above n 9. 71 The Anglo-Irish Agreement was done at Hillsborough on 15 November 1985. See United Nations Treaty Series, vol 1413, no I-23668. It entered into force on 29 November 1985. 72 See the Agreement concerning the International Fund for Ireland (the ‘IFI Agreement’) signed on 18 September 1986. See United Nations Treaty Series, vol 1515, no I-26244. 73 See Art 5 of the IFI Agreement. 74 [2006] OJ L409/86.

EU Federalism in 3-D 27 already relied on that Treaty provision in order to grant structural funding having the same objectives as those pursued by the IFI (the PEACE programme).75 Given that ‘[a]ctions to support intra-community reconciliation in Ireland inevitably constitute an integral part of the cohesion policy precisely because, without reconciliation and mutual understanding between the communities, there can be no economic and social cohesion in the areas in question’,76 the European Parliament submitted that Regulation No 1968/2006 should have been adopted under ex Article 159 EC (now Article 175 TFEU). It further argued that ex Article 159 EC (now Article 175 TFEU) specified neither the sectors in which ‘specific actions’ outside the structural funds might be instituted, nor the forms which those actions might take. Hence, the promotion of a harmonious development and the reduction of disparities between the levels of development of the various regions of the EU could be carried out by funding the activities of an international organisation such as the IFI. From an institutional perspective, the action for annulment brought by the European Parliament may be explained by the fact that its powers are greater under ex Article 159 EC (now Article 175 TFEU) than under ex Article 308 EC (now Article 352(1) TFEU). Whilst ex Article 308 EC provided for the consultation procedure,77 Article 159 EC (now Article 175 TFEU) empowers the European Parliament to stand on an equal footing with the Council: the European Parliament and the Council act as co-legislators. The Council did not, however, share the views of the European Parliament. For the Council, ex Article 159 EC (now Article 175 TFEU) did not provide for the powers to act required by the IFI’s activity and could not therefore supply an appropriate legal basis for the grant of the EU financial contributions to the IFI. According to the Council, the objectives of Regulation No 1968/2006 were to promote peace and reconciliation between nationalists and unionists in Ireland, which is an objective falling outside the scope of ex Article 158 EC (now Article 174 TFEU). A distinction had thus to be drawn between the objectives of the PEACE programme and those of the IFI: the IFI aims at reconciliation in order to facilitate cohesion, whereas the PEACE programme aims at cohesion in order to facilitate reconciliation. Stated differently, the Council opined that ‘the objective of economic and social cohesion is merely a consequence of the pursuit of reconciliation secured by the action of [the IFI, which is] an international organisation

75 Parliament v Council (International Fund for Ireland Case), above n 9, para 17, ‘The PEACE programme is [an EU] initiative established under the Structural Funds. Pursuant to paragraph 22 of Annex II to Regulation No 1083/2006, that programme is implemented as a cross-border programme within the meaning of Article 3(2)(c) of Regulation No 1083/2006 under the objective of European territorial cooperation. It includes, notably, actions to promote cohesion between communities, in order to promote social and economic stability in the areas concerned. The eligible area corresponds to the whole of Northern Ireland and the border counties of Ireland.’ 76 Ibid para 23. 77 But see Art 352(1) TFEU, which increases the powers of the European Parliament by replacing the consultation procedure with the consent procedure.

28 Koen Lenaerts external to the [EU]’.78 In addition, the Council pointed out that ex Article 159 EC (now Article 175 TFEU) might be relied upon only if financial aid was administered in accordance with the regulatory framework laid down in ex Title XVII of the EC Treaty (now Title XVIII of the TFEU). Therefore, even if one were to assume that the IFI’s priority was economic cohesion rather than reconciliation, neither Title XVIII of the TFEU nor the EU regulatory framework could apply to the IFI, an international organisation of which the EU is not a member. Moreover, the Council rejected the use of a dual legal basis, since the objective of economic and social cohesion was incidental to the promotion of peace and reconciliation in Ireland.79 From an institutional perspective, the Council’s reliance on ex Article 308 EC (now Article 352(1) TFEU) may be explained by the fact that the Council enjoys greater powers under the former Treaty provision than under ex Article 159 EC (now Article 175 TFEU). Arguably, ex Article 308 EC (now Article 352(1) TFEU) protects the interests of the Member States better than a ‘flexible and expansive interpretation’ of ex Article 159 EC (now Article 175 TFEU). Indeed, in cases where it is not easy to ascertain whether an EU measure is limited to promoting social and economic cohesion or whether it goes beyond that objective, the interests of the Member States are better safeguarded under ex Article 308 EC (now Article 352(1) TFEU), given that under that Treaty provision Member States enjoy a right of veto. The ECJ agreed with neither the European Parliament nor the Council.80 It concurred with the European Parliament that Regulation No 1968/2006 promoted social and economic cohesion. Whilst the 1985 Anglo-Irish Agreement sought to promote peace and reconciliation, the ECJ found that the IFI Agreement, which created the IFI, sought to promote social and economic cohesion. However, the ECJ agreed with the Council in that Regulation No 1968/2006 did not prevent the use by the [IFI of the EU’s] contribution to cover actions which, while complying with the objectives of the IFI Agreement, extend beyond the scope of the [EU’s] policy on economic and social cohesion or, at least, are not managed in accordance with the criteria applied by the [EU] within the framework of that policy.81

Hence, ex Article 159 EC (now Article 175 TFEU) did not suffice to cover all the activities of the IFI for which EU funding was used.

78

Parliament v Council (International Fund for Ireland Case), above n 9, para 34. Ibid para 34. 80 But see ibid, Opinion of AG Bot. The Advocate General urged the ECJ to uphold the action for annulment brought by the European Parliament. 81 Ibid para 59. Indeed, since the EU was a member neither of the IFI nor of its Board, the EU did not have any means of preventing the IFI from using EU funds for purposes other than the promotion of social and economic cohesion (ibid para 60). In addition, Regulation No 1968/2006 did not exhaustively determine the activities to be financed (ibid para 61). Neither did it set out the substantive conditions with regard to the activities to be financed by those contributions which diverged from those set out in Art 2 thereof (ibid para 62). 79

EU Federalism in 3-D 29 Next, the ECJ proceeded to examine whether the EU legislator should have had recourse to both ex Article 159 EC (now Article 175 TFEU) and ex Article 308 EC (now Article 352(1) TFEU) as a dual legal basis of the contested regulation. In order to rely on ex Article 308 EC, the ECJ noted that the contested regulation had to relate to ‘the operation of the common market’.82 That was indeed the case. In light of recital 17 of Regulation No 1968/2006, the ECJ observed that the latter sought ‘to bring about economic improvements in disadvantaged areas of two Member States and thus relates to the functioning of the common market’.83 Accordingly, the contested regulation should have been adopted under the concomitant use of those two Treaty provisions. This meant that the adoption of the contested regulation would have required both the application of the co-decision procedure and for the Council to act unanimously. The ECJ thus annulled Regulation No 1968/2006, whilst maintaining its legal effects until the entry into force, within a reasonable time, of a new regulation adopted on an appropriate legal basis. It is worth noting that such a regulation was adopted on 15 December 2010, ie after the entry into force of the Treaty of Lisbon.84

C. Concluding Remarks Parliament v Council (the IFI Case) is a positive contribution to the case law of the ECJ on the choice of the appropriate legal basis. It demonstrates that the ECJ strives to find the right balance between two fundamental constitutional principles, namely federalism and separation of powers. It also reveals that the ECJ decided to adopt a solution that—as such—was abandoned by the authors of the Treaty of Lisbon as it effectively imposed on the European Parliament and the Council a requirement to adopt a new regulation in accordance with the ordinary legislative procedure, combined with a unanimous vote within the Council.85 Indeed,

82 Ibid para 67. It is worth noting that the Treaty of Lisbon has replaced the expression ‘in the course of the operation of the common market’ by ‘within the framework of the policies defined in the Treaties’. This means that, unlike ex Art 308 EC, Art 352(1) TFEU is no longer confined to the realm of the common market but may be relied upon in the context of other EU policies, such as the Area of Freedom, Security and Justice. But Art 352 TFEU contains three additional paragraphs: Art 352(2) TFEU provides that, in relation to the monitoring of the principle of subsidiarity, ‘the Commission shall draw national Parliaments’ attention to proposals based on this Article’; Art 352(3) TFEU provides that ‘[m]easures based on this Article shall not entail harmonisation of Member States’ laws or regulations in cases where the Treaties exclude such harmonisation’; lastly, Art 352(4) TFEU states that ‘[t]his Article cannot serve as a basis for attaining objectives pertaining to the [Common Foreign and Security Policy]’. See generally M Dougan, ‘The Treaty of Lisbon 2007: Winning Minds, not Hearts’ (2008) 45 CML Rev 617, 654 ff. 83 Parliament v Council (International Fund for Ireland Case), above n 9, para 68. 84 European Parliament and Council Regulation (EU) 1232/2010 concerning European Union financial contributions to the International Fund for Ireland (2007 to 2010) [2010] OJ L346/1. Regulation No 1232/2010 was adopted under Arts 175 TFEU and 352(1) TFEU. 85 This is exactly what the European Parliament and the Council ended up doing when they adopted a new regulation, see above n 84.

30 Koen Lenaerts unlike the EC Treaty,86 no provision of the TEU or TFEU combines unanimity within the Council with the co-decision procedure, ie now the ordinary legislative procedure.87 But this does not mean that the ECJ acted in conflict with the view of the authors of the Treaties. On the contrary, since the latter did not oppose the use of a dual legal basis, it is a matter for the ECJ to decide whether the use of two (or more) Treaty provisions as legal bases is required. It is worth noting that in Parliament v Council (the IFI Case) the ECJ did not refer to the Titanium dioxide line of case law,88 according to which the use of a dual legal basis is not possible ‘where the procedures laid down for each legal basis are incompatible with each other or where the use of two legal bases is liable to undermine the rights of the [European] Parliament’.89 Formulated positively, this amounts to subjecting the use of a dual legal basis to two cumulative conditions. Whilst the ruling of the ECJ in Parliament v Council (the IFI Case) definitely fulfils the second condition, the first one did not seem readily to be met. Indeed, the solution adopted by the ECJ enhances the powers of the European Parliament, as the new regulation had to be adopted under the co-decision procedure, whereas the decision-making procedure of ex Article 159 EC (now Article 175 TFEU) and that of ex Article 308 EC (now Article 352(1) TFEU) provided for different voting requirements within the Council. The fact, however, that the use of a dual legal basis was the only means of preserving simultaneously the principle of separation of powers and the political safeguards of federalism may explain why the condition of compatibility of the applicable decision-making procedures set out in Titanium dioxide was not held to be decisive in Parliament v Council (the IFI Case). Had the ECJ interpreted ex Article 159 EC broadly, the competences of the EU would have been expanded without the Member States having a right of veto, circumventing the political safeguards laid down in ex Article 308 EC. In the same way, had it relied on ex Article 308 EC alone, the EU would have expanded its competences whilst weakening the powers of the European Parliament.90 It follows that only recourse to Article 352(1) TFEU in conjunction with another Treaty provision was capable of ensuring that the expansion of EU competence would not amount to disturbing 86 See, eg ex Art 42 EC; ex Art 47(2) EC; ex Art 151(5) EC and compare them, respectively, to Art 48 TFEU; Art 53(1) TFEU; Art 167(5) TFEU. 87 Dougan, above n 82, 640 (arguing that ‘variants of the co-decision procedure which currently instruct the Council to act by unanimity will see QMV become fully applicable’). 88 Klamert, above n 65, fn 63. 89 Commission v Council (Titanium dioxide), above n 65, paras 17–21; Case C-94/03 Commission v Council (Rotterdam Convention) [2006] ECR I-1, para 52. 90 In this regard, the approach followed in Parliament v Council (the IFI Case) follows that of the ECJ in Kadi, above n 60. In that case, the ECJ upheld the concomitant use of ex Arts 60 EC, 301 EC and 308 EC as legal bases, in spite of the fact that ex Arts 60 EC and 301 EC excluded the intervention of the European Parliament and provided for qualified majority voting within the Council. The ECJ reasoned that ‘adding [ex] Article 308 EC to the legal basis of the contested regulation enabled the European Parliament to take part in the decision-making process relating to the measures at issue which are specifically aimed at individuals whereas, under [ex] Articles 60 EC and 301 EC, no role is provided for that institution’ (ibid para 235).

EU Federalism in 3-D 31 the horizontal and vertical ‘checks and balances’ put in place by the authors of the Treaties. The preservation of the overall constitutional equilibrium was held to be more important than the strict compatibility of the decision-making procedures laid down in the Treaty provisions concerned. The latter problem could indeed be solved in practical terms.91

IV. THE DIAGONAL DIMENSION: THE EMANCIPATION OF EU CITIZENSHIP

A. General Observations A feature shared by both the Treaty provisions on free movement and those on EU citizenship is that there must be ‘a link’ or nexus with EU law. For the fundamental freedoms, no link exists where the situation at issue is purely internal.92 As the ECJ pointed out in Saunders, [t]he provisions of the Treaty on freedom of movement for workers cannot therefore be applied to situations which are wholly internal to a Member State, in other words, where there is no factor connecting them to any of the situations envisaged by [EU] law.93

Originally, the incorporation of the provisions on EU citizenship into the Treaty in 1993 changed nothing in this respect. In this regard, in paragraph 23 of Uecker and Jacquet, the ECJ held that citizenship of the Union … is not intended to extend the scope ratione materiae of the Treaty also to internal situations which have no link with [EU] law … Any discrimination which nationals of a Member State may suffer under the law of that State falls within the scope of that law and must therefore be dealt with within the framework of the internal legal system of that State.94

It has been argued that the ‘cross-border’ link requirement amounts to a reformulation of the principle of conferral for the judicial enforcement of Treaty limits

91

See above nn 84 and 85. In relation to goods, see eg Case 98/86 Mathot [1987] ECR 809; and Case 286/81 Oosthoek’s Uitgeversmaatschappij [1982] ECR 4575. As to establishment, see eg Case 204/87 Bekaert [1988] ECR 2029; Joined Cases C-54/88, C-91/88 and C-14/89 Nino [1990] ECR I-3537. As to workers, see eg Case 175/78 R v Saunders [1979] ECR 1129; Joined Cases 35/82 and 36/82 Morson and Jhanjan v State of the Netherlands [1982] ECR 3723; Case C-132/93 Steen II [1992] ECR I-2715. Regarding services, see eg Case C-108/98 RI.SAN. [1999] ECR I-5219; Case C-97/98 Jägerskiöld [1999] ECR I-7319; and Case C-245/09 Omalet, judgment of 22 December 2010, not published in the reports. As to capital, see Joined Cases C-515/99, C-519/99 to C-524/99 and C-526/99 to C-540/99 Reisch and Others [2002] ECR I-2157. In relation to citizenship, see Joined Cases C-64/96 and C-65/96 Uecker and Jacquet [1997] ECR I-3171; Case C-127/08 Metock and Others [2008] ECR I-6241. 93 Saunders, above n 92, para 11. 94 Uecker and Jacquet, above n 92, para 23. See also Case C-148/02 Garcia Avello [2003] ECR I-11613, para 26; Case C-403/03 Schempp [2005] ECR I-6421, para 20; Case C-192/05 Tas-Hagen et Tas [2006] ECR I-10451, para 23; C-212/06 Government of the French Community and Walloon Government [2008] ECR I-1683, para 39; Case C-499/06 Nerkowska [2008] ECR I-3993, para 25. 92

32 Koen Lenaerts imposed upon the Member States.95 As Ritter explains, extending the application of the fundamental freedoms to purely internal situations would result in an incursion into national competences, which would deprive the Member States of the power to regulate the factors of production by reference to policy objectives other than those recognised as legitimate by EU law.96 The truth is that determining the presence or absence of a link with EU law has significant repercussions for the vertical allocation of powers. The more broadly the ‘link’ with EU law is defined, the wider the substantive scope of EU law will be. From a federal perspective, a broad definition of the link with EU law significantly restricts the exercise of competences pertaining to the Member States.97 On the contrary, a restrictive definition leaves more room to the national authorities. The case law relating to free movement and EU citizenship indicates that the ECJ has not been too strict in defining the required link with EU law. Five types of case illustrate this point. First, the ECJ has applied EU law to situations where barriers are erected to insulate a territory from other parts of the same Member State.98 Secondly, the ECJ has also adopted a relaxed approach when examining whether a contested national measure has a deterrent effect on the exercise of EU rights.99 Thirdly, cases like D’Hoop show that EU law applies when free movers return to their own Member State.100 Fourthly, it is possible for a person to invoke the EU rights of a third party, in so far as there is a ‘direct link’ between the legal position of that person and the rights of the third party.101 Last but not least, the ECJ has applied the Treaty provisions on EU citizenship in cases where there is no physical movement from one Member State to another, but where there are other elements that serve as a sufficient connecting factor to EU law. Since the absence of physical movement played an important role in the rationale of the ECJ in its recent judgments in Rottmann, Ruiz Zambrano and McCarthy,102 it is worth looking at how a link with EU law was established in those situations. For example, in Garcia Avello,103 one of the first cases of this trend, Mr Garcia Avello and Ms Weber—a Spanish-Belgian couple living in

95 Others have argued that it enshrines the principle of subsidiarity. See P Oliver, ‘Some Further Reflections on the Scope of Articles 28–30 (ex 30–36) EC’ (1999) 36 CML Rev 783. 96 C Ritter, ‘Purely Internal Situations, Reverse Discrimination, Guimont, Dzodzi and Article 234’ (2006) 31 EL Rev 690, 692. 97 S O’Leary, The Evolving Concept of Community Citizenship: From the Free Movement of Persons to Union Citizenship, 13 European Monographs (Amsterdam, Kluwer Law International, 1996) 276. 98 See, eg, Case C-163/90 Legros and Others [1992] ECR I-4625; Joined Cases C-363/93 and C-407/93 to C-411/93 Lancry and Others [1994] ECR I-3957, Joined Cases C-485/93 and C-486/93 Simitzi [1995] ECR I-2655; and Case C-281/98 Angonese [2000] ECR I-4139. 99 Case C-60/00 Carpenter [2002] ECR I-6279; Case C-1/05 Jia [2007] ECR I-1. 100 See Case C-224/98 D’Hoop [2002] ECR I-6191. See also Case C-520/04; Turpeinen [2006] ECR I-10685, Joined Cases C-11/06 and C-12/06 Morgan and Bucher [2007] ECR I-9161; Case C-353/06 Grunkin-Paul [2008] ECR I-7639. 101 Schempp, above n 94. 102 See Case C-135/08 Rottmann (ECJ, 2 March 2010); Ruiz Zambrano, above n 10; Case C-434/09 McCarthy (ECJ, 5 May 2011). 103 Garcia Avello, above n 94.

EU Federalism in 3-D 33 Belgium—decided to follow Spanish law when naming their son and daughter; they used the first surname of the father followed by the first surname of the mother (‘Garcia Weber’). However, their application to register the children under that name was rejected by the Belgian Registrar for Births, Marriages and Deaths, on the ground that, in Belgium, children bear their father’s surname. Before the ECJ, Belgium argued that there was no link with EU law as the situation of the children of Mr Garcia Avello was purely internal: they were two Belgian nationals residing in Belgium. However, the ECJ rejected that argument. First, it observed that the children of Mr Garcia Avello were also Spanish nationals lawfully residing in Belgium.104 Secondly, the ECJ held that Belgian law contravened Articles 18 and 20 TFEU by putting children with dual nationality at a disadvantage, given that they would have to cope with difficulties resulting from having two different names, both on the professional and on the private level, in using, in one Member State of which they were nationals, the documents or diplomas obtained in another Member State of which they were also nationals.105 In the same way, in Zhu and Chen,106 the ECJ ruled that Article 21 TFEU precluded British authorities from deporting Mrs Chen, a Chinese citizen, who was the mother of Catherine Zhu, an Irish infant, and who had sufficient resources to support herself and her daughter. The fact that Catherine—who had acquired Irish nationality as a result of being born in Northern Ireland—had never left the UK had no bearing on the ECJ’s findings.107

B. Focusing on ‘the Status of Citizen of the Union’: Rottmann Rottmann is a landmark judgment which opens a new stage in the law on EU citizenship. In that case, the ECJ focuses on the ‘status of citizen of the Union’ rather than on the existence of elements demonstrating a cross-border situation. The facts of the case are as follows. Whilst being the subject of judicial investigations in Austria, Dr Rottmann, an Austrian national, moved to Germany in 1995. Two years later, Austria issued an arrest warrant against him. In February 1999, he acquired German nationality by naturalisation, which meant losing simultaneously his Austrian nationality. However, in August 1999, Austria informed Germany of the arrest warrant issued against Dr Rottmann. Taking the view that, 104 Ibid paras 27–28, where the ECJ pointed out that the fact that the children also had Belgian nationality was irrelevant. In any event, the ECJ indicated that Belgium could not deny recognition of their Spanish nationality ‘by imposing additional requirements, with a view to the exercise of fundamental freedoms provided for in the Treaty’. See Case C-369/90 Micheletti and Others [1992] ECR I-4239, para 10. 105 Garcia Avello, above n 94, para 36. 106 Case C-200/02 Zhu and Chen [2004] ECR I-9925. 107 Ibid para 19 (holding that ‘[t]he situation of a national of a Member State who was born in the host Member State and has not made use of the right to freedom of movement cannot, for that reason alone, be assimilated to a purely internal situation, thereby depriving that national of the benefit in the host Member State of the provisions of [EU] law on freedom of movement and of residence’).

34 Koen Lenaerts by withholding that information, Dr Rottmann had obtained German nationality by deception, Germany revoked that nationality, and since the original nationality did not revive, Dr Rottmann became stateless. Dr Rottmann challenged that decision before the German courts. In essence, the referring court asked the ECJ the following question: Is it contrary to Article 20 TFEU for a Member State to withdraw from a citizen of the Union the nationality of that State, acquired by naturalisation and obtained by deception, inasmuch as that withdrawal deprived the person concerned of the status of citizen of the Union, and of the benefit of the rights attaching thereto, by rendering him stateless, acquisition of that nationality having caused that person to lose the nationality of his Member State of origin? At the outset, several Member States posited that the rules on the acquisition and loss of nationality do not fall within the scope of application of EU law, as those rules remain within the realm of national sovereignty.108 However, the ECJ dismissed that argument, recalling that the powers retained by the Member States may be circumscribed by the substantive law of the EU. For the case at hand, this meant that the Member States must, when exercising their powers in the sphere of nationality, have due regard to EU law.109 Moreover, Germany and Austria urged the ECJ to declare the preliminary reference inadmissible as all the elements of the case at hand were confined to a single Member State, namely, Germany. At the time the contested decision was adopted, Dr Rottmann was a German national, living in Germany, to whom an administrative act was addressed by a German authority.110 However, Advocate General Poiares Maduro opined that the ECJ should reject that objection of admissibility, focusing on ‘the origins of [Dr] Rottmann’s situation’.111 For the Advocate General, the exercise by [Dr] Rottmann of his right, as a citizen of the Union, to move and reside in another Member State had an impact on the change in his civil status: it was because he transferred his residence to Germany that he had been able to satisfy the conditions for acquiring German nationality, namely, lawful habitual residence within that country’s territory.112

Like the Advocate General, the ECJ dismissed the argument put forward by Germany and Austria. In so doing, however, it adopted a different approach. In contrast to Advocate General Poiares Maduro, the ECJ ‘disregard[ed Dr Rottmann’s] earlier move and look[ed] exclusively to the future effects that withdrawal of German citizenship would have by rendering [Dr] Rottmann stateless’.113 In the key passage of the judgment, the ECJ held that

108 109 110 111 112 113

Rottmann, above n 102, para 37. Ibid para 41. Ibid para 38. Ibid, Opinion of AG Poiares Maduro, para 11. Ibid para 13. Ruiz Zambrano, above n 10, Opinion of AG Sharpston, para 78.

EU Federalism in 3-D 35 [i]t is clear that the situation of a citizen of the Union who, like the applicant in the main proceedings, is faced with a decision withdrawing his naturalisation, adopted by the authorities of one Member State, and placing him, after he has lost the nationality of another Member State that he originally possessed, in a position capable of causing him to lose the status conferred by Article [20 TFEU] and the rights attaching thereto falls, by reason of its nature and its consequences, within the ambit of [EU] law.114

In reaching that conclusion, the ECJ stressed once again that ‘citizenship of the Union is intended to be the fundamental status of nationals of the Member States’.115 With a view to transforming this postulate into a living truth, the ECJ places weight on the status of citizen of the Union as such rather than on free movement. In light of Rottmann, even if there is no movement across a frontier, national measures which deprive an individual of his or her status of citizen of the Union, and thereby of the rights attaching to that status, fall within the scope of application of the Treaty provisions on EU citizenship.

C. A Ground-breaking Judgment: Ruiz Zambrano Rottmann left open the question whether the absence of a cross-border element prevents the application of Article 20 TFEU where a national measure does not deprive an individual of his or her status as an EU citizen but, in practical terms, produces the same effect. Stated simply, was the rationale underpinning Rottmann to be limited to national measures the application of which resulted in a legal deprivation of the status of citizen of the Union, or did it also extend to measures that deprived an EU citizen of the genuine enjoyment of the substance of the rights attaching to that status? In Ruiz Zambrano, the ECJ was confronted with that very question. In that case, the referring court asked, in essence, whether Mr Ruiz Zambrano, a Colombian national residing illegally in Belgium, could rely on the Treaty provisions on EU citizenship with a view to obtaining a derivative right of residence as the father of two Belgian children. By obtaining such a derivative right, Mr Ruiz Zambrano also sought to obtain a work permit to which, as an illegal immigrant, he was not entitled under Belgian law. The referring court also asked whether EU law opposes reverse discrimination, implicitly on the assumption that the situation of Mr Ruiz Zambrano was to be qualified as purely internal. The facts in Ruiz Zambrano may be distinguished from those in Zhu and Chen.116 Whilst both cases dealt with EU citizens who were infants, had never exercised their right to free movement and were dependent on a relative in the ascending line who was a national of a third country, only Ruiz Zambrano involved EU citizens residing in the Member State of which they were nationals. 114 115 116

Rottmann, above n 102, para 42. Ibid para 43. Zhu and Chen, above n 106.

36 Koen Lenaerts All intervening Member States and the Commission argued that the situation of Mr Ruiz Zambrano was one that was purely internal to Belgium: the son and daughter of Mr Ruiz Zambrano were two Belgian citizens who had never exercised their right to move. However, Advocate General Sharpston did not share that view. For her, cases like Garcia Avello, Zhu and Chen and Rottmann showed that the Treaty provisions on EU citizenship had been applied in spite of the fact that ‘the element of true movement was either barely discernable or frankly non-existent’.117 She also criticised the traditional approach of the ECJ by saying that ‘[l]ottery rather than logic would seem to govern the exercise of EU citizenship rights’.118 As to the question whether the situation of Mr Ruiz Zambrano was to be treated as a purely internal one, she urged the ECJ to read Rottmann in conjunction with its findings in Zhu and Chen. The Advocate General conceded that, unlike the contested decision in Rottmann, the Belgian legislation at issue did not deprive the son and daughter of Mr Ruiz Zambrano of their status as EU citizens. However, bearing in mind that the deportation of Mr Ruiz Zambrano to Colombia (or to any other third country) would require his son and daughter to leave the territory of the EU with him, Advocate General Sharpston argued that [t]hat would, in practical terms, place [such children] in a position capable of causing them to lose the status conferred [by their citizenship of the Union] and the rights attaching thereto. [Hence,] the children’s situation ‘falls, by reason of its nature and its consequences, within the ambit of EU law.’119

Just as with the situation of Catherine Zhu in Zhu and Chen, it would be impossible for the son and daughter of Mr Ruiz Zambrano to exercise their rights to move and reside in any Member State without the support of their father.120 Therefore, Advocate General Sharpston concluded that Article 20 TFEU was applicable to a situation such as that of Mr Ruiz Zambrano and his Belgian son and daughter. Moreover, she found that the Belgian legislation at issue was an interference with the rights of the children to move and reside in the EU, which could not be justified as it failed to comply with the principle of proportionality.121 The ECJ agreed with the Advocate General, ruling that Article 20 TFEU was applicable. It began by noting that Article 3 of the Citizens’ Rights Directive (the CRD),122 which defines the persons benefiting from the rights contained therein, was not applicable to the case at hand, since that provision applies to EU citizens ‘who move to or reside in a Member State other than that of which they are a national, and to their family members’.123 Next, the ECJ held that, since the son 117

Ruiz Zambrano, above n 10, Opinion of AG Sharpston, para 77. Ibid para 88. 119 Ibid para 95. 120 Ibid para 96. 121 Ibid paras 98 ff. 122 European Parliament and Council Directive 2004/38/EC on the right of citizens of the Union and their family members to move and reside freely within the territory of the Member States amending Regulation (EEC) 1612/68 and repealing Directives 64/221/EEC, 68/360/EEC, 72/194/EEC, 73/148/EEC, 75/34/EEC, 75/35/EEC, 90/364/EEC, 90/365/EEC and 93/96/EEC [2004] OJ L158/77. 123 Ruiz Zambrano, above n 10, para 39. 118

EU Federalism in 3-D 37 and daughter of Mr Ruiz Zambrano were Belgian nationals, Article 20 TFEU conferred the status of citizen of the Union on them.124 After stressing that ‘citizenship of the Union is intended to be the fundamental status of nationals of the Member States’,125 the ECJ ruled, in the key passage of the judgment, that ‘Article 20 TFEU precludes national measures which have the effect of depriving citizens of the Union of the genuine enjoyment of the substance of the rights conferred by virtue of their status as citizens of the Union’.126 For the case at hand, this meant that a refusal to grant a right of residence to a third country national with dependent minor children in the Member State where those children were nationals and resided, and also a refusal to grant such a person a work permit, had such an effect.127 Thus, Mr Ruiz Zambrano had a derivative right of residence in Belgium. Whilst it is too early to forecast the full impact of Ruiz Zambrano on EU citizenship, I should like, however, to make six observations. First, it follows from Ruiz Zambrano that Article 20 TFEU applies to EU citizens who reside in their home Member State but have not exercised their right to move, provided that the national measure at issue deprives that citizen of ‘the genuine enjoyment of the substance of the rights conferred by virtue of their status as citizens of the Union’ (in French: ‘privation de la jouissance effective de l’essentiel des droits conférés par leur statut de citoyen de l’Union’). Contrary to the traditional approach, Ruiz Zambrano stresses that a link with EU citizenship may exist in spite of the absence of physical movement across the border of the home Member State. Secondly, although the ECJ referred to ‘the genuine enjoyment of the substance of the rights conferred by virtue of ’ EU citizenship, I believe that the rationale underpinning Ruiz Zambrano also applies to national measures the effect of which is to deprive an EU citizen of the genuine enjoyment of the substance of some, but not all, of the rights attaching to that status. Indeed, for example, the national measure at issue in Ruiz Zambrano did not deprive Mr Ruiz Zambrano’s son and daughter of their right to seek diplomatic and consular protection in the territory of a third country as provided for by Article 23 TFEU. Thirdly, Ruiz Zambrano illustrates that EU citizenship not only operates under a pro-integrationist rhetoric, but also emulates the rationale under which human rights apply. In light of Ruiz Zambrano, any EU citizen is entitled to say ‘civis europeus sum’ [to all the Member States, including his or her own,] and to invoke [the Treaty provisions on EU citizenship] in order to oppose any [deprivation of the genuine enjoyment of the substance of the rights conferred by virtue of EU citizenship].128

124

Ibid para 40. Ibid para 41. 126 Ibid para 42 (referring to Rottmann, above n 102, para 42). 127 Ruiz Zambrano, above n 10, para 43. 128 Case C-168/91 Konstantinidis [1993] ECR I-1191, Opinion of AG Jacobs, para 46 (adapted as indicated in square brackets). 125

38 Koen Lenaerts Like human rights in democratic societies, the rights attaching to EU citizenship operate as a protective shadow that stays with EU citizens, regardless of whether they move or stand still. Hence, the fact that the deprivation of those rights takes place in a context lacking a cross-border dimension is not decisive. Fourthly, in contrast to its previous ruling in Rottmann and the Opinion of AG Sharpston, the ECJ held that Article 20 TFEU opposed a national measure such as that at issue in the main proceedings, without first determining whether it complied with the principle of proportionality. However, this does not mean that the new approach developed in Ruiz Zambrano makes no room for that principle. In my view, the reason why the national measure in question was not examined under the principle of proportionality lies in the fact that Belgium did not provide any justification as to the compatibility of that measure with Article 20 TFEU. Instead, it limited itself to arguing that Article 20 TFEU did not apply to the case at hand. Be that as it may, it seems very difficult for a national measure which causes the de facto loss of the status of citizen of the Union, to pass muster under the proportionality principle, given that ‘citizenship of the Union is intended to be the fundamental status of nationals of the Member States’. Fifthly, an a contrario interpretation of Ruiz Zambrano suggests that there is no link with EU law where there is no deprivation of ‘the genuine enjoyment of the substance of the rights conferred by virtue of [the] status as citizens of the Union’. Such an interpretation leaves open two different, albeit interconnected, questions. The first question is whether the expression ‘national measures which have the effect of depriving citizens of the Union of the genuine enjoyment of the substance of the rights conferred by virtue of their status as citizens of the Union’ is to be interpreted as a synonym of the expression ‘national measures which are liable to hinder or make less attractive the exercise of [rights attaching to the status of citizen of the Union] guaranteed by the Treaty’ (broad interpretation),129 or whether the expression ‘national measures which have the effect of depriving citizens of the Union of the genuine enjoyment of the substance of the rights conferred by virtue of their status as citizens of the Union’ refers to situations in which EU citizens have de facto lost a right attaching to the status of citizen of the Union (restrictive interpretation). As to the second question, it seems that the deprivation of a right attaching to the status of EU citizenship must relate to the ‘substance’ of those rights. Needless to say, a restrictive interpretation of the expression ‘national measures which have the effect of depriving citizens of the Union of the genuine enjoyment of the substance of the rights conferred by virtue of their status as citizens of the Union’ would render the term ‘substance’ redundant, as a de facto loss of a right attaching to the status of citizen of the Union would, by definition, affect its substance. Conversely, the term ‘substance’ would be of paramount importance if

129 That expression is commonly used by the ECJ in the context of the Treaty provisions on free movement. See, eg, Case C-19/92 Kraus [1993] ECR I-1663, para 32.

EU Federalism in 3-D 39 the expression ‘national measures which have the effect of depriving citizens of the Union of the genuine enjoyment of the substance of the rights conferred by virtue of their status as citizens of the Union’ were to be interpreted broadly. In accordance with such a broad interpretation, the ECJ will be called upon to draw the distinction between violations affecting the substance of a right attaching to EU citizenship and violations which do not. Last but not least, it is worth noting that, unlike Advocate General Sharpston, the ECJ did not address the issue of reverse discrimination. Perhaps, once it held that the situation of Mr Ruiz Zambrano was not purely internal, it reasoned that it was no longer necessary to determine the role played by reverse discrimination in the context of EU citizenship. However, one might also argue that the ECJ evaluated the two alternatives put forward by Advocate General Sharpston, choosing an approach based on the status of citizen of the Union over one that relied on the prohibition of reverse discrimination. It is true that in the case at hand both approaches led to the same solution. That will not always be so, however: reverse discrimination may arise even in situations where there is no deprivation for EU citizens of ‘the genuine enjoyment of the substance of the rights conferred by virtue of their status as citizens of the Union’.

D. The Sequel: McCarthy In McCarthy,130 the ECJ had for the first time the opportunity to apply its new approach to EU citizenship. In so doing, it clarified some of the issues that Ruiz Zambrano had left open, most notably the interaction between Articles 20 and 21 TFEU, and the interpretation of the expression ‘national measures which have the effect of depriving citizens of the Union of the genuine enjoyment of the substance of the rights conferred by virtue of their status as citizens of the Union’. The facts of the case may be summarised as follows. Mrs McCarthy, a dual Irish and UK national, was born and had always lived in the UK, ie she had never exercised her right of free movement. Mrs McCarthy married a Jamaican national who lacked leave to remain in the UK in accordance with that Member State’s immigration laws. In order to prevent his deportation, Mrs and Mr McCarthy applied to the Secretary of State for a residence permit and a residence document as, respectively, a Union citizen and the spouse of a Union citizen. However, their application was rejected on the ground that Mrs McCarthy was neither economically active nor self-sufficient, as she was a recipient of State benefits. The referring court asked, in essence, whether Article 3(1) of Directive 2004/38 or Article 21 TFEU [was] applicable to the situation of a Union citizen who [had] never exercised [her] right of free movement,

130

McCarthy, above n 102.

40 Koen Lenaerts who [had] always resided in a Member State of which [she was] a national and who [was] also a national of another Member State.

If so, the referring court also asked whether that Union citizen could be considered to be a legal resident for the purposes of Article 16 of Directive 2004/38, in spite of the fact that she did not satisfy the requirements of Article 7 thereof. The ECJ began by recalling that a Union citizen in a situation such as that of Mrs McCarthy is not covered by the concept of ‘beneficiary’ as provided for by Article 3(1) of Directive 2004/38. In addition to relying on the wording of that provision, the ECJ found that a teleological and contextual interpretation of Article 3(1) also led to that outcome. It noted that the residence of a person in the Member State of which he or she is a national cannot be made subject to conditions, since this would run counter to the principle of international law,131 recognised by EU law, which precludes a Member State from refusing its own nationals the right to enter its territory and remain there for any reason, or expelling its own nationals from its territory, or refusing their right to reside in that territory, or making such right conditional.132 Accordingly, since Directive 2004/38 sets out the conditions governing the exercise of the right to move and reside freely within the territory of the Member States, it cannot apply to Union citizens who enjoy an unconditional right of residence due to the fact that they reside in the Member State of which they are a national.133 As to the applicability of Article 21 TFEU, the ECJ held that EU law does not apply to situations ‘which have no factor linking them with situations governed by [EU] law and which are confined in all relevant respects within a single Member State’.134 However, the ECJ pointed out that from the fact that a Union citizen, like Mrs McCarthy, has not made use of her right of free movement, it does not follow that her situation is, for that reason alone, to be considered as purely internal.135 Quoting the key passage of Ruiz Zambrano, it held that ‘Article 20 TFEU precludes national measures which have the effect of depriving Union citizens of the genuine enjoyment of the substance of the rights conferred by virtue of that status’.136 Needless to say, among the rights attaching to the status of citizen of the Union, there is the right to move and reside freely within the territory of the Member States, which is enshrined in Article 21 TFEU. However, the ECJ held that no element of the situation of Mrs McCarthy, as described by the national court, indicates that the national measure at issue in the main proceedings has the effect of depriving her of the genuine enjoyment of the substance of the rights associated with her status

131 That principle is reaffirmed in Art 3 of Protocol No 4 to the European Convention for the Protection of Human Rights and Fundamental Freedoms, signed in Rome on 4 November 1950. 132 McCarthy, above n 102, para 29. 133 Ibid para 34. 134 Ibid para 45. 135 Ibid para 46. 136 Ibid para 47.

EU Federalism in 3-D 41 as a Union citizen, or of impeding the exercise of her right to move and reside freely within the territory of the Member States, in accordance with Article 21 TFEU.137

Indeed, the failure by UK authorities to take into account Mrs McCarthy’s Irish nationality had in no way affected her right to move and reside freely within the EU. Next, the ECJ went on to distinguish the facts of the case at hand from those in Ruiz Zambrano and Garcia Avello. In contrast to Ruiz Zambrano, the ECJ observed that the national measure at issue in the main proceedings did not have the effect of obliging Mrs McCarthy to leave the EU.138 As to Garcia Avello, the ECJ explained that what mattered in that case was not whether the discrepancy in surnames was the result of the dual nationality of the persons concerned, but the fact that that discrepancy was liable to cause serious inconvenience for the Union citizens concerned that constituted an obstacle to freedom of movement that might be justified only if it was based on objective considerations and was proportionate to the legitimate aim pursued.139 Stated differently, the ECJ ruled that dual nationality is not in itself a sufficient connecting factor with EU law.140 Accordingly, the situation of a person such as Mrs McCarthy had no factor linking it with any of the situations governed by EU law and was thus confined in all relevant respects within a single Member State. In accordance with the ruling of the ECJ in McCarthy, one might argue that a combined reading of Articles 20 and 21 TFEU suggests that, in order for a national measure to fall within the scope of EU law, the measure must produce either a ‘deprivation effect’ or an ‘impeding effect’. In my view, the ‘impeding effect’ refers to the traditional line of case law according to which the application of the Treaty provisions on EU citizenship requires the existence of a crossborder link. Thus, the ‘impeding effect’ requires a cross-border link, but it does not require the national measure in question to cause the loss, in practice, of the status of citizen of the Union. As Garcia Avello shows, it suffices that the national measure at issue is liable to cause ‘serious inconveniences’ to a right attaching to the status of citizen of the Union. By contrast, as Ruiz Zambrano made clear, the ‘deprivation effect’ does not depend on the existence of such a link but focuses on the rights attaching to the status of EU citizen. Or, in other words, the ‘deprivation effect’ does not require a cross-border link but requires the national measure to cause more than ‘serious inconveniences’. That effect requires a de facto loss of one of the rights attaching to the status of citizen of the Union. It follows from the foregoing that the ‘impeding’ and ‘deprivation’ effects are subject to different requirements which are not, however, mutually exclusive. It is still possible for a

137

Ibid para 49 (emphasis added). Ibid para 50. 139 Ibid para 52 (referring to Case C-353/06 Grunkin and Paul [2008] ECR I-7639, paras 23–24 and 29). 140 McCarthy, above n 102, para 54. 138

42 Koen Lenaerts national measure which applies in a cross-border context to cause the loss of the status of EU citizen, thus producing both types of effect. Furthermore, it is worth noting that the ECJ did not expressly refer to Zhu and Chen. However, that silence should not be interpreted as a sign of inconsistency. On the contrary, in my view, a close reading of Zhu and Chen suggests that the latter is actually consistent with McCarthy. In that case, one should recall that the application of the national measure in question would have caused a ‘deprivation effect’: just like the children of Mr Ruiz Zambrano, the deportation of Mrs Chen would have forced Catherine Zhu to leave the territory of the Union. The deportation of her mother would indeed have had the effect of depriving her of the genuine enjoyment of the substance of the rights associated with her status as a Union citizen. Hence, her Irish nationality provided a sufficient connecting factor with EU law, but not because she was an Irish national living in the UK but owing to the fact that her Irish nationality allowed her to benefit from the rights attaching to her status as an EU citizen. Accordingly, since the national measure at issue caused the de facto loss of a right attaching to her status as an EU citizen, namely her right to move, that measure fell within the scope of Article 21 TFEU.141 Moreover, it is true that in Ruiz Zambrano, the ECJ grounded the ‘deprivation effect’ in Article 20 TFEU instead of having recourse to Article 21 TFEU. However, given that Article 21 TFEU limits itself to giving expression to a right already laid down in Article 20(2)(a) TFEU, one might argue that Article 21 TFEU also opposes a national measure which has ‘the effect of depriving a Union citizen of the genuine enjoyment of the substance of [the right to move]’.142

E. Concluding Remarks In light of Rottmann, Ruiz Zambrano and McCarthy, the rights attaching to the status of citizen of the Union are self-standing, ie they may be relied upon even in situations lacking a ‘cross-border’ link. This means that EU citizens enjoy a catalogue of EU rights, as provided for in Article 20 TFEU, with which both the EU political institutions and the Member States must comply. Rottmann, Ruiz Zambrano and McCarthy may be examined from a vertical as well as from a horizontal perspective. Vertically, by divorcing the rights laid down in Article 20 TFEU from the rhetoric of free movement, the ECJ expanded the scope of application of those rights which now apply even in the absence of a cross-border element. By refusing to qualify the situation of Mr Ruiz Zambrano as ‘purely internal’ (ie a situation with no links to EU law), Belgian authorities

141 Today, an EU citizen in the same situation as that of Catherine Zhu would fall within the scope of Art 3(1) of Directive 2004/38, as that person, unlike Ms McCarthy, would have only the Irish nationality. Hence, he or she would be an Irish national challenging an administrative decision adopted by UK authorities. 142 Ruiz Zambrano, above n 10, para 42.

EU Federalism in 3-D 43 were bound to comply with EU law. Horizontally, the ECJ has the delicate task of interpreting the rights listed in Article 20 TFEU with due regard to the prerogatives of the EU legislator. That is why, in Ruiz Zambrano and in McCarthy, the ECJ first examined whether Directive 2004/38 was applicable before interpreting Articles 20 and 21 TFEU. Most importantly, from a diagonal perspective, the ECJ must be cautious when drawing the contours of the Treaty provisions on EU citizenship. Otherwise, it would be emulating the ‘incorporation doctrine’ of the United States Supreme Court.143 If the ECJ were to apply the rationale underpinning Ruiz Zambrano and, at the same time, to interpret Article 20 TFEU broadly so as to include new rights, this would clearly impinge upon the prerogatives of both the EU legislator and the Member States. That is why, in McCarthy, the ECJ examined thoroughly whether the national measures fell within the scope of Article 21 TFEU. That is also why the ECJ decided to draw the distinction between national measures producing a ‘deprivation effect’ and national measures producing an ‘impeding effect’. Accordingly, by limiting the rationale underpinning Ruiz Zambrano to the former type of national measures, the ECJ is actually seeking to strike the right balance between EU citizenship understood as ‘the fundamental status of nationals of the Member States’144 and the powers of the EU and national legislators.

V. CONCLUSION

As a conceptual framework of reference, federalism must be examined in the EU setting beyond the rhetoric of the nation-State. In my view, federalism should be understood as the balance of power between the search for ‘unity’ and the protection of ‘diversity’. The study of that balance of power should not be limited to the principle of conferral. In addition, the balance between ‘unity’ and ‘diversity’ should be examined from a multi-dimensional perspective. Vertically, one should look not only at the allocation of regulatory powers between the Union and the Member States, but also at how EU law affects the constitutional and political structure of the latter. For example, in section II., I attempted to show how the ECJ strives to accommodate EU rules on State aids to the special characteristics of Member States with a federal structure. Notably, the ECJ has adapted the concept of ‘selective aid measures’ to that category of Member States. Yet in order to avoid distortions caused by unfair competition that would eventually give rise to the fragmentation of the internal market, the adaptation of State aid rules to Member States with a federal structure is subject to strict conditions. Thus, the ECJ honours two conflicting principles, namely, the

143 See, in relation to fundamental rights, A Knook, ‘The Court, the Charter, and the Vertical Division of Powers in the European Union’ (2005) 42 CML Rev 367. 144 Ruiz Zambrano, above n 10, para 41.

44 Koen Lenaerts respect for the national identity of Member States with a federal structure and the effectiveness of EU State aid rules. Horizontally, it was submitted that the application of the balance of power between ‘unity’ and ‘diversity’ does not pertain to the exclusive province of the judiciary, but that the political process may also contribute to maintaining that balance in accordance with the intention of the authors of the Treaties. Additionally, as the IFI Case demonstrates, the ECJ can interpret EU law in such a way as to protect the political safeguards of federalism, namely, the right of veto. Diagonally, in light of Rottmann, Ruiz Zambrano and McCarthy, the application of the Treaty provisions on EU citizenship is not conditioned upon the existence of a cross-border element. The rights attaching to the status of citizen of the Union apply regardless of whether the citizen moves or stands still. Yet the ECJ limited the rationale underpinning those judgments to cases where the national measure in issue produces a ‘deprivation effect’. In so doing, the ECJ sought not to blur the contours of the Treaty provisions on EU citizenship. Otherwise, it would have encroached upon the prerogatives of the EU and national legislators. Hence, the ECJ sought to strike the right balance between the protection of individual rights granted by EU law and the powers of the EU and national legislators.

2 Federalism and Jurisdiction PAVLOS ELEFTHERIADIS*

I. INTRODUCTION

F

EDERALISM AS A political doctrine can mean many different things in different places. If you are in a strong unitary State, it will be de-centralising. If you are in a loose association of States, it will be the opposite. In the context of debates about the EU, where we start from international treaties, it generally means moving away from the legal framework of an international organisation and the gradual centralisation of power under some constitutional framework, with the corresponding loss of Member State competences. In one of the earliest discussions of federalism in the EU, Koen Lenaerts wrote as follows: The other dimension, the vertical division of powers between the Community and the Member States, was equally constitutionalized when the Court had to explain in the landmark Van Gend & Loos ruling of 1962 that although the EEC Treaty presents itself as a compact among sovereign States, it is in reality—due to its substance—a ‘constitution’ of a central legal order, federally related to the legal orders of the Member States.1

The term ‘federally’ is linked to the hierarchy introduced by the idea of the ‘new legal order’. The direct effect of the Treaties follows the logic of a central legal order imposing uniformity onto the Member States. Federalism in the EU generally calls for more federation, not less. But what is a federation?

II. THE QUESTION OF FEDERATION

European Union scholarship is rather ambiguous about federations and federal ideas. In a well-known article, Kalypso Nicolaïdis argued that the EU must be

* University Lecturer in Law and Fellow in Law, Mansfield College, University of Oxford. I am very grateful to the participants at the conference at Leuven in May 2011, where I presented an earlier version of this essay. I am grateful to all the participants for their comments and suggestions, and especially to Geert De Baere, Elke Cloots, Helder De Schutter, Triantafyllos Gouvas and George Pavlakos. 1 K Lenaerts, ‘Constitutionalism and the Many Faces of Federalism’ (1990) 38 American Journal of Comparative Law 205, 208.

46 Pavlos Eleftheriadis understood as a federal union, but not as a federal State.2 The key to her argument was that Europe is not supposed to be a single ‘demos’ but a community of different ‘demoi’, hence a ‘demoicracy’. She commented on the Constitutional Treaty then under discussion as follows: ‘[T]oday’s constitution does not call for a homogeneous community or for laws grounded on the will of a single European demos. Rather, it makes mutual respect for national identities and institutions one of its foremost principles.’3 The key here is that the institutional architecture does not centralise power; it allows the various parts to retain their independence. Is that federalism? Nicolaïdis relied heavily on the fact that under the Constitutional Treaty Member States had an EU law right to withdraw. She drew the conclusion that the Treaty ‘firmly establishes the EU as a federal union, rather than a federal state, which … is defined by the very denial of that right’.4 It is clear, I think, that Nicolaïdis’s conclusions are correct. Given that the option for all Member States to leave the Union is always open according to international law, we cannot call the structure one of statehood. All members remain sovereign States while at the same time creating some political institutions at the centre, such as a Court, a Commission and a Council. This will necessarily be the case for as long as the Union is based on treaties of public international law. So the EU is certainly a union of States, not a State itself. But can we call it ‘federal’? A federation has been defined by a leading political scientist as a compound polity combining constituent units and a general government, each possessing power delegated to it by the people through a constitution, each empowered to deal directly with the citizens in the exercise of a significant portion of its legislative, administrative, and taxing powers, and each directly elected by its citizens.5

In a confederation, by contrast, the institutions are not organised according to a single constitution, neither do they seek their legitimacy in the representation of a people. Instead, power is exercised by the delegates of the constituent units and is justified on the basis of their reciprocal agreements, not on the basis of an independent electoral base. So, ‘by contrast with federation, in which each government operates directly on the citizens, in confederations the direct relationship lies between the shared institutions and the governments of the Member States’.6 These are successful definitions that are widely shared among political scientists and lawyers. 2

K Nicolaïdis, ‘We, the Peoples of Europe’ (2004) 83 Foreign Affairs 97, 102. Ibid 102. 4 Ibid, 105. 5 RL Watts, ‘Federalism, Federal Political Systems, and Federation’ (1998) 1 Annual Review of Political Science 117, 121. For important essays on federalism and the EU, see K Nicolaïdis and R Howse (eds), The Federal Vision: Legitimacy and Levels of Governance in the United States and the European Union (Oxford, Oxford University Press, 2001); S Fabbrini (ed), Democracy and Federalism in the European Union and the United States: Exploring Post-National Governance (London, Routledge, 2005). 6 Watts, above n 5, 121. 3

Federalism and Jurisdiction 47 This account of federation is close to what Lenaerts was describing, ie, a constitution imposing order in a vertical fashion, as the Court of Justice has consistently said since the 1960s. It seems not to cover Nicolaïdis’s idea of a federal ‘union’, which sounds much closer to a ‘confederation’. It is obvious from a combination of the points made by Lenaerts and Nicolaïdis that the EU is somewhere in between the federation and a confederation. But where exactly? If we were political scientists, we could perhaps refrain from passing judgement. We might be able to say that it is a sui generis case. In due course its success and imitation elsewhere, or its unceremonious demise, would tell political scientists all they needed to know. But the issue is not in the main one of classification or of descriptive accuracy. We are not seeking to explain a political phenomenon. We are seeking to understand and interpret a constitutional arrangement of political power with real effects everywhere. Whether the EU is to be seen as federal or not is primarily an issue of political legitimacy, constitutional interpretation and legal doctrine. Our interpretation starts from the facts on the ground, but the ultimate task is interpretive.7 By choosing an appropriate intellectual framework, we are opting for a particular way of reading the existing law or outlining ethical duties, so that our legal deliberations about the free movement of goods and persons, competition law, the law of State aid or the protection of the environment will be affected. Thus a ‘federalist’ reading of direct effect, supremacy or citizenship will produce a different reading of the law from one that is ‘internationalist’. The questions before us are practical, not theoretical.

III. HISTORICAL FEDERALISM

Some theorists of the EU continue to see its future as one where the legitimacy of the EU will always be compromised until it adopts a single constitutional theory in the way of a federal State. This theory was most clearly expressed by Joschka Fischer, a former German Foreign Minister. In his famous speech in Berlin in May 2000, Fischer outlined a federalist vision for the ‘finality’ of European integration. The word ‘finality’ captures very well the reference to an historical aim. As is well known, Fischer’s institutional model is one where a new constitution would redraw the lines separating States and the Union, with the firm intention of strengthening the powers of the Union and giving them federal features. The background to Fischer’s institutional proposals is the following historical argument: ‘Quo vadis Europa?’ is the question posed once again by the history of our continent. And for many reasons, the answer Europeans will have to give, if they want to do well by

7 I discuss this process in detail in P Eleftheriadis, ‘The Idea of a European Constitution’ (2007) 27 Oxford Journal of Legal Studies 1.

48 Pavlos Eleftheriadis themselves and their children, can only be this: onwards to the completion of European integration. A step backwards, even just stand still or contentment with what has been achieved, would demand a fatal price of all EU Member States and of all those who want to become Members; it would demand a fatal price above all of our people. This is particularly true for Germany and the Germans.8

The problems that, for Fischer, the Union is called to solve are not just everyday political problems. They have to do with the ‘history of our continent’. Our task is nothing less than finding Europe’s place in modern history. Europe’s historical destiny is to be decided and debated as a whole, as a single problem. Fischer concluded that it is entirely clear that Europe will only be able to play its due role in global economic and political competition if we move forward courageously. The problems of the 21st century cannot be solved with the fears and formulae of the 19th and 20th centuries.

Again, the question of European integration is placed in a long-term historical context with the particular ‘role’ of Europe in global economic and political competition. Fischer’s federalism is historical in nature. A similar narrative informs Jürgen Habermas’s arguments on the desirability and status of the European Constitution. In his 2001 essay ‘Why Europe Needs a Constitution’, Habermas outlined an argument for a European Constitution that is partly historical and partly political.9 The main objective, for Habermas, was to maintain the specific form of social solidarity achieved in Europe, against the pressures of globalisation and ‘neoliberalism’. He wrote as follows: The question therefore is: can any of our small or medium, entangled and accommodating nation-states preserve a separate capacity to escape enforced assimilation to the social model now imposed by the predominant global economic regime? This model is informed by an anthropological image of ‘man’ as rational chooser and entrepreneur, exploiting his or her labour-power; by a moral view of society that accepts growing cleavages and exclusions; and by a political doctrine that trades a shrinking scope of democracy for freedoms of the market. These are the building blocks of a neo-liberal vision that does not sit well with the kind of normative self-understanding so far prevalent across Europe as a whole.10

The argument, then, is that in order to protect their achievements against the forces of globalisation and seek a ‘re-regulation’ of global economy, European nations ‘have a reason for building a stronger Union with greater international influence’.11 This means not only stronger internal institutions, but also the capacity

8 See J Fischer, ‘Vom Staatenverbund zur Föderation—Gedanken über die Finalität der europäischen Integration’ in Y Mény, C Joerges and J Weiler (eds), What Kind of Constitution for what Kind of Polity? Responses to Joschka Fischer (Florence, Robert Schuman Centre, 2000). 9 J Habermas, ‘Why Europe Needs a Constitution’ (2001) 11 New Left Review 5. For further arguments in the same direction, see J Habermas, The Inclusion of the Other in P de Greiff and C Cronin (eds), Studies in Political Theory (Cambridge, Mass, MIT Press, 1998). 10 Habermas, ‘Why Europe Needs a Constitution’, above n 9, 12. 11 Ibid.

Federalism and Jurisdiction 49 to act in the international scene: ‘With a view to the future of a highly stratified world society, we Europeans have a legitimate interest in getting our voice heard in an international context that is at present dominated by a vision quite different from ours.’12 Both for social cohesion and for articulating an international voice we need to promote closer European integration through a new constitution. But what are the preconditions for closer integration through a new constitution? It is clear that we cannot create society afresh by fiat. Habermas concedes that any closer institutional integration in Europe needs to build on a common consciousness or a minimum of cultural and political cohesion. Without a pan-European ‘civic solidarity’ the project will fail, since it will not be endorsed by the public it is intended to serve. What is needed is that European citizens share, therefore, something like a public sphere of a vibrant democracy: ‘There will be no remedy for the legitimation deficit, however, without a European-wide public sphere—a network that gives citizens of all Member States an equal opportunity to take part in an encompassing process of focused political communication.’13 This public sphere, and the civic solidarity it brings about, support State institutions but are also created by them in a ‘circular process’. This is the key to Habermas’s argument for a new constitution. History tells us that constitutions can act as catalysts in the creation of nations. Habermas notes that national consciousness emerged as much from the mass communication of formally educated readers as from the mobilisation of enfranchised voters and drafted soldiers. It has been shaped as much by the intellectual construction of national histories as by the discourse of competing parties, struggling for political power.

This is the argument for the constitution. It is a catalyst for creating the kind of social and political preconditions that will make Europe stronger. It is not an end in itself, but a means for defending Europe against globalisation and asserting its voice in the international scene. So Habermas says that ‘the European project can be seen as a common attempt by the national governments to recover in Brussels something of the capacity for intervention that they have lost at home’.14 Europe, therefore, ought positively to promote and, if necessary, actively engineer a Continental public culture and a European public sphere through both institutional and cultural measures. What are we to make of this argument? There are some obvious difficulties with it. First, why is social democracy the privileged identity of Europe? Are other political positions any less ‘European’? Secondly, how can a constitution be a historical catalyst? And thirdly, what is to be lost by the creation of a panEuropean public sphere? This is clearly a reversal of the argument made by Larry

12 13 14

Ibid. Ibid 17. Ibid 14.

50 Pavlos Eleftheriadis Siedentop in Democracy in Europe.15 Siedentop noted that the preconditions for a genuine democracy—the kinds of things that de Tocqueville noticed in his study of America—are lacking in Europe: there is no common political tradition, no common language, no particular identification with other communities, no mutual trust. Siedentop’s conclusion is that we should hold up the federalist project, at least until the conditions are closer to being met. Habermas’s argument is the reverse: we must proceed with federalism precisely because we need to accelerate the process of ‘creating a public sphere’ at the European level. A common public sphere is the result of constitution-building. The argument is instrumental. We impose the constitutional model for the union, in order to accelerate the formation of an adequate European political identity. It makes no difference for this argument that the structure of the Treaties as they stand may contradict the strong federalist model. But such an argument cannot work in law. Any legal and constitutional interpretation must stay loyal to the materials as they stand. Habermas’s argument, whatever its merits, says nothing to constitutional law. It is a call for political action, not an argument for constitutional analysis. Moreover, Habermas’s call for political union ignores the important political values that each political society in each Member State has achieved for itself. Many of the smaller States in particular achieved their political integration after wars of liberation or civil wars or other important events. Their public life is informed by these common memories. It would be a great loss to public life if a common political structure were to replace these domestic constitutions. In a similar vein, John Rawls asked these questions in his letter to Philippe Van Parijs: One question the Europeans should ask themselves, if I may hazard a suggestion, is how far-reaching they want their union to be. It seems to me that much would be lost if the European union became a federal union like the United States. Here there is a common language of political discourse and a ready willingness to move from one state to another. Isn’t there a conflict between a large free and open market comprising all of Europe and the individual nation-states, each with its separate political and social institutions, historical memories, and forms and traditions of social policy. Surely these are great value to the citizens of these countries and give meaning to their life. The large open market including all of Europe is aim of the large banks and the capitalist business class whose main goal is simply larger profit. The idea of economic growth, onwards and upwards, with no specific end in sight, fits this class perfectly. If they speak about distribution, it is [al]most always in terms of trickle down. The long-term result of this—which we already have in the United States—is a civil society awash in a meaningless consumerism of some kind. I can’t believe that that is what you want.16

15

L Siedentop, Democracy in Europe (London, Penguin, 2000). J Rawls and P Van Parijs, ‘Three Letters on the Law of Peoples and the European Union’ (2003) Revue de philosophie économique 7. 16

Federalism and Jurisdiction 51 IV. CONSTITUTIONAL FEDERALISM

A different type of federalism keeps closer to law. For this type of argument federalism is a viable interpretation of the materials as they stand. Federalism advises us to interpret them in an appropriate light informed by certain political values. Walter Van Gerven, for example, has argued that the political principles according to which the EU ought to be organised are those of an ‘accountable executive government’, assuming that the EU already exercised executive government under a single constitutional framework.17 Although Van Gerven concedes that the EU is not a State and not on the way to becoming a federal State,18 he argues that the appropriate constitutional framework for the EU is one where the values of domestic democracy are strictly adhered to. In his concluding chapter Van Gerven compares parliamentary and presidential systems of government. He then considers which one of them may be the most appropriate form of government for the EU, as if the problems addressed by domestic and EU institutions were the same.19 He concluded that a European democratic government should encompass an elected parliament, a ‘multiparty consensus government’ relying upon the Commission, a Commission that is fully accountable to Parliament, a Council of Ministers and a European Council.20 The argument is political and forward-looking, but it is obvious that if the EU is to be understood under the lenses of democratic theory, the more democracy we have the better. But Van Gerven’s arguments suffer from one glaring omission. His list of appropriate institutions has no room for what is known as the ‘intergovernmental conference’. At present, this is the highest authority in EU law. It is the body that enjoys full powers to redraft the Treaties at will, and has the power to bring about these changes if it gathers the ratification of all the Member States according to public international law. If there is an EU constitution then the intergovernmental conference is its author. This is because the Union is a creature of public international law and the Treaties are subject to the will of the Member States. How can Van Gerven have missed the highest law-making authority in the EU? It is because there is no way of accommodating the process of creating treaties into any democratic framework. Treaty-making cannot be captured by any ‘federal’ analysis or any model of democracy. Van Gerven’s analysis has no room for the absolute equal standing of all Member States qua subjects of public international law (neither does he mention the other ways in which States have equal standing in the EU institutions).21 Such equality in treaty-making does not fit any possible

17 W Van Gerven, The European Union: A Polity of States and Peoples (Oxford, Hart Publishing, 2005). 18 Ibid 60. 19 Ibid 309–74. 20 Ibid 373–74. 21 See P Eleftheriadis, ‘The Standing of States in the European Union’ in N Tsagourias (ed), Transnational Constitutionalism: International and European Perspectives (Cambridge, Cambridge University Press, 2007).

52 Pavlos Eleftheriadis democratic theory or theory of executive ‘accountability’. When we include this fundamental institutional feature of the Union, the democracy model fails.

V. PLURALIST FEDERALISM

I now turn to the version of federalist federalism that I consider the most popular. This is the idea that some form of federalism is an appropriate constitutional interpretation of EU law while at the same time allowing for the equal standing of States that Van Gerven did not accommodate. This combines federalism and pluralism.22 How is this possible? Koen Lenaerts first spoke of it in his 1990 article on federalism: Devolutionary federalism … refers to a constitutional order that redistributes the powers of a previously unitary State among its component entities; these entities obtain an autonomous status within their fields of responsibility. The principal concern is to organize diversity in unity. The system will only be in balance when the shift from a single-sovereign towards a multi-sovereign structure is no longer resented as a merely centrifugal movement threatening national cohesion.23

Lenaerts used this idea to describe features of the European Communities that signalled disintegration rather than integration. He used federalism as a formal idea, which occasionally is deployed for integration and occasionally for disintegration. He concluded that European federalism is ‘dual federalism’ according to which the Court of Justice ‘pursues a constitutional model in which the legal orders of the Community and the Member States are as strictly separated as possible’.24 Rather than signalling the failure of federalism, for Lenaerts the existence of two or more ‘legal orders’ in the Union signalled just another variation for federalism. One of his conclusions was as follows: Federalism is present whenever a divided sovereignty is guaranteed by the national or supranational constitution and umpired by the supreme court of the common legal order. Both sovereigns enact laws with direct operation upon those to whom they are addressed. There is—in principle at least—no political oversight of one sovereign over the other. The normal enforcement mechanism of the constitutional balance is of a judicial nature and in case of conflict between validly enacted federal and componententity laws, the former prevail over the latter.25

22 On pluralism and the EU, see N Walker, ‘The Idea of Constitutional Pluralism’ (2002) 65 MLR 317; N MacCormick, Questioning Sovereignty: Law, State and Nation in the European Commonwealth (Oxford, Oxford University Press, 1999); N Krisch, Beyond Constitutionalism: The Pluralist Structure of Postnational Law (Oxford, Oxford University Press, 2010); P Eleftheriadis, ‘Pluralism and Integrity’ (2010) 23 Ratio Juris 365. 23 Lenaerts, above n 1, 206. 24 Ibid 260. 25 Ibid 263.

Federalism and Jurisdiction 53 Lenaerts’ solution to the problem faced by Van Gerven was therefore to say that the federation does not need political oversight, but only judicial oversight over the allocation of competences. The States do not owe accountability to any political institution; but their actions are subject to the judgment of the Court of Justice. There are many ‘legal orders’, but one dominant court. All that we need to assume is that the Court of Justice is the federal court whose judgments prevail over the national courts. Can this argument work? I do not think so. We saw above that Van Gerven’s constitutional federalism could not accommodate the international nature of the Treaties. It could not show that there are any accountability mechanisms or any other democratic tests for the intergovernmental conference. Because the diplomatic representatives appear as delegates of equal States under public international law, no common European institutions can test them. They are all separately accountable to their own constitutional arrangements (a framework often occurring in many aspects of the institutional law of the EU, such as, for example, the election of Members of the European Parliament). Their actions are subject only to public international law and domestic law. The EU is not even attempting to be democratic in the way it produces its fundamental laws. The same problem affects, in my view, the constitutional federalism of Lenaerts. There is no way of accommodating the public international law elements of the Union through the supposed jurisdiction of the Court of Justice. The amendment of treaties is something wholly beyond the control of the Court of Justice. This means that a major law-making power is something wholly outside what Lenaerts calls a ‘federal court’ and wholly outside the Union’s supposedly ‘federal’ constitutional law. But other institutional elements are also exempt. For example, the composition of the European Parliament is also determined exclusively by State law and adjudicated by State courts.26 The domestic compliance of the Member States’ institutions with EU law is also in principle outside the jurisdiction of the Court of Justice. The Court receives references for a preliminary ruling but cannot hear claims from persons against their States. When the Commission brings an action against a Member State it launches a judicial procedure that is in principle international. So the Court adjudicates between States and the Commission, but it is not a supreme European Court. The Court of Justice has very strong and important powers; but its civil procedure and scheme of jurisdiction have not replaced those of the Member States. However, if there is no central constitutional law and no central scheme of jurisdiction, there is no federation. VI. THE QUESTION OF JURISDICTION

Can there be a federation where there is no single scheme of jurisdiction? The argument offered by Lenaerts could perhaps be read in a different way. The 26

See, eg, Case C-208/03P Le Pen v European Parliament [2005] ECR I-6051.

54 Pavlos Eleftheriadis separate legal orders may well work under the jurisdiction of different courts. We may understand pluralism to refer not only to the constitutional frameworks, but also to the courts entrusted with their interpretation and application. This is perhaps what some authors refer to as ‘radical pluralism’. This would be a true innovation, in that it would have to change the generally accepted definition of a federation offered by Ronald Watts. Can there be plural systems of jurisdiction in a single federation? This would entail not only that there were distinct orders of courts, but also that their respective constitutional arrangements would remain independent from each other—as they are in the EU. In the United States, there is no doubt that the federal courts and the federal constitution enjoy supremacy over state courts. In the EU, as both Van Gerven and Lenaerts admit, the two systems work in parallel, without a single system of jurisdiction. The question we are asking is this: Can we accommodate this equal standing of the jurisdictions within a federal constitutional interpretation? I should like to argue now that this is constitutionally impossible. It is impossible not in the sense that we cannot have something like it in practice. A constitutional crisis or an anomaly can create such phenomena. In this sense the co-existence of two jurisdictions is historically possible. But it is impossible in the sense that it is rationally intolerable: it contradicts fundamental assumptions about constitutionalism. No amount of pluralism and no amount of ‘heterarchy’ can turn an international model to a federal one. And to say that the central concepts of domestic constitutional law apply to bodies working in the mode of international law can lead to paradoxical results. The problem is that the international model relies on a dualist scheme of jurisdiction, ie a division of labour between national and international, whereas a federal constitutional model relies on a single scheme of jurisdiction. The former is the model of the EU. The latter is the model for States. But why do we say that whenever we have independent jurisdictions we have an international model? I need to spend some time clarifying how we argue for this point. One way of arguing for the proposition that constitutional orders must have a single scheme of jurisdiction may be on the basis that a single jurisdiction has very good effects. It may be said that a single jurisdiction contributes to better outcomes in terms of well-being, substantive moral interests, the good life or some other suitable index of success. Such an economic argument can never succeed, however. If we approach the problem from the point of view of good outcomes, European institutions will at most be legitimate for a given period of time, when they are and to the extent they are ‘successful’, and illegitimate for another period of time, when they fall below a certain threshold of success. There is no stability to any economic justification of institutions. This does not fit our experience of the legitimacy and stability of institutions, national or international. We praise this inflexibility and formality of institutions, and we often explain this in terms of the virtues of the rule of law. Economic arguments cannot accommodate this fact. Another possible and popular answer is one based on ‘popular sovereignty’. This answer tells us that whatever political framework we endorse, it must express

Federalism and Jurisdiction 55 the ‘will of the people’, and the will of the people is expressed only through a single scheme of power and jurisdiction. Popular sovereignty, for this argument, needs a single system of jurisdiction for its adequate expression. Because we need a single scheme, we need to choose between States or the Union. This is the view recently advocated by the German Constitutional Court. The international treaties that brought about the Communities and now the single EU do not mark the exercise of proper ‘pouvoir constituant’. Hence, there is no ‘democratic legitimacy’ in Europe’s institutions. In its Lisbon judgment the Court said: ‘The citizens’ right to determine, in equality and freedom, public authority with regard to persons and subject-matters through elections and other votes is the fundamental element of the principle of democracy.’27 It elaborated the principle as follows: The Basic Law does not grant the bodies acting on behalf of Germany powers to abandon the right to self-determination of the German people in the form of Germany’s sovereignty under international law by joining a federal state. Due to the irrevocable transfer of sovereignty to a new subject of legitimisation that goes with it, this step is reserved to the directly declared will of the German people alone.28

Later, in paragraph 281, the point is made very clearly about the EU: ‘Also in their elaboration by the Treaty of Lisbon, no independent people’s sovereignty of the citizens of the Union in their entirety results from the competences of the European Union’. Here the question of legitimacy is a question of authorship. Because there is a German people and this people has brought about the German constitution, that constitution is legitimate. In paragraph 346, the Court applies the point to the EU itself and notices that there is no adequate European ‘people’. The citizenship of the Union is derived solely from the will of the Member States and does not constitute a people of the Union which would be competent to exercise self-determination as a legal entity giving itself a constitution.29 The Court concludes that there can be no European federation and no European jurisdiction. I do not think such an argument works. It has many flaws. I will here mention only the most important.30 If the constitution is legitimate because approved by the ‘will of the people’ then it is always unstable. What happens if the ‘people’ changes its mind—and how do we know? Stability is achieved for the German Constitutional Court by its dogmatic assertion that the people’s ‘will’ is somehow eternally encapsulated in the text of the constitution as read by the Federal

27

Re Ratification of the Treaty of Lisbon, 2 BvE 2/08, 30 June 2009, [2010] 3 CMLR 13, para 212. Ibid para 228. 29 Ibid para 346. 30 I say more about it in P Eleftheriadis, ‘The Moral Distinctiveness of the European Union’ 9 International Journal of Constitutional Law (2011) 695. For the German Court’s latest ‘sovereignty’ pronouncements, see the recent very important judgment in Re the Euro Bailout, BVerfG, 2 BvR 987/10, 2 BvR 1485/10, 2 BvR 1099/10, 7 September 2011; for comment, see P Eleftheriadis, ‘The Euro and the German Courts’ 128 Law Quarterly Review (2012) 216. 28

56 Pavlos Eleftheriadis Constitutional Court. This is an absurd proposition. The Federal Constitutional Court is aware of that instability, and in a way it welcomes it: From the continuing sovereignty of the people which is anchored in the Member States and from the circumstance that the states remain the masters of the Treaties, it follows—at any rate until the formal foundation of a European federal state and the change of the subject of democratic legitimisation which must be explicitly performed with it—that the Member States may not be deprived of the right to review adherence to the integration programme.31

But the very idea of a ‘continuing sovereignty’ upsets any thought of stable constitutional law, and even the very idea of ‘jurisdiction’. Whatever a court decides may then be subject to the ‘continuing sovereignty’ of the relevant people. When we join the ideas of sovereignty and constitutional law, sovereignty always wins. The German Constitutional Court’s constitutional theory seems to me entirely incoherent.32 In addition, the Court’s argument makes no room for international law. If we take its theory of popular sovereignty at face value then international law has no authority whatsoever. We are left with a simple national ‘monism’, where international law is not law at all. It follows that the Treaties have no authority whatsoever. The people, or ‘das Volk’, can lawfully ignore them. But then the theory proves too much. For these reasons, the question of jurisdiction is not to be addressed either by an economic argument, or by a theory of ‘popular sovereignty’. For different reasons, both arguments misunderstand the role of constitutional law and institutions, and offer us little on the way to a successful interpretation of EU law. Neither argument accounted for the stability, permanence and formality of political and judicial institutions. Constitutional interpretation must proceed from a different argument.

VII. JURISDICTION AND LEGITIMACY

We need a different argument. How can we account for the stability of institutions, especially against the background of changing economic and social conditions? It is obvious that the unity of jurisdiction and the stability of institutions go hand in hand. In order to address the question we must return to some philosophical fundamentals. Why do we need fixed institutions? Why not have perfectly fluid collective arrangements among equally empowered and constantly negotiating citizens?

31

Re Ratification of the Treaty of Lisbon, above n 27, para 334. The general reasons for this incoherence are explained in P Eleftheriadis, ‘Law and Sovereignty’ (2010) 29 Law and Philosophy 535. 32

Federalism and Jurisdiction 57 The answer is given by one of the first accounts of the nature of a constitutional order in Aristotle’s Politics. Having outlined the idea of the rule of law and the separation of powers, Aristotle gives the following advice regarding the preservation of States: In all well-attempered governments there is nothing which should be more jealously maintained than the spirit of obedience to law, more especially in small matters; for transgression creeps in unperceived and at last ruins the state, just as the constant recurrence of small expenses in time eats up a fortune.33

But why should we have fixed rules? Are they not certain always to get it wrong, being rigid generalisations that miss the particularity of the moment? Yes, this is precisely the point, says Aristotle. Laws will always be imperfect. They tend to miss the details of the particular circumstances.34 He writes, ‘the best man, then, must legislate, and laws must be passed, but these laws will have no authority when they miss the mark, though in all other cases retaining their authority.’35 But this, he will immediately argue, is one of the reasons for having laws. The reasoning is this: laws get it wrong in a particular way; individual decisions get it wrong in an entirely different and more damaging way. If we had a constant negotiating process, we should decide things on the basis of what appeared to the majority to be just and fair—or to be promoting its interests. Given what we know about human nature, any such decision would be considered to be wrong by many, and not least by the minority who voted against it. Both Plato and Aristotle have warned that political disputes of this kind threaten the very viability of a political community. Unless we have a framework for enlisting the support of the occasional losers, the very structure of decisionmaking is threatened.36 In the Republic, Socrates asks if a group of people could act together if they did not respect common ground-rules of conduct and communication in order to establish at least some basic trust. He concluded that without rules of this kind, any co-operation between them would fail. So Socrates asks: ‘Do you think that a city, an army, a band of robbers or thieves, or any other tribe with a common unjust purpose would be able to achieve it if they were unjust to each other?’37 The body would hardly exist without common rules that everyone

33 Aristotle, Politics, 1307b (B Jowett trans) in R McKeon (ed), The Basic Works of Aristotle (New York, The Modern Library, 2001) 1246. 34 Plato writes, eg: ‘[L]aw could never accurately embrace what is best and most just for all at the same time, and so prescribe what is best. For the dissimilarities between human beings and their actions, and the fact that practically nothing in human affairs ever remains stable, prevent any sort of expertise whatsoever from making any simple decision in any sphere that covers all cases and will last for all time.’ See Plato, Statesman, 294a (CJ Rowe trans) in JM Cooper (ed), Plato. Complete Works (Indianapolis, Ind, Hackett, 1997) 338. 35 Aristotle, Politics, 1286a, above n 33, 1200. 36 See B Williams, ‘Plato against the Immoralist’ in M Burnyeat (ed), The Sense of the Past: Essays in the History of Philosophy (Princeton, NJ, Princeton University Press, 2008) 97–107. 37 Plato, The Republic (GMA Grube trans, revised by CDC Reeve) in Cooper (ed), above n 34, R 351c.

58 Pavlos Eleftheriadis obeyed: ‘Injustice, Thrasymachus, causes civil war, hatred, and fighting among themselves, while justice brings friendship and a sense of common purpose. Isn’t that so?’38 Thrasymachus agrees and helps Socrates into this conclusion: Apparently, then, injustice has the power, first, to make whatever it arises in—whether it is a city, a family, an army, or anything else—incapable of achieving anything as a unit, because of the civil wars and differences it creates, and, second, it makes that unit an enemy to itself and to what is in every way its opposite, namely, justice.39

So a group of people trying to become a body but consisting of principled immoralists who set out only to serve their self-interest with utter disregard of what justice requires of them, will fail to constitute a parliament or an army, or even a band of robbers. They are enemies to themselves. Their incompatible ends will destroy them. Remember that here Plato is not speaking of virtue. He is speaking of principles of just conduct that can be formulated publicly. These public rules are required for a body to exist. Without the practise of justice, there is no collective agent. This is the basis for an argument for the value of the rule of law. Aristotle and Plato defend the rule of law not on the basis that it always secures justice, but on the basis that it secures legitimacy. The wrongness of our collective decisions may well not render them illegitimate if they have been taken according to constitutional principles acceptable by all, so in principle respecting equal citizenship on the basis of reciprocity. Thus it is not the justice of constitutions that grounds them, but their legitimacy. And it is a condition for legitimacy that it is based on a single constitutional scheme of institutions and courts, effectively a single scheme of jurisdiction. Why? Because without such a single scheme, we cannot guarantee reciprocity and cannot guarantee co-operation. Reciprocity can be credible only if two conditions are met. First, the institutions of decision-making are public and publically known to apply equally to all. Secondly, the persons who occupy the positions of official power ought to be known and accountable, and ultimately be trusted by those whose lives they are changing. They ought to be able to understand the particularities of everyone’s situation. Waldron explains this very well as a requirement of just political representation.40 The philosophical argument for the rule of law is not, therefore, simply an argument for law by a rule-book. Rather, it is a constitutional principle about the appropriate function of offices and institutions. The laws become agreeable in all their imperfection only if they are administered by officers appropriately trained in the law, suitably distant from any dispute but also close enough to understand the interests at hand. The personal virtues and qualities of these officers are therefore part of the ideal. They are essential, if the

38 39 40

Ibid R 351d. Ibid R 352a. J Waldron, ‘Representative Lawmaking’ (2009) 89 Boston University Law Review 335, 345.

Federalism and Jurisdiction 59 law is to be applied without passion or self-interest. Again Aristotle put this very effectively: For magistrates there must be—this is admitted; but then men say that to give authority to any one man when all are equal is unjust. Nay, there may be indeed be cases which the law seems unable to determine, but in such cases can a man? Nay, it will be replied, the law trains officers for this express purpose, and appoints them to determine matters which are left undecided by it, to the best of their judgment. Further, it permits them to make any amendment of the existing laws which experience suggests. Therefore he who bids the law rule may be deemed to bid God and Reason alone rule, but he who bids man rule adds an element of the beast; for desire is a wild beast, and passion perverts the minds of rulers, even when they are the best of men. The law is reason unaffected by desire.41

This is not just ‘ideal’ theory. The persons appointed can work only ‘to the best of their judgment’, implying that their judgment will not always be perfect. The institutional framework for decision-making is supposed to accommodate our human limitations, and indeed derives some of its force from human frailty and weakness of will. In Aristotle’s version, which has been repeated by numerous philosophers through the ages, the rule of law is not just a rule of rules. It is the conscious organisation of human life according to basic principles of fairness that allows us to have a collective life. What follows from these thoughts is that, for a constitutional order to exist, its citizens must hold themselves to be under a political obligation to obey it even where they disagree with its requirements. They must be able to say that the institutions, with all their expected imperfections and errors, are preferable to self-help; they are legitimate, even though they may be occasionally wrong or even unjust. Citizenship and legitimacy are therefore intimately related. It follows, I think, that a constitutional framework and a system of jurisdiction must be single and not double or plural. A plural scheme of decision-making and adjudication destroys the basis for legitimacy—even if it does not necessarily lead to unjust actions.

VIII. STATES OR THE UNION?

Lenaerts took the Court of Justice to have created a new legal order. I am not sure that the idea of a ‘new legal order’ helps us understand these legal processes, but I shall assume it does for the purposes of this argument. Let us assume that this new legal order may be seen as the basis of a constitutional scheme of jurisdiction. It does seem to compete with the similar order of States. But which order is legitimate, the State or the EU? Saying that there ought to be a single scheme to enable legitimacy to do its work, does not tell us which one of the two should be preferred.

41

Aristotle, above n 33, 1286b, 1202.

60 Pavlos Eleftheriadis Why does it have to be the State? Is it simply because this is now effective? Why not say that when the EU legal order and the national legal order conflict, there is a free choice about which one should be the single basis of legitimacy? This takes us back to familiar arguments in political philosophy. There must be for each person one scheme of government and jurisdiction that generates the duties of citizenship. If there were more than one, then the purpose of the State’s institutions in guiding action and achieving co-ordination would be frustrated. Given the pluralism of views about justice and the partiality we all show to our own causes, we must have in place a single system of institutions attempting to bring about justice. But which one? Jeremy Waldron has noted that there is a special tie with the place where we live and its laws. So he argues that legitimacy must be exclusive and salient in the following sense: What must be established is that there is a good reason to recognize this organization, as opposed to any rival organization, as the one to do justice in the given territory or with regard to the claims that are at issue. To the extent that such reasons exist, the organization is ‘legitimate’. Legitimacy, then, is an exclusive characteristic: only one organization may be legitimate with regard to a given set of claims or with regard to the issues of justice arising in a given territory.42

The way this answer applies to the EU is, I think, very simple. The Member States offer a more or less legitimate constitutional scheme. Their order is effective and it works. This should be enough to establish their claim to legitimacy. Remember, the test is legitimacy, not perfect justice or perfect efficiency. As long as the established order is legitimate, any further worry about additional benefits of a new federalist legal order is redundant. The question does not arise. We could of course bring about a new legal order, but this should happen through the currently legitimate (State) legal and legislative process.

IX. THE QUESTION OF A UNION OF PEOPLES

Most political theorists who have looked into the issue have pointed out that the EU is nowhere near the level of political community required for a democracy or a federal entity.43 As Professor Anand Menon has summarised the point: ‘Far from being some kind of state in the making, the Union is a tool of existing states.’44 The EU relies on States and not the other way round. Not everyone sees it this way, of course. The Court of Justice of the European Union has long spoken of a new legal order and of an autonomous constitutional

42

J Waldron, ‘Special Ties and Natural Duties’ (1993) 22 Philosophy and Public Affairs 3, 22. See Siedentop, above n 15; A Moravcsik, ‘In Defence of the Democratic Deficit: Reassessing Legitimacy in the European Union’ (2002) 40 Journal of Common Market Studies 603–24; A Milward, The European Rescue of the Nation State, 2nd edn (London, Routledge, 2004). 44 A Menon, Europe: The State of the Union (London, Atlantic Books, 2008) 213. 43

Federalism and Jurisdiction 61 order. Advocate General Poiares Maduro expressed some such view when he said that the EU Treaties have established a ‘new legal order’, beholden to, but distinct from the existing legal order of public international law. In other words, the Treaty has created a municipal legal order of trans-national dimensions, of which it forms the ‘basic constitutional charter’.45

But if the Treaty has created a ‘municipal’ legal order, it has, then, replaced the municipal legal orders of the Member States, or at least sits alongside them as a parallel system of law and courts. I have argued above, however, that such a ‘pluralist’ and ‘federalist’ view of the EU legal order is inconsistent with basic principles of constitutional law. It misunderstands the political and ethical requirement for a single scheme of civil and criminal jurisdiction. A set of political institutions may be legitimate for a given group of people in a given territory at a given time only if it is the only one in force, as Aristotle and Plato showed, and as Waldron has recently reminded us. If legitimacy is exclusive in this sense then EU law and State law cannot both be sources of legitimacy or available schemes of constitutional authority. This interpretation is much closer to EU law as it stands today. A careful look at the Treaties shows that they do not endorse a federalist or pluralist model. They carefully outline the division of labour between European law and institutions, and State law and institutions.46 They are best understood under a dualist framework, parallel to that which applies to international law. But, then, where is the authority of EU law based? Is it entirely dependent on States? If political obligation is based on the singleness of the political institutions that properly apply to us, then any international source of political obligation must be an impossibility. What happens to supremacy and direct effect if we go down this route? Is this not an equally unfamiliar result, denying the progress of EU law over the last 50 years? In my view this conclusion does not follow. The way to avoid the conclusion is by rejecting the German Constitutional Court’s national monism as equally mistaken. Most constitutional orders—and in the best view, the German legal order as well—organise their relations with the EU on a dualist basis. By being dualist, they are open to international law and the EU. By explicitly endorsing the EU Treaties as something with constitutional significance, the Member States of the EU are also cosmopolitan. It is part of their constitutional order that they accept that they owe legal and constitutional duties to all outsiders, including to those States with which they have entered into treaties of public international law and including all the persons who derive rights directly from the EU Treaties. So the authority of international and EU law is not simply decided by national authorities on an ad hoc basis; it is built into the same national scheme

45 Joined Cases C-402/05 P and C-415/05 P Yassin Abdullah Kadi v Council of the European Union and Commission of the European Communities [2008] ECR I-6351, Opinion of AG Poiares Maduro, para 21. 46 I have argued for that in some detail in P Eleftheriadis, ‘The Structure of European Union Law’ (2010) 12 Cambridge Yearbook of European Legal Studies 121.

62 Pavlos Eleftheriadis of constitutional principle. In a way, all EU obligations are now equally domestic obligations. European Union law has a double foundation: both the fact of the international treaties and the fact of the constitutional endorsement by appropriate domestic institutions and processes. In this sense, the EU is a Union of Peoples, with important effects in the constitutional orders of all Member States. Our remaining constitutional challenge is to develop such a view of the EU as a cosmopolitan union of peoples. I think that the best possible argument must take seriously Rawls’s warning about the future of the EU and the need to recognise the virtues of existing political institutions and the cultures that sustain them. I can only sketch a possible argument here.47 Everyone can understand the difficulties and challenges of creating institutions of justice. Because all persons in the world are in the same predicament, burdened by the same passions and aware of the fragility of any political order, all of them understand the difficulties in setting up a constitution. Aristotle and Plato speak eloquently about the individual virtues required for the excellence of any free State or ‘polis’, but the modern political philosophers are not unaware of it either. Michael Walzer, for example, has explained that the moral standing of States depends on the political communities that underpin them and on a idea of a ‘communal integrity’ which ‘derives its moral and political force from the rights of contemporary men and women to live as members of a historic community and to express their inherited culture through political forms worked out among themselves’.48 This moral standing of States has been explained by Kant through the metaphor of a tree with its own trunk and roots: [F]or a state is not (like the land on which it resides) a belonging (patrimonium). It is a society of human beings that no one other than itself can command or dispose of. Like a trunk, it has its own roots; and to annex it to another state as a graft is to do away with its existence as a moral person and to make a moral person into a thing, and so to contradict the idea for the original contract, apart from which no right over a people can be thought.49

The metaphor derives its force from Kant’s earlier statements that the citizens of a legitimate State have to exhibit the appropriate moral qualities in setting up institutions of the rule of law and due process. This entails that jurisdiction has an international dimension. Any theory of legitimate jurisdiction entails the recognition of reciprocal duties owed by all States

47 I develop these points more fully in P Eleftheriadis, ‘Citizenship and Obligation’ in J Dickson and P Eleftheriadis (eds), The Philosophical Foundations of European Union Law (Oxford, Oxford University Press, forthcoming). 48 M Walzer, ‘The Moral Standing of States: A Response to Four Critics’ (1980) 9 Philosophy and Public Affairs 209, 211. 49 I Kant, ‘Toward Perpetual Peace’ (1795) 8:344 in MJ Gregor (ed and trans), Practical Philosophy (Cambridge, Cambridge University Press, 1996) 318.

Federalism and Jurisdiction 63 to all persons, citizens or non-citizens. Kant refers to this as the cosmopolitan right to ‘conditions of universal hospitality’. Any foreigner can claim before any State official that he be treated with respect and without violence or mistreatment: What he can claim is not the right to be a guest (for this a special beneficent pact would be required, making him a member of the household for a certain time), but the right to visit; this right, to present oneself for society, belongs to all human beings by virtue of the right of possession in common of the earth’s surface on which, as a sphere, they cannot disperse infinitely, but must finally put up with being near one another; but originally no one had more right than another to be on a place on the earth.50

Wherever there is jurisdiction, there ought to be ‘hospitality’. The argument is not based on belonging to political community, but on the separate idea of the moral community between all the persons of the world. Our duties are to each other. This argument shows that any legitimate domestic jurisdiction needs common or agreed rules that deal with foreign jurisdictions and foreign nationals with mutual respect. It then follows that an egalitarian international law is based on the moral equality of all persons and not just States. These are very general principles, the details of which of course will have to be filled out, just like the details of an egalitarian constitution need to be filled out by constitutionmaking, legislation and adjudication. But they show how the justice of the domestic constitution and the framework of international law are elements of the same project of justice. We cannot achieve one without the other. So the constitutional interpretation of the law of the EU is continuous with the development of domestic constitutional law. The projects are entirely parallel. European Union law, just like international law, does not have, strictly speaking, its own jurisdiction. It does not compete with States as orders of public law, ie schemes of law application, adjudication and enforcement. And because EU law does not compete with any State legal order, it is able to coexist with all of them without contradiction.

X. CONCLUSION

If these arguments are correct, we do not need the idea of federation for an appropriate constitutional interpretation of EU law. Federal States constitute a single scheme of government and jurisdiction on the basis of a coherent set of constitutional principles. The EU is no such thing. The idea of a pluralist legal system or dual sovereignty is equally implausible. It follows also that what Nicolaïdis has called ‘federal union’ is not really federal at all. In her account, the union has no single scheme of government and no single framework of jurisdiction. This account is correct, but it suggests a model based on an international recognition 50

Ibid 329.

64 Pavlos Eleftheriadis of the EU by the constituent States. So a much better approach is to say that the EU is a union of peoples under an international framework of justice and legitimacy. To say that it is international is neither to challenge the EU’s moral standing nor to diminish its importance. Arguments to such effect entirely misunderstand the power and legitimacy of international institutions.

3 Federalism, the EU and International Law On the Possible (and Necessary) Role of Subsidiarity in Legitimate Multilevel Trade Governance ALEXIA HERWIG*

I. INTRODUCTION

F

EDERALISM HAS BEEN offered as a framework for analysis of the EU legal order, ie as a useful way for thinking about questions of who does what, where and how.1 Similar questions have been raised in international law.2 The principle of subsidiarity is amongst legal mechanisms that purport to answer the ‘who’ and ‘where’ questions. However, in spite of recognition that the federalist questions of ‘who’, ‘where’, ‘what’ and ‘how’ are useful to pose in multilevel governance systems, legal principles for the allocation of authority are much contested, and the principle of subsidiarity is no exception to this. Subsidiarity has critics, but also defenders. In this contribution, I want to engage with one sceptical view about the role of the principle of subsidiarity for deciding upon and reasoning about competences, and law’s legitimacy in multilevel governance contexts for integrated/liberalised markets in the EU and the world trade order. At the outset, I should say that I am sympathetic to at least parts of the diagnostic of the problem with the positivisation of the principle of subsidiarity. What I want to suggest in this contribution, though, is that one needs to distinguish a critique of the principle as enshrined in and applied through positive law from a critique of one amongst several possible constructions of the principle of subsidiarity, * Assistant Professor in Subsidiarity and Multilevel Governance, Faculty of Law, University of Antwerp. 1 K Nicolaïdis and R Howse (eds), The Federal Vision: Legitimacy and Levels of Governance in the United States and the European Union (New York, Oxford University Press, 2001). 2 See T Broude and Y Shany (eds), The Shifting Allocation of Authority in International Law: Considering Sovereignty, Supremacy and Subsidiarity (Oxford, Hart Publishing, 2008).

66 Alexia Herwig and finally from a critique of the best possible construction of the principle of subsidiarity. The last type of critique, I want to suggest, cannot be valid. Ultimately, the suggestion is, it is not the principle of subsidiarity that is problematic, for it is intertwined with our reasoning about law’s validity, but how it is given substance through application.

II. MEANINGS ASCRIBED TO SUBSIDIARITY

Many different meanings are ascribed to subsidiarity or the subsidiarity principle. It is useful to consider them briefly here in order to delimit the topic. In the EU, the subsidiarity principle reads: Under the principle of subsidiarity, in areas which do not fall within its exclusive competence, the Union shall act only if and in so far as the objectives of the proposed action cannot be sufficiently achieved by the Member States, either at central level or at regional and local level, but can rather, by reason of the scale or effects of the proposed action, be better achieved at Union level. …3

Gráinne de Búrca distinguishes between three altogether different meanings of subsidiarity, but more have been suggested. She understands ‘democratic subsidiarity’ as the attempt to bring decision-making as close to the citizen as possible.4 She also associates the protection of citizens’ rights with this notion of ‘democratic subsidiarity’.5 George Bermann in fact discusses the two (accountability and individual liberty) separately.6 Pursuant to this construction, the subsidiarity principle expresses a preference for localised decision-making. De Búrca also develops the notion of ‘executive subsidiarity’, according to which subsidiarity is about the protection of executive prerogatives.7 Pursuant to this understanding, subsidiarity is about limiting as far as possible the encroachment of the higher level upon the affairs of the lower level.8 This then gives ample room to lower levels of decision-making to pursue various differing goals, such as the protection of culture, traditions, civic virtues, market efficiency, etc, as long as they are able to achieve objectives sufficiently. The underlying goals served by executive subsidiarity may be various. Bermann points to the preservation of identities, the promotion of diversity and allowing sub-State level government

3

Art 5(3) TEU. G de Búrca, ‘Reappraising Subsidiarity’s Significance after Amsterdam’, Harvard Jean Monnet Working Paper no 7/99, 13. For a similar, democratic understanding of subsidiarity, see GA Bermann, ‘Taking Subsidiarity Seriously’ (1994) 94 Columbia Law Review 331, 339. 5 De Búrca, above n 4, 14. 6 Bermann, above n 4, 340–41. 7 De Búrca, above n 4, 14. This is echoed by Bermann, who considers that, amongst other values, the subsidiarity principle serves to protect the allocation of competences internal to EU Member States. See Bermann, above n 4, 342. 8 Bermann, above n 4, 342. 4

Federalism, the EU and International Law 67 of EU Member States to govern matters in light of the fact that they are not well represented in Union decision-making.9 ‘Procedural subsidiarity’ has also been suggested by commentators, including de Búrca. ‘Procedural subsidiarity’ is not about zero sum games in the allocation of regulatory authority (one wins, one loses). It is rather reflected in the form and content of directives that leave flexibility to Member States as to how to achieve the objectives set by the Union.10 According to Lazer and Mayer-Schoenberger’s understanding of the term, ‘[p]rocedural subsidiarity fosters a legislative centre in which the Community prescribes the goals and, with some flexibility, the means to achieve them. National law-makers find themselves in a Community-wide “regulatory competition”. ’11 For de Búrca, the use of particular legal instruments does not exhaust the meaning of ‘procedural subsidiarity’. Instead, the concept is seen to go to the question of whether legislative activity is needed, and what makes it appropriate at this level or a higher or lower level without limits to the levels involved.12 It thus places an onus of justification upon public institutions,13 a line that I shall also take in this contribution. Another construction of the subsidiarity principle may be termed ‘effectiveness subsidiarity’. In their discussion of the Amsterdam Protocol on the Application of the Principles of Subsidiarity and Proportionality, Lazer and Mayer-Schoenberger consider that the vision of subsidiarity set forth therein has two parts. The first part is concerned with substantive necessity (objectives not being sufficiently achieved), the second with effectiveness (objectives being better achieved by the Community/Union).14 Antonio Estella’s construction of the subsidiarity principle is quite close to this. For him, the subsidiarity principle is about absolute and comparative differences in effectiveness in the achievement of objectives. Thus, he sees the first part (objectives not being sufficiently achieved) as being about the lack of effectiveness in absolute terms.15 However, he points out, this alone cannot justify regulatory action by the Union.16 The Union must also show that its actions would bring some added value compared to Member State action.17 Lastly, there is ‘deliberative subsidiarity’. Some have submitted that the subsidiarity principle may be seen as a principle of democratic inclusion, as a reflection

9

Ibid 341. De Búrca, above n 4, 30–33; D Lazer and V Mayer-Schoenberger, ‘Blueprints for Change: Devolution and Subsidiarity in the United States and the European Union’ in Nicolaïdis and Howse (eds), above n 1, 139. 11 Lazer and Mayer-Schoenberger, above n 10, 139. 12 De Búrca, above n 4, 3, 6, 8, 13 and 19–22. 13 Ibid 8. 14 Lazer and Mayer-Schoenberger, above n 10, 134. They do not, however, subscribe to the view that this is how the subsidiarity principle ought to be construed. 15 A Estella, The EU Principle of Subsidiarity and its Critique (Oxford, Oxford University Press, 2002) 94. 16 Ibid. 17 Ibid. 10

68 Alexia Herwig of the principle that all those affected by a decision should have a say in it.18 Thus, whenever there are cross-border effects, decision-making should shift to the level that provides for decision-making procedures that are inclusive of the affected European interests. Since EU Member States do not allow non-nationals to vote for federal-level elections, this will then be the EU level by default. The word subsidiarity is derived from the Latin ‘subsidiarius’, ie reserve and support, which provides a good clue to its meaning in the current context.19 Under this interpretation, Union action is supposed to help Member States, but only in so far as help is needed. Since the integration of markets requires joint decisions and Member States cannot take these even though they do dispose of the implementing means, Union action is only needed to set objectives, making directives an appropriate regulatory tool. Based on this short overview of possible definitions, we can conclude that the principle has been accorded many meanings, none of which has so far emerged as convincing. At its most basic, subsidiarity implies a notion of lexical ordering: one comes first, then another with respect to the fulfilment of certain tasks. What changes, the dependent variable, is the substantive or procedural criterion with the help of which such ordering is to be achieved.

III. CRITIQUES OF SUBSIDIARITY

A. Fundamental Critiques of Subsidiarity Various critiques have been developed against the principle of subsidiarity. I shall mention some of them briefly here, in order to illustrate their difference from the critique by Robert Howse and Kalypso Nicolaïdis that I want to discuss in greater detail in this contribution. Gareth Davies criticises the subsidiarity principle as a vacuous masking principle.20 Since the only objectives to be achieved by an action are set as EU objectives, namely the establishment of a barrier-free EU market, the EU decision-making level ipso facto becomes the only appropriate one. Member States, acting alone, can never achieve harmonisation, and it is precisely the divergence in regulations which creates the problem for free circulation and undermines the EU objectives in the first place. Moreover, since almost anything is somehow linked to trade,

18

NW Barber, ‘The Limited Modesty of Subsidiarity’ (2005) 11 ELJ 308, 312 and 316–18. See in this connection the interesting notion of ‘mutual subsidiarity’ developed by Landy and Teles, pursuant to which national and EU decision-makers are to help each other perform their tasks. I think this captures part of the meaning of subsidiarity, but does not answer the question of who is to do what. M Landy and SM Teles, ‘Beyond Devolution: from Subsidiarity to Mutuality’ in Nicolaïdis and Howse (eds), above n 1, 414–15. See also K Nicolaïdis, ‘Conclusion’ in Nicolaïdis and Howse (eds), above n 1, 459–64. 20 G Davies, ‘Subsidiarity: The Wrong Idea, in the Wrong Place, at the Wrong Time’ (2006) 43 CML Rev 63, 68, 72 and 75–77. 19

Federalism, the EU and International Law 69 there are no real limits on the exercise of EU decision-making.21 According to this criticism, the subsidiarity principle is utterly unhelpful, and what is really at issue— whether or not EU objectives are worthwhile pursuing in light of conflicting objectives—is not tackled by it. According to Estella, the subsidiarity principle fails to achieve its countermajoritarian purpose because it is not a justiciable principle. With respect to the material criteria for subsidiarity, Estella argues that the further criteria spelled out in the Amsterdam Protocol on Subsidiarity and Proportionality—the existence of a transnational dimension and market distortion—are too indeterminate and ambiguous, and require political rather than legal reasoning.22 It was for fear of losing its legitimacy that the then European Court of Justice (ECJ) shied away from engaging in a material subsidiarity analysis that would have required politico-normative assessment rather than legal reasoning.23 This part of Estella’s critique needs to be considered together with the second part. According to Estella, there are two dimensions to legitimacy in federalism, one vertical (the subsidiarity principle), one horizontal (appropriate decisionmaking procedures). Estella criticises the subsidiarity principle as being exclusively about the vertical dimension because it poses a binary question—‘Should we act?’—which allows for only a ‘yes’ or ‘no’ answer.24 But, he maintains, as long as there is no true European demos, the more fundamental legitimacy question to be posed is how decision-making procedures ought to be designed to protect this plurality and diversity (the horizontal dimension of legitimacy); and according to Estella, only deliberative decision-making achieves this.25 Again, one would have to conclude that the subsidiarity principle is utterly unhelpful because it poses the wrong question. In contrast to these fundamental critiques of the subsidiarity principle, Howse and Nicolaïdis’s position accepts that there is something valuable about subsidiarity in the transnational context, but it objects that it is articulated in the wrong place. For that reason, their critique is particularly challenging to examine.

B. Howse and Nicolaïdis’s Critique Howse and Nicolaïdis’s critique is directed at trade governance through the World Trade Organisation (WTO). They argue that the subsidiarity principle as a legal principle is too constitutionalist, inasmuch as it is concerned with the right allocation and exercise of authority between global, regional, State and local institutions. They point out that issues are cross-cutting, and therefore invariably mix up

21 22 23 24 25

Ibid. Estella, above n 15, 80, 82, 96 and 159–66. Ibid 159–66. Ibid 177. Ibid 48–49.

70 Alexia Herwig levels of governance and require joint solutions.26 A fixed division of competences together with formal legal principles of supremacy and subsidiarity would not, according to Howse and Nicolaïdis, protect and promote democratic politics.27 Supremacy would create problems, since hands-tying via supranational constitutional law is undemocratic because of agency costs.28 Ex post monitoring of agents is difficult because of the secrecy of negotiations and information deficits on the part of the principals.29 Ex ante restrictions on the mandate of agents are difficult because they reduce the ability of agents to make concessions, and, I would add, can lead to a breakdown of negotiations which may not have been at all what the agents (or most of them) wanted.30 The de-democratisation stems from the fact that international rules with little or no democratic pedigree are then given supremacy over national—more or less—mundane law based on political processes there. But on the other hand, allocating the competences by default to the national level does not solve the democratic conundrum either, as Christian Joerges reminds us, because nation-States suffer from a democracy deficit of their own for their failure to include ‘the Other’ when their legislation produces trans-boundary impacts.31 In the EU context, Howse and Nicolaïdis argue, the principles of supremacy and subsidiarity should not be understood as an allocation of competences and vertical relations of governance. Rather, the principles of subsidiarity and supremacy should be understood as horizontal transfers of sovereignty, which basically function to bring the Member States into new relations to each other. Thus, supremacy serves to enhance the representation of foreigners inside each Member State, while subsidiarity serves to mitigate the impact of foreign unaccountable laws in another Member State.32 According to Howse and Nicolaïdis, ‘[w]e need to understand supremacy and subsidiarity as heuristics or boundary conditions for the dialogue between levels of governance, as a baseline for accountability for governance beyond the state.’33 In short, supremacy and subsidiarity are more about creating processes in which accountability becomes possible. The next step of Howse and Nicolaïdis’s argument is based on the premise that the legitimacy of governance today stems from democracy, constitutional 26 R Howse and K Nicolaïdis, ‘Democracy without Sovereignty: The Global Vocation of Political Ethics’ in Broude and Shany (eds), above n 2, 164. 27 Ibid 164. 28 Ibid 168. 29 Ibid 64. 30 A co-author and I have argued elsewhere that the unanimity rule in international law-making structurally advantages positions that seek to preserve the status quo and is, for that reason, not in keeping with the ideal of democratic equality. See A Herwig and T Hüller, ‘Zur normativen Legitimität der Welthandelsordnung’ in M Hilf and T Niebsch (eds), Perspektiven des internationalen Wirtschaftsrechts (Stuttgart, Boorberg Verlag, 2008) 133. 31 C Joerges, ‘Constitutionalism in Postnational Constellations: Contrasting Social Regulation in the EU and in the WTO’ in C Joerges and E-U Petersmann (eds), Constitutionalism, Multilevel Trade Governance and Social Regulation (Oxford, Hart Publishing, 2006) 494. 32 Howse and Nicolaïdis, ‘Democracy without Sovereignty’, above n 26, 164 and 166–67. 33 Ibid 167.

Federalism, the EU and International Law 71 settlement and technical expertise or effectiveness (getting things right, one might say), with democracy ensuring feedback links to reflexive participatory mechanisms to ensure law remains in line with what people want.34 Now, for the world trade order, Howse and Nicolaïdis claim, subsidiarity has become ‘colonised’ by reliance on technical expertise and supremacy in the form of attempts to constitutionalise WTO law as a set of higher law—both, mind you, disavowing space for ‘normal’ politics. These two sources of legitimacy are therefore in need of a third one—the politics of democracy—to make them responsive to people’s beliefs, values and expectations; but instituting global democracy would not be possible either.35 According to Howse and Nicloaïdis, neither a ‘legitimacy fix’ through a reallocation of competences with the help of the formal legal principles of subsidiarity and supremacy, nor ‘architectural’ approaches focused on decision-making structures and procedures will be enough.36 Such architectural fixes of designating and cabining off decision-making competences to levels of governance, and ensuring that they do not stray into other areas, fail to capture the interdependence of policy fields and the ongoing management of this interdependence.37 Further, whether or not which level of governance should decide on an issue is politically contested over time. An ‘architectural’ fix would stifle this political contestation, where what is needed is re-politicisation and the possibility for moving back and forth between centralisation and decentralisation within each policy area. Flexibility, not rules, might be the slogan here. As a solution, Howse and Nicolaïdis advocate a political ethic, focused on the behaviour of actors and the beliefs, values and expectations that inspire them.38 What is needed is ‘performance art rather than architecture’, or a focus away from the stage to the performance of the actors.39 And further, ‘political structures are simply shells, more or less conducive to such democratic ethos’.40 Ultimately, such a political ethos must measure itself against whether it conforms with the ideal of self-legislation and autonomy that underlies the whole idea of democracy and majoritarian decision-making.41 What is necessary is that those in the disagreeing minority can nevertheless accept the decision taken through majority rule and thus be legitimately coerced.42 In terms of remedies, Howse and Nicolaïdis first advocate inclusiveness, which may mean actual participation (such as amicus briefs in WTO dispute settlement) but also openness to other values, as the Appellate Body of the WTO displayed

34 35 36 37 38 39 40 41 42

Ibid 169–70. Ibid 171 (need for third source of legitimacy) and 168 (critique of global democracy). Ibid 171–72. Ibid 182. Ibid 172 and 181. Ibid 172 and 182. Ibid 185. Ibid 185–86. Ibid.

72 Alexia Herwig in its decisions in EC—Hormones, US—Shrimp and EC—Asbestos, and a constructive engagement with the views of civil society.43 Secondly and thirdly, there is a need for opportunities for review and revision, and for checks and balances, which may usefully arise from fragmentation, ie a lack of coherence of international legal norms and a spirit of contestation on the part of other international organisations.44 Lastly, compromise, mitigation and compensation may be used to make certain outcomes more acceptable to the losing minority.45

C. Howse and Nicolaïdis’s Reasoning There are several ways of interpreting the argument by Howse and Nicolaïdis. On the one hand, they might simply be saying that competences within multilevel governance constructs like the EU and the WTO cannot easily be allocated ab initio to this or that level of governance. Those constructs result in legitimate governance because issues typically cut across levels of governance, and excluding them definitely from the influence and considerations of a particular level of governance might not lead to an adequate solution of the problem. On this reading, one would need political deliberative processes first, in order to work out which norms of which level of governance and in which proportions should finally govern an issue. As a matter of applied theorising, I fully agree with this view. Within multilevel governance constructs, knowledge about all relevant considerations and their respective weights in a particular issue is often limited. Who could have known beforehand, for instance, that the regulation of genetically modified organisms (GMOs) would raise not only issues of food and environmental safety (and which are relevant considerations in the WTO Agreement on the Application of Sanitary and Phytosanitary Measures (‘the SPS Agreement’)), but also, intertwined with them, questions about access to seeds, developing country farming, consumer choice, and the coexistence of conventional and organic agriculture, which are not as such part of the SPS Agreement? On another interpretation, the reasoning behind Howse and Nicolaïdis’s rejection of subsidiarity and supremacy is a non-ideal theory critique of governance through the WTO. In political theory, some authors acknowledge that a conception of justice is dependent on certain empirical pre-conditions.46 In ideal theory, the designation of principles of social and political justice occurs where all the factual background conditions are in place for these principles to be realised in full (eg, people speak the same language, desire to live together in a polity with

43

Ibid. Ibid 187–89. 45 Ibid 190. 46 J Rawls, A Theory of Justice, revised edn (Oxford, Oxford University Press, 1999) 398–99 and 434–41. 44

Federalism, the EU and International Law 73 the others, are capable of moral reasoning). In non-ideal theory, these empirical preconditions are absent, and consequently the principles of justice are inapplicable to such a situation. Turning to multilevel economic governance, subsidiarity would be underinclusive and therefore undemocratic, in the sense that outside affected interests are not represented in national decision-making. And while higher-level decisionmaking might formally include all affected interests, it would be undemocratic because of the remoteness of decision-making, and substantially under-inclusive because many stakeholders cannot in fact participate due to knowledge and resource asymmetries. Against this backdrop, merely adjusting levels of governance through principles of subsidiarity and supremacy would indeed not remedy these underlying problems of democracy. However, Howse and Nicolaïdis’s contribution may also be read as scepticism as regards the ability of law to address problems legitimately. They repeatedly suggest that one should focus on promoting the right ethics or attitude, and allowing for political contestation unencumbered by formal decision-making rules to shape governance through the WTO. To this extent, I beg to differ from their analysis. What I should like to show in the following paragraphs is that an expanded understanding of subsidiarity is part of the validity of law, and that the ‘cures’ Howse and Nicolaïdis advocate must rely on law to structure them and ensure they function in the desired manner, and therefore presuppose subsidiarity. It is helpful to recall the legal definition of subsidiarity: Under the principle of subsidiarity, in areas which do not fall within its exclusive competence, the Union shall act only if and in so far as the objectives of the proposed action cannot be sufficiently achieved by the Member States, either at central level or at regional and local level, but can rather, by reason of the scale or effects of the proposed action, be better achieved at Union level. …47

What this in fact says is that decisions shall be taken at the level that will achieve the optimal decision, with there being a rebuttable presumption that this is the lower level of decision-making. I should like to suggest that what makes law valid is informed by such an expanded notion of subsidiarity. For instance, negative liberties, positivised as constitutional or human rights, may be understood as an expression of subsidiarity. They give the right holders immunities from governmental interference in respect of the ordering of their private affairs but also put limits on it. They express the judgment that people may desire to pursue different life plans, about whose specific desirability and justification reasonable people may reasonably disagree. For instance, one individual may want to become a minister in a church, another one an atheist social worker and yet another one a green CEO. Different people will hold different views, and for different reasons, on which of these life plans is worthier and it seems impossible to come up with a single socially shared 47

Art 5(3) TEU.

74 Alexia Herwig value that could provide an ordering of life plans. On this interpretation, the individual is deemed best placed to decide which life plan he or she wishes to pursue, as long as collective regulation is not required by reason of coercive externalities on others in the pursuit of their life plans. Reserving space for individual decision-making is consistent with the idea of autonomy. However, when externalities reduce individual choices in a coercive manner, only action (regulation) at the collective level can achieve autonomy precisely by reason of the scale or effect of the proposed action, since it makes possible accountability to those affected by the externality. Also, on a broadly Kantian account, justified legal coercion involves giving inter-subjectively valid reasons. If these cannot be given then the individual (ie, lower level) should decide and coercion, including legal coercion, is not appropriate.48 When might it be the case that legal coercion is justified by inter-subjective reasons? Others have suggested that this will occur when there has been a prior coercive act involving using others wilfully as a means to one’s ends, to which the law responds with actions enjoying inter-subjective authorisation.49 In this case, the higher level of governance (ie, the collective) has a claim to being involved, and any solution to the problem must involve reasons that are accepted as valid inter-subjectively. For present purposes, I shall assume that lack of deontological consideration, effectiveness in terms of limiting options and sufficient social relevance to take the matter out of the realm of private morality must be present for triggering collective action and burdens of justification. What objective is it, then, that supplies the criterion for judging which level of governance is best equipped to take a decision? On this Kantian account of subsidiarity, the objective to be achieved is to arrive at an outcome qua procedures that is consistent with our equal autonomy and capacity as moral agents. If my account is so far persuasive, several things must follow for the law. First, if the law is to be legitimate, the subsidiarity principle as I here elaborate it here must actually be present in the law, that is, the law must be able to provide for the possibility of identifying such coercive acts. Secondly, the principle must be 48 G Pavlakos and J Pauwelyn, ‘Principled Monism and the Normative Conception of Coercion under International Law’ in M Evans and P Koutrakos (eds), Beyond the Established Legal Orders: Policy Interconnections between the EU and the Rest of the World (Oxford, Hart Publishing, 2011) 318, 323–24, 326 and 329. 49 Ibid 318 and 323. But see ibid at 330, for a more restrictive definition of coercion as a trigger for legal regulation and reason-giving, which suggests that coercive acts must be effective enough to put the coercee into a situation of Zugzwang. Zugzwang comes from chess, where there is an obligation to move a piece each time, and refers to a situation where the player has no option but to move in a way that disadvantages herself. In their chapter of the cited volume, Pavlakos and Pauwelyn give the example of border controls and immigration restrictions that leave individuals from developing countries with the severely restricted choice of life in misery (and possibly danger) in their countries, or attempting illegal (and possibly dangerous) border-crossings. What Pavlakos and Pauwelyn here refer to is thus a sub-set of action-directing actions that use others as a means to one’s end. Pavlakos and Pauwelyn base themselves on AJ Julius’s conception of action-direction, articulated in AJ Julius, ‘Getting People to Do Things’, draft paper presented at a workshop on the Relevance of Coercion, Centre for Law and Cosmopolitan Values, University of Antwerp, April 2009.

Federalism, the EU and International Law 75 present all the way down, ie at the very abstract or general level and at the very specific level. On some level, Howse and Nicolaïdis share my understanding of subsidiarity because they claim that subsidiarity, understood here broadly as meaning that governance is legitimate if it is addressed by the appropriate people in the appropriate space of governance and in the appropriate way, can be helpful.50 Nevertheless, they reject the idea that there should be a formal legal principle of subsidiarity, which suggests that subsidiarity is an extra-legal principle. It is of course consistent to affirm the principle of subsidiarity and its relevance for law in the abstract and to criticise its actual application. However, I do not think that this is what Howse and Nicolaïdis do, because they display a more general scepticism towards the subsidiarity principle, and the solutions they advocate focus more on the behaviour of actors than on institutional and legal design.

D. Howse and Nicolaïdis’s Solutions Howse and Nicolaïdis’s cures for the multilevel governance problématique emphasise virtues over law. Clearly, a shift towards a more virtuous behaviour and attitude on the part of agents and decision-makers in multilevel economic governance is helpful towards realising a more legitimate system of governance. A just system of governance (or government) that is supported by a sense of justice by the people subject to it will be more stable and will have a greater chance of affecting human behaviour in positive ways than one in which people violate the rules on every possible occasion. In this sense, virtuous behaviour is an important facilitator en route to establishing a just, legitimate and feasible system of government; but it can neither take its place, nor be interpreted so demandingly that any kind of deviance defeats the feasibility of justice. Where I think the problem lies is that Howse and Nicolaïdis place (almost) exclusive faith in virtuous behaviour. How is such a personal ethic to guarantee that the agents in fact adopt virtuous behaviour? We need laws, institutes and procedures to coerce others if necessary, but this automatically raises the question of their justification and, as I have argued above, subsidiarity-type questions. The other problem I have with respect to ‘good-governance-throughcontestation’ type arguments,51 is that what emerges therefrom may not be a right outcome. As Howse and Nicolaïdis observe, ‘[o]utcomes that result from the exclusion per se of interests and values of individuals or groups from the 50

Howse and Nicolaïdis, above n 26, 185. Amongst these, I would count Krisch and—to a lesser extent—Teubner and Fischer Lescano. See N Krisch, Beyond Constitutionalism: The Pluralist Structure of Postnational Law (Oxford, Oxford University Press, 2010); N Krisch, ‘The Pluralism of Global Administrative Law’ (2006) 17 European Journal of International Law 1, 247; G Teubner and A Fischer-Lescano, ‘Regime-Collisions: The Vain Search for Unity in the Fragmentation of Global Law’ (2004) 25 Michigan Journal of International Law 999. 51

76 Alexia Herwig decision-making process are likely to be inconsistent with the democratic ideal of political equality’.52 Now, who is to guarantee that outcomes that result from unresolved contestation, fragmentation or regime collisions meet this standard if no procedures and standards for justification-giving, nor any rules protecting the position of those unable or less able to influence decisions, are put forth? Howse and Nicolaïdis argue that subsidiarity and supremacy should be about creating processes for accountability, but what they offer looks quite meagre. Is my vision too unifying, static and prone to a fetish about rules? I do not think so. First, I too would agree that there is a need for revision of even the question of whether an issue creates externalities or coerces others (ie, the very question subsidiarity tackles), and certainly of any actual law and its pursuit of intermediate objectives. But this can be done via institutional rules that put in place sunset clauses or establish committees to re-evaluate an issue in an objective, transparent and fair manner. Secondly, my argument, according to which externalities or coercion need to be resolved at that decision-making level that actually encapsulates those affected, is quite consistent with the proliferation of diversity, because it allows constituent units to agree to disagree as long as the reasons for doing so are intelligible to the other side. This, I would submit, is how we should understand the ‘unity in diversity’ motto of the EU.

IV. THE SUBSIDIARITY PRINCIPLE IN WTO AND EU LAW

A. WTO Law If the subsidiarity principle, as I understand it here, is necessarily intertwined with the validity of law, it must also be detectable in positive law. In the WTO context, I suggest that it may be found particularly in those norms that trigger the application of the higher-level law and set in motion (or could set in motion) processes of accountability. Thus, it is in norms such as national treatment, most-favoured nation treatment, export or domestic content contingency of subsidies or the specificity of subsidies, or the requirements of Articles XX GATT and XIV GATS and specifically here the assessment of alternative measures to the one at issue that the subsidiarity principle would manifest itself. In this section, I want briefly to consider some examples from WTO law that may be seen as reflective of my understanding of subsidiarity.53

52

Howse and Nicolaïdis, above n 26, 186. I use the three cases to illustrate my conception of coercion and to show that my understanding of subsidiarity can help us to make sense of these cases. The following discussion does not claim to be an analysis of all relevant cases, which would have to examine cases on the provisions mentioned in the main body of the text. 53

Federalism, the EU and International Law 77 One example is the Appellate Body decision on likeness in EC—Asbestos.54 The case concerned less favourable treatment of like products through domestic regulations in the EU. Canada, a producer of asbestos, alleged that the French ban on asbestos and asbestos-containing products violated Article III:4 of the GATT 1994 because asbestos was treated worse than its substitute products, which were not banned. The first element of Article III:4 requires that products be ‘like’. The majority of the Appellate Body in EC—Asbestos considered that the determination of likeness was fundamentally a determination about the nature and extent of a competitive relationship between imported and domestic products.55 The Appellate Body turned to the Border Tax Criteria for assessing likeness; these focus on the physical characteristics, consumer tastes, end-uses and the tariff classification of the products under consideration, possibly extending to crosselasticity of demand.56 It considered the Border Tax Criteria useful for structuring the analysis of likeness, but also found that the clear and established significant health risks of asbestos, compared to the much lesser risk of substitute products, were relevant considerations for the determination of physical characteristics and consumer tastes, and pointed to the unlikeness of the products.57 Weighing up all the Border Tax Criteria, the Appellate Body came to the conclusion that Canada had failed to show that asbestos and the substitute products were indeed like products.58 What is one to make of this decision? Surely, the French ban limited the options of Canadian asbestos producers who, under the ‘all-affected-interests’ principle, would have a claim to being included in reason-giving processes, for which supranational and international law may function as a substitute. What I should like to suggest is that the decision may be read as recognising that the French ban was not an instance of deliberate disregard of the interests of outsiders located in Canada. Given the clear health risks of asbestos, any deontological justification capable of being freely assented to by moral agents would have necessarily come to the conclusion that the protection of the life and health of the population enjoys primacy over the moral interest in realising economic gains. Because of the absence of coercion, as I define it here, there was no a legal case to be made. Secondly, consider the famous US—Shrimp decision. This concerned a US ban on shrimp fished in a manner that endangered sea turtles, and a requirement for producers to use nets with turtle-excluding devices59 in order to get access to the US market. The Appellate Body found the chapeau of Article XX GATT to be violated because the US did not allow shrimp fished with techniques comparable in effectiveness to the turtle-excluding devices to become certified, did not propose 54 European Communities—Measures Affecting Asbestos and Asbestos Containing Products, Appellate Body Report (adopted 5 April 2001) WT/DS135/AB/R. 55 Ibid para 99. 56 Ibid paras 101–02. 57 Ibid paras 113, 118 and 121. 58 Ibid para 141. 59 These are trap doors allowing the turtles to escape from the nets.

78 Alexia Herwig negotiations on shrimp conservation to all affected countries, offered longer phase-in terms to some but not to others, and failed to administer its certification system in a way that gave due process to outside foreign shrimp producers.60 In its assessment of whether the measure constituted unjustifiable discrimination in the sense of the chapeau, the Appellate Body noted that [p]erhaps the most conspicuous flaw in this measure’s application relates to its intended and actual coercive effect on the specific policy decisions made by foreign governments, Members of the WTO. Section 609, in its application, is, in effect, an economic embargo which requires all other exporting Members, if they wish to exercise their GATT rights, to adopt essentially the same policy (together with an approved enforcement program) as that applied to, and enforced on, United States domestic shrimp trawlers.61

The fact that the US did not allow for certification of producers using fishing methods that were prima facie similar in effectiveness, coupled with the lack of due process for contesting a decision and being given reasons for exclusion from the US market, is a perfect illustration of actions using others as a means to an end. The US did not recognise at all the moral interest in economic gains of foreign shrimp producers (because of the lack of due process), nor the fact that conditions and fishing techniques in the shrimp-exporting countries may have been different, and thereby deprived them in an important way of their autonomy as moral agents. Thus, the Appellate Body noted that [i]t may be quite acceptable for a government, in adopting and implementing a domestic policy, to adopt a single standard applicable to all its citizens throughout that country. However, it is not acceptable, in international trade relations, for one WTO Member to use an economic embargo to require other Members to adopt essentially the same comprehensive regulatory program, to achieve a certain policy goal, as that in force within that Member’s territory, without taking into consideration different conditions which may occur in the territories of those other Members.62

Further, the exclusion of many foreign shrimp producers made it easier for US shrimp producers to bear the cost of the turtle-excluding devices by preserving their ability to sell on the domestic market, and contributed to levelling the playing field vis-à-vis those foreign producers allowed to sell in the US, since these had to incur similar costs of introducing the turtle-excluding devices. In other words, the US was using the foreign shrimp producers as a means to an end. The Appellate Body may have been suggesting as much when it stated: The resulting situation is difficult to reconcile with the declared policy objective of protecting and conserving sea turtles. This suggests to us that this measure, in its application, is more concerned with effectively influencing WTO Members to adopt essentially the same comprehensive regulatory regime as that applied by the United States to its

60 United States—Import Prohibition of Certain Shrimp and Shrimp Products, Appellate Body Report (adopted 12 October 1998) WT/DS58/AB/R, paras 162–84. 61 Ibid para 161. 62 Ibid para 164.

Federalism, the EU and International Law 79 domestic shrimp trawlers, even though many of those Members may be differently situated. We believe that discrimination results not only when countries in which the same conditions prevail are differently treated, but also when the application of the measure at issue does not allow for any inquiry into the appropriateness of the regulatory program for the conditions prevailing in those exporting countries.63

The Appellate Body also faulted the US for not having attempted international negotiations on the conservation of sea turtles, emphasising that consensual, multilateral processes with a joint determination of whether turtle-excluding devices were suitable, marked out the equilibrium between rights and obligations of WTO members.64 We may interpret this as an affirmation that a process of inter-subjectively valid justification was required, and that only the higher level of decision-making encompassing those subject to the coercion could have guaranteed this. Thirdly, consider the Appellate Body decision in Korea—Beef. Amongst other measures, a requirement that stores retailing foreign beef display signs indicating that they sold foreign beef, combined with a dual retail system for domestic and imported beef, was challenged. The Panel had determined that the dual retail system ‘encourages the perception that imported and domestic beef are different, when they are in fact like products belonging to the same market’, which gave a competitive advantage to domestic beef ‘based on criteria not related to the products themselves’.65 The Appellate Body found that the sign requirement did not constitute less favourable treatment in the sense of Article III:4 GATT because adverse consumer reactions to the signs and the foreign beef were only incidental. It stated that ‘[c]ircumstances like limitation of “side-by-side” comparison and “encouragement” of consumer perceptions of “differences” may simply be incidental effects of the dual retail system without decisive implications for the issue of consistency with Article III:4’.66 It also considered that the encouragement of consumer perceptions brought about by the dual retail system did not necessarily imply a competitive disadvantage.67 In respect of the dual retail law, the Appellate Body noted that [w]e are aware that the dramatic reduction in numbers of retail outlets for imported beef followed from the decisions of individual retailers who could choose freely to sell the domestic product or the imported product. The legal necessity of making a choice was, however, imposed by the measure itself. The restricted nature of that choice should be noted. The choice given to the meat retailers was not an option between remaining with the pre-existing unified distribution set-up or going to a dual retail system … Thus, the reduction of access to normal retail channels is, in legal contemplation, the effect 63

Ibid para 165. Ibid para 170. 65 Korea—Measures Affecting Imports of Fresh, Chilled and Frozen Beef, Panel Report (adopted 10 January 2001) WT/DS161/R, WT/DS169/R, para 6.34. 66 Korea—Measures Affecting Imports of Fresh, Chilled and Frozen Beef, Appellate Body Report (adopted 10 January 2001) WT/DS161/AB/R, WT/DS169/AB/R, para 141. 67 Ibid para 141. 64

80 Alexia Herwig of that measure. In these circumstances, the intervention of some element of private choice does not relieve Korea of responsibility under the GATT 1994 for the resulting establishment of competitive conditions less favorable for the imported product than for the domestic product.68

The Appellate Body also noted that a dual retail system that is solely the result of choices by private firms is not unlawful.69 While there was no formal coercion to make a choice for domestic beef, there was coercion to make an either/or choice, and there may well have been economic incentives for retailers to choose domestic beef. The factual result of the Korean measure was thus a ‘freezing’ of the choice made because of sunk costs involved once a choice had been made. Now, it may well be that the Korean Government pursued self-interested motives of protecting the domestic Hanwoo beef market through its regulation against cheaper foreign competition. However, what the Appellate Body is saying is that the sign requirement itself is not limiting the options of foreign beef producers, since the consumer preferences already existed independently of the measure and were not positively shaped by it. There are two ways of reading this: either the measure is not coercive enough, since foreign beef producers remain free to make efforts to change consumer perception in their favour; or whatever effects are imposed on foreign beef producers do not attain the required social relevance to become a matter for legal regulation.70 Lastly, the recent panel report in US—Tuna is interesting. The case raised the question of whether a US Act, a Code of Federal Regulations and a ruling by a US federal appeals court prescribing the conditions under which tuna could be voluntarily labelled as dolphin-safe could be considered as a mandatory technical regulation in the sense of Annex 1.1 of the WTO’s Agreement on Technical Barriers to Trade (TBT Agreement), with there being strong consumer preferences for dolphin-safe tuna in the US market. The majority of the panel considered that the mandatory character could be established if the measures regulated in a binding and exclusive manner the conditions under which a product could be labelled as dolphin-free.71 Put negatively, the measures prohibit a product from being labelled as dolphin safe that does not meet the conditions. It considered this criterion to be met because the US measures were subject to enforcement, laid

68

Ibid para 146. Ibid para 149. 70 Of course, private preferences in the free market can also create problems of impermissible coercion, eg racist hiring preferences, and require legal regulation. However, these are quite different in that they leave the individuals optionless or with severely reduced options if the preference is widespread, and deprive them of important means of participating in social life and the ability to realise their life plans. This degree of option-reduction is not present in the case at hand. What is more, the idea of a market hinges essentially on the possibility that consumers may choose not to purchase products, and that producers may be deprived of revenue or even go out of business. Consumer selection as the driver behind competitive efficiency gains of firms is thus a constitutive element of market creation, with which WTO law is fundamentally concerned. 71 United States—Measures Concerning the Importation, Marketing and Sale of Tuna and Tuna Products, Panel Report (not yet adopted) WT/DS381/R, paras. 7.117, 7.130-3, 7.139. 69

Federalism, the EU and International Law 81 down specific conditions under which any claim about dolphin-protective fishing techniques could be made.72 Note that the operative reasoning of the panel speaks of ‘any claim’ whereas the measures only regulate labels or marks affixed to tuna products.73 It may thus be possible in the US to make claims about dolphin-safe fishing techniques through commercial communications other than labels or marks affixed to the product. What is more, any voluntary labelling standard will also lay down criteria for the award of a label that will inevitably exclude some (otherwise, there would be no point to label) as the dissenting panellist points out in the dissent in paragraph 7.151. It seems that the panel’s operative reasoning could also apply to voluntary standards, unless the decisive reason for why the US measures were considered mandatory is the fact that they lay down only one way of establishing dolphin-safety. But then one might ask how restrictive is restrictive enough, and from whose viewpoint, for there to be a mandatory requirement? Would, for instance, a scheme that allows three different ways of certification not be restrictive enough simply in virtue of numbers or by reason of the quality and reasonableness of the conditions? There then seems to be a danger that the panel is mixing up questions of possible justifications, relevant under Article 2.2 or 2.4 of the TBT Agreement, with questions of enforceability and bindingness. The dissenting opinion in contrast considers that only laws that require products to have certain characteristics as the condition for being allowed to be sold on the market would qualify as mandatory.74 On the issue of the de facto mandatory character, the dissent considers in essence that the consumer preferences that make it so that selling tuna whose impact on dolphins is uncertain or negative cannot be attributed to the US.75 From this short overview of WTO dispute settlement reports, we can see that they may be interpreted with the help of my conception of subsidiarity, and that it offers some analytical value. It therefore does not seem far-fetched to claim that, empirically speaking, the subsidiarity principle is already at least sometimes reflected in positive law.

B. Implications for the EU Principle of Subsidiarity An obvious consequence for EU law emanating from my discussion is that the subsidiarity principle is a legal principle that is implicated when the Court of Justice of European Union adjudicates the boundaries of Article 34 TFEU, when it reviews whether EU legislative measures have the correct legal basis, or when it decides whether or not the free movement provisions are applicable to the actions of non-State national entities. In addition, the subsidiarity principle must inform 72 73 74 75

Ibid paras 7.142–5. Ibid para 2.6. Ibid paras 7.152–4. Ibid paras 7.176–8.

82 Alexia Herwig decisions at the Treaty-making and legislative stage about whether the Union should enjoy certain competences by reason of cross-border coercive effects of sufficient social seriousness, and whether it should exercise them in a concrete case. Contrary to what Davies claims, this would allow for contestation over whether or not the Union objectives linked to the creation of the internal market should be pursued in a given case. If Union action is fundamentally supposed to be about securing the equal moral autonomy of agents, however, the procedures by which Union and Member State decisions are taken necessarily comes into focus. The subsidiarity principle must then link elements of vertical and horizontal—or perhaps better federal and procedural—legitimacy in one assessment. The two are part of the same overarching question addressed by subsidiarity. To this extent, I think Estella’s critique of the subsidiarity principle is founded upon a mistaken distinction. Where he is right, though, is that the subsidiarity principle as enshrined in Article 5(3) TEU may be interpreted so as to display a bias towards the status quo, because it leads to a binary decision on which level of governance should act. It permits an interpretation in the actual case of EU multilevel governance in the positive law that assumes that all is well if the level with the least degree of deficits in terms of autonomypreservation (compare the notion of objectives being better achieved above a certain threshold) gets to exercise a competence, regardless of the fact that this level could be improved further still through fairer decision-making procedures. What the subsidiarity principle, as I understand it here, would call for is a differentiated decision, which may very well indicate that the higher level of decision-making should act but identify deficits in procedures if they exist. Such an analysis is not at all unusual for courts to perform, even if they ultimately render a ‘yes’ or ‘no’ decision, but it could also be undertaken through political procedures such as the one envisaged for national parliaments under the Lisbon Treaty.

V. CONCLUSION

In this contribution, I have argued that the principle of subsidiarity is intertwined with the validity of law, and that law (and therefore subsidiarity) must be part of the solution to the problem of legitimate multilevel governance of international trade and EU law. However, the principle of subsidiarity provides only a framework of analysis for working out the problem of law’s validity, with the substantive and procedural standards governing an issue requiring further elaboration. In this sense, subsidiarity is only the beginning of the analysis, and an exclusive focus on subsidiarity to solve the legitimacy problématique of international governance would be misguided.

4 The Court of Justice as a Federal Constitutional Court: A Comparative Perspective MONICA CLAES* AND MAARTJE DE VISSER**

I. INTRODUCTION

[T] ‘

UCKED AWAY IN the fairytale Kingdom of Luxembourg,’ Eric Stein famously declared in 1980, the European Court of Justice had ‘wrought a federal type constitution for Europe’.1 The time is ripe to revisit this conceptualisation of the Court of Justice as a constitutional court in a federal polity. We do so by adopting a comparative perspective, asking what inspiration may be drawn from the way in which federal constitutional courts at the national level have been designed and function. To be clear, our intention is not to engage in yet another debate on whether or not the EU has a constitution, or whether the Union may be characterised as a federation. Such an endeavour would presume the thorny preliminary question of definitions—of ‘constitution’, ‘constitutional’, ‘federation’ to name but the most obvious concepts—either from a normative perspective, or based on empirical insights. Indeed, our aim is more modest and starts from a couple of assumptions which are necessary in order to make the comparison meaningful. These will be elaborated in section II. below. Section III. inquires into the composition, relevant procedures and caseload of several national constitutional courts, and section IV. considers how the Court of Justice fares in this respect. Section V. concludes.

* Professor of European and Comparative Constitutional Law, Maastricht University. This contribution was made possible thanks to the financial support provided by the European Research Council (ERC). ** Assistant Professor, Maastricht University. 1 E Stein, ‘Lawyers, Judges and the Making of a Transnational Constitution’ (1981) 75 American Journal of International Law 1.

84 Monica Claes and Maartje de Visser II. TWO ASSUMPTIONS

As indicated in the introduction, our examination of the Court of Justice in this contribution proceeds on the basis of two assumptions: the first concerning the federal character of the EU; the second relating to the constitution of the Union.

A. The Federal Character of the EU The Union does have a number of characteristics which allow it to be considered as belonging to the federal type. Federalism is traditionally associated with States, and with the constitutional relation between sub-State entities within a State.2 As such, in the European tradition of federalism, with its focus on indivisible sovereignty, the main objection against the use of the federal concept in the context of the EU does not relate so much to the federal element per se, but to the fact that the EU is not a State. However, there are as many flavours of federalism as there are federal systems. In that sense, it might be more helpful to speak of ‘archetypes’ of federal structures. The Union has a mixed or compound structure, combining international and national elements.3 This would be the typology of the EU as a federation in the American tradition, which allows for the recognition of federations based on a treaty.4 What is more, the EU has federal features, such as the division of competences between the central or European level and the Member State entities, which is laid down in a document that is beyond the reach of either level to change using ordinary legislative procedures. Accordingly, the Union can qualify as a federation under quite a few definitions of that concept and the EU may be—and has been— profitably analysed in federal terms.5 Our first assumption, then, is that the Union has a number of characteristics which allow it to be considered as belonging to the federal type. Of course, the federal character of the EU is contested. It has been explicitly rejected by the German Bundesverfassungsgericht.6 It is not mentioned in the Treaties, and one might even say it has been blacklisted in the political debate. The reason for this lies in the keen desire to avoid an implicit choice for an existing prototype, which may open wounds or result in protracted debates as 2

R Schütze, From Dual to Cooperative Federalism (Oxford, Oxford University Press, 2009) 69. Ibid 70. 4 Ibid. 5 See, eg M Burgess, Federalism and the EU: Building of Europe 1950–2000 (New York, Routledge, 2000); D McKay, Designing Europe: Comparative Lessons from the Federal Experience (Oxford, Oxford University Press, 2001); L Goldstein, Constituting Federal Sovereignty: The European Union in Comparative Context (Baltimore, Md, Johns Hopkins University Press, 2001); A Menon and MA Schain, Comparative Federalism: The European Union and the United States in Comparative Perspective (Oxford, Oxford University Press, 2006); S Dosenrode, Approaching the European Federation? (Aldershot, Ashgate, 2007); Schütze, above n 2. 6 89 BVerfG 155, decision of 12 October 1993 (Maastricht Urteil). 3

The Court of Justice as a Federal Constitutional Court 85 to the finalité of the European integration project. At the same time, it should be acknowledged that constitutional vagueness in this respect may have its merits: it allows for the accommodation of conflicting interests and a gradual fleshing out of the relations between the various levels, as, inter alia, the Spanish experience illustrates.7 In contrast, a federalist discourse would raise expectations about the nature of polity that it organises. Moreover, there are differences from existing mature federal systems. So, for instance, full-fledged federalism requires full discipline and compliance with the rules that have been commonly agreed upon in the constitution. At the end of the day, those rules are judicially enforced by the constitutional court, if political negotiation proves unsuccessful. True federalism thus seems incompatible with constitutional pluralism, understood as competing claims for ultimate authority.8 Federalism can cope with multi-level constitutionalism and constitutional plurality. Yet, ultimately, the sub-national constitutions, as well as the organs enforcing them, must comply with the federal constitution, which is considered by all levels as the supreme law of the land. The supremacy of the federal constitution is essential in a full-fledged federal system.9 While we acknowledge that European federalism is not perfect, we consider it appropriate and useful to analyse the EU in federal terms.

B. The Constitution of the EU Furthermore, we shall assume that it is appropriate to use the constitutional concept in the context of the EU. As with the concept of federalism, there are a number of ingredients in the concept which critics consider essential when it comes to the notion of ‘Constitution’, and which lead them to deny the constitutional character of the European Treaties. Yet, as with federalism, the critique ultimately boils down to the argument that the Union lacks the capacity of a State. There are various versions of this line of reasoning, which are well known and will not be rehearsed here. For some, only States can have constitutions.10 For others, constitutions presuppose a ‘We, the people’ and hence a demos. For yet others, an EU Constitution could be only a very restrictive and formalistic version of a constitution-type set of rules, because the Member States—the true sovereigns of the Union—already have a constitution, which still is the highest norm

7 J Garcia Roca, Autonomías territoriales y forma de Estado: Contribución al estudio de la forma territorial del Estado espanõl (Madrid, Libro, 1985). 8 On constitutional pluralism, see eg M Avbelj and J Komárek (eds), Constitutional Pluralism in the European Union and Beyond (Oxford, Hart Publishing, 2012). 9 This is why Von Bogdandy and Schill refer to the EU as a federative but not hierarchical structure; see A Von Bogdandy and S Schill, ‘Overcoming Absolute Primacy: Respect for National Identity under the Lisbon Treaty’ (2011) 48 CML Rev 1417. 10 But see A Arnull, C Barnard, M Dougan and E Spaventa, A Constitutional Order of States (Oxford, Hart Publishing, 2011).

86 Monica Claes and Maartje de Visser of their land, the expression of the self of the people and their will to be governed by and under that (national) Constitution. In addition, the Member States have quite explicitly abandoned the constitutional concept in relation to the EU, first as convened in the European Council11 and later also within their domestic sphere. The Member States stripped the Reform Treaty of its constitutional attributes (title, human rights bill, language such as ‘laws’ and ‘ministers’), so as to be able legitimately to proceed to ratification of the new old-style Treaty without a referendum. After all, the new Treaty was no longer ‘constitutional’ in nature. This rejection of the constitutional concept should not be taken lightly: after all, the Member States are the Herren der Verträge and are still crucial components of the ‘constituting power’ of the EU. Constitutions arguably are only truly constitutional if they are recognised as such, and continue to be accepted by those governed under it. To be sure, there certainly are differences from national constitutions. Moreover, the relationship between the EU Constitution and national constitutions is one of the most intricate legal questions of European integration. Yet, assuming a functional definition of the term, it can still be maintained that the Union may be considered a polity governed by a body of law that functions largely as the Constitution of the European Union, in a manner comparable to national constitutions in States.12 If one accepts these assumptions, it is appropriate to analyse whether the Court of Justice of the European Union (CJEU) may properly be seen as a federal constitutional court. This is what the following sections set out to do.

III. KEY FEATURES OF NATIONAL FEDERAL CONSTITUTIONAL COURTS

In this section, we take a closer look at national constitutional courts with a federal mandate to identify some of their core characteristics. Given the nature of the inquiry, there is only a limited number of courts which can feature in the comparison, only a limited number of EU Member States being federal States. Indeed, only Germany, Austria and Belgium consider themselves explicitly to be federal States and will therefore certainly be included in the examination below. Spain and Italy may not subscribe to the federal terminology, and in fact the use of such terminology is simply unacceptable to some of the constitutional actors in those countries. For our purposes, however, we adopt a broad definition of a federation, namely, ‘a vertical divided powers system based on a constitutional document which lies beyond the reach of any of the levels acting alone under the procedure for the adoption of ordinary legislation’. Using this approach, we shall

11

Presidency Conclusions of European Council of 23 June 2007, 11177/1/07 REV 1. See, eg Joined Cases C-402/05P and C-415/05P Yassin Abdullah Kadi and Al Barakaat International Foundation v Council and Commission [2008] ECR I-6351, Opinion of AG Poiares Maduro. 12

The Court of Justice as a Federal Constitutional Court 87 therefore also incorporate the Italian Corte costituzionale and the Spanish Tribunal Constitucional in our analysis.

A. Situational Logic and Composition At the risk of stating the obvious, it should first be noted that in the countries mentioned, the choice has been made to entrust the resolution of vertical competence disputes to a judicial organ. This is in fact a common approach and may be seen as an indication of the wish, or even the need, to approach questions regarding the demarcation of the precise ambit of competences of the sub-units and those of the central bodies as legal questions and, hence, to take them out of the realm of politics.13 Indeed, in Belgium the introduction of what is today called the constitutional court was prompted by the process of federalisation initiated in the 1970s.14 Constitutional courts with a federal mandate are always located at the central level. This may give rise to a perception of ‘bias’ in favour of the central level when it comes to vertical competence adjudication, with concomitant risks for the compliance with constitutional court judgments by the federated entities, should they dispute those judgments’ legitimacy. However, such a ‘bias’ may already be reflected in the constitution itself, so that the constitutional court is simply giving effect to the constitutional settlement, rather than imposing its own reading on ‘neutral’ provisions. For example, it has been argued in relation to the Austrian constitutional court that [t]he Court’s adjudication is perhaps rather more centralist than federalist, but this is also due to the rather centralistic concept of the Austrian Federal Constitution and to the extremely fragmented division of powers which in practice has a centripetal effect.15

It will be seen later that the composition of these courts and the design of the procedure for competence adjudication sometimes incorporate features that can fuel feelings of a perceived bias. It should further be mentioned that the constitutional courts examined here do not just occupy themselves with resolving competence disputes between the federation and federated entities. Their mandates are broader than that. Constitutional

13 See also M Shapiro, ‘The Success of Judicial Review and Democracy’ in M Shapiro and A Stone Sweet (eds), On Law, Politics, and Judicialization (Oxford, Oxford University Press, 2002). 14 D Reynders, Prévention et règlement des conflits. La genèse de la Cour d’arbitrage (Brussels, Dossier du CRISP, 1983); JP Lagasse and S Moureaux, La Cour d’Arbitrage: Juridiction Constitutionnelle: Commentaire de la Loi du 28 Juin 1983 (Brussels, Larcier, 1984). The Belgian constitutional court was initially known as the ‘Cour d’arbitrage’ (court of arbitration). Its name was changed in the 2007 constitutional revision. 15 A Gamper and F Palermo, ‘The Constitutional Court of Austria: Modern Profiles of an Archetype of Constitutional Review’ (2008) 3 Journal of Comparative Law 64, 78. On the constitutional system of Austria, see generally M Stelzer, The Constitution of the Republic of Austria—A Contextual Analysis (Oxford, Hart Publishing, 2011).

88 Monica Claes and Maartje de Visser courts also control the legislature by assessing laws against the constitution (using heads of review other than the competence provisions, such as fundamental rights provisions) and can protect the rights of individuals in specific cases by means of constitutional complaint procedures (Verfassungsbeschwerde, recursos de amparo). Turning to the composition of constitutional courts with a federal mandate, the Belgian court is the only one which clearly reflects the federal element in its make-up. Half of the justices of this constitutional court should be former MPs drawn from the federal, regional or community parliaments. None of the other countries has adopted a similar approach. For instance, in Germany, Article 36(1) of the Basic Law states that civil servants employed by the highest federal authorities are to be drawn from all the Länder in an appropriate proportion, so as to ensure that the Länder have an influence on federal authorities. However, the Bundesverfassungsgericht is the only highest federal authority which is not subject to this principle of regional proportionality. At any rate, there is no serious theory in Germany that claims that the Federal Constitutional Court is primarily dictated by the interests of the federation on account of its being a federal organ.16 In some of the countries under consideration, the senate (as representative of the federated entities) has the right to nominate a certain number of the judges. It cannot be excluded that the senate uses its right of nomination as a means of achieving some degree of regional representation n the court.17 Furthermore, in Austria, the idea is that the representation of the interests of the federated entities is ensured by the requirement that three of the judges and two substitutes must be permanently resident outside Vienna (the seat of the constitutional court). However, ‘in practice its impact is very limited’.18 This is a point worthy of note, given that the composition of constitutional courts is often organised so as to ensure that relevant interests are well represented. This, in turn, is expected to enhance the legitimacy of the court in the eyes of the groups so represented. Thus, in Belgium, Germany and Italy, a certain proportion of the justices of the constitutional court are drawn from the ordinary (supreme) courts.19 This is typically explained by the wish to ensure a good modus

16 U Steiner, ‘The Role of the Federal Constitutional Court within Germany’s Federal Structure’, at the international seminar Constitutional Courts and Federalism in Buenos Aires, 26 and 27 March 2008, available on . 17 So D Kommers and R Miller, ‘Das Bundesverfassungsgericht: Procedure, Practice and Policy of the German Federal Constitutional Court’ (2008) 2 Journal of Comparative Law 194, 199 (‘the participation of the Bundesrat in the selection of the Court’s Justices was meant to ensure that the Court was, at least with respect to staffing, steeped in Germany’s federalism’). 18 Gamper and Palermo, above n 15, 68. 19 In Belgium, 6 of the 12 judges must be former members of the Cour de cassation, the Conseil d’Etat, law professors or clerks (Belgian Special Law on the Constitutional Court, Art 34). In Germany, of the 16 members of the court, at least 6 must be former members of the federal judiciary with at least 3 years of experience (German Law on the Federal Constitutional Court, Arts 31 and 34). In Italy, the eligibility criteria for elevation to the constitution bench prescribe that appointees must be

The Court of Justice as a Federal Constitutional Court 89 vivendi between the ordinary courts and the constitutional court, since the latter is in a number of ways dependent on the co-operation of the former. The constitutional courts rely on the ordinary courts, for instance, to supply them with preliminary questions so that they can exercise their mandate of constitutionally reviewing laws. Another issue is whether ordinary courts accord judgments of the constitutional court binding or merely persuasive authority.20 Similarly, most constitutional courts examined here can count former politicians among their members. The rationale for this arrangement is twofold: (a) it should help raise awareness within the court of the political dimension inherent in its work, which may call for a more lenient standard of review at times, or at least point towards judicial sensitivity in crafting judgments with policy implications; and (b) it may breed (more) acceptance of the court’s judgments by the political institutions.21 In this light, and given the perceived bias in favour of the federal level on the part of constitutional courts with a federal mandate, one might therefore expect a similar acknowledgement of the need to ensure that the interests of the federated entities are reflected in the composition of the constitutional court as well. It should also be pointed out that judicial behaviour of the individual justices on the bench is not observable in these cases. In Belgium, Austria and Italy, judgments are always collegiate and no provision is made for separate opinions.22 The possibility of dissenting and concurring opinions does, however, exist in Germany and Spain.23 Yet, for instance, in relation to Germany, Donald Kommers has commented that the institutional bias against personalised judicial opinions has tended to minimise published dissents. There have been only 134 since they were first

‘chosen from among magistrates, including those in retirement, from supreme, ordinary and administrative courts’ (Italian Constitution, Art 135(2)). Similarly, in Spain, 2 of the 12 judges of the Tribunal Constitucional are appointed by the general council for the judiciary, which in practice also ensures the representation of (former) judges of the ordinary courts at the level of the Tribunal (Spanish Constitution, Art 159(1)). 20 These are in particular salient issues in Italy, where the possibilities to seize the Corte costituzionale directly by means of an action for annulment are limited and preliminary references traditionally account for the bulk of the court’s workload. See T Groppi, ‘The Italian Constitutional Court: Towards a “Multilevel System” of Constitutional Review?’ (2008) 3 Journal of Comparative Law 100. 21 On the composition of constitutional courts, see also V Ferreres Comella, Constitutional Courts and Democratic Values—A European Perspective (New Haven, Conn, Yale University Press, 2009) ch 5; Venice Commission, The Composition of Constitutional Courts (Science and technique of democracy collection, no 20, Strasbourg, Council of Europe Publishing, 1997). 22 In Austria, Art 30(1) of the Act on the Constitutional Court (Verfassungsgerichtshofgesetz) stipulates that the votes cast by the judges ‘are not public’, and this is taken to prevent judges from being able to publish dissenting opinions. See, eg Stelzer, above n 15, 203. Stelzer indicates that this issue is not free from debate and that scholars have expressed themselves in favour of allowing dissenting opinions. 23 German Law on the Federal Constitutional Court (Gesetz über das Bundesverfassungsgericht), Art 30(2); Ley Orgánica 2/1979 del Tribunal Constitucional, Art 90(2).

90 Monica Claes and Maartje de Visser allowed in 1970.24 Such dissenting opinions would make the personal preferences of individual justices visible. This may in general be considered detrimental to the acceptability and legitimacy of the Court’s judgments in sensitive cases. What is more, as far as the adjudication of vertical competence conflicts is concerned, the publication of vocal dissenting opinions could fuel tensions between the echelons of government, rather than reduce them. After all, the neutrality of the law and the court, as intended impartial arbiter, was an important element in deciding to entrust the court, rather than politics, with resolving such disputes. Lastly, it cannot be excluded that the expression of dissenting opinions may also have ramifications for compliance with the judgment by the ‘losing’ level, thus further jeopardising the legitimacy of the constitutional court within the overall system.

B. Design of the Procedure for Adjudicating Vertical Competence Disputes Having thus set the scene, this section considers how and by whom the constitutional courts under study may be seized of jurisdiction to adjudicate vertical competence disputes. In Germany, under Article 93(1)(3) of the Basic Law, the Bundesverfassungsgericht may act as an arbiter in conflicts of competence between the federation and individual Länder (Bund-Länder-Streitigkeiten). Disputes are most likely to arise out of disagreements concerning the administration of federal law by the Länder and associated oversight practised by the federation.25 Only a Land government and the federal Government may be parties to this procedure,26 which is subject to a six-month time limit. Although it enjoyed a short spell of popularity in the early years of the Bundesverfassungsgericht’s existence, the federation and the Länder now tend to resort to the action for annulment to resolve jurisdictional conflicts. This is in large part due to the more lenient admissibility conditions of that procedure, both in terms of standing (one-third of the Bundestag members may act as applicant) and the absence of a deadline within which such an action must be filed.27

24

Kommers and Miller, above n 17, 205. See also Arts 84–85 of the German Basic Law. The oversight procedure works as follows. If the Government has identified a defect in the administration of a federal law by a Land, it first notifies the Land of its findings. If the perceived defects are not corrected, the Government and the Land may petition the federal council (the Bundesrat), which will decide whether a violation has occurred. The decision of the Bundesrat may be contested before the constitutional court (German Law on the Federal Constitutional Court, Art 70). In addition, if a Land fails to comply with its obligations under the Basic Law or federal law, the federal Government may take measures to compel such compliance. This procedure has, to date, never been used, given the possibility for the two echelons to bring their disputes before the constitutional court. 26 German Law on the Federal Constitutional Court, Art 68. 27 See also W Heun, The Constitution of Germany—A Contextual Analysis (Oxford, Hart Publishing, 2011) 73. He gives the example of disputes regarding fiscal matters. 25

The Court of Justice as a Federal Constitutional Court 91 Article 138 of the Austrian Constitution states that the constitutional court pronounces on competence conflicts between the federation and the Länder.28 A distinction is made in this respect between concrete and abstract disputes. In the former case, there is a disagreement between the federation and the Länder in relation to a specific matter.29 The disagreement may produce an affirmative conflict of jurisdiction, whereby both levels assert the authority to deal with the same administrative matter, or concern a negative conflict of competence, with both the federation and the Land declining to issue a ruling or render a decision. Affirmative conflicts may be brought before the constitutional court by the federal Government and the Land government. The initiation of litigation immediately suspends the operation of the administrative procedure pending before the federal and Land authorities.30 Negative conflicts may be referred by the party whose claim was dismissed by both the federation and the Land.31 In the event of an abstract dispute, the constitutional court may be asked by the federation or the Länder to determine whether a draft law or regulation exceeds the competences allocated to the level seeking to adopt this law or regulation.32 Different from the other courts examined here, the Austrian constitutional court may thus engage in a priori review in the context of vertical conflict resolution. It should be noted that, in this situation, the court may assess the draft law or regulation only for compatibility with the constitutional distribution of competences between the federation and the Länder, and may not broaden the inquiry to examine the constitutionality of the draft law more generally. The Belgian Constitution does not provide for a special procedure for the judicial resolution of disputes over the vertical division of competences, akin to the German Bund-Länder Streitigkeiten. Claims that one of the government levels has exceeded its constitutional competences in law-making may instead be pleaded during the course of an a posteriori action for annulment, or be brought before the constitutional court indirectly, as a ground for review included in a preliminary reference from an ordinary court. Annulment actions may be initiated by the Council of Ministers, or by the government of a community or region, as well as by the presidents of the central, community or regional assembly, at the request of two-thirds of their members.33 In addition, individuals who can prove an interest may also file an annulment action, challenging laws adopted by any of the three government levels as being ultra vires. Annulment actions must in

28 The provision also grants the court the power to decide other competence disputes, ie those between courts and administrative authorities, those between courts and those between two Länder. 29 Austrian Constitution, Art 138(1)(c). 30 Austrian Act on the Constitutional Court, Art 47(4). 31 Ibid Art 50(1). 32 Austrian Constitution, Art 138(2); Austrian Act on the Constitutional Court, Arts 53–56. 33 Belgian Constitution, Art 142, read together with Belgian Special Law on the Constitutional Court, Art 2.

92 Monica Claes and Maartje de Visser principle be initiated within six months of the publication of the contested act in the official journal.34 Under Article 161(c) of the Spanish Constitution, the constitutional tribunal may hear petitions regarding ‘conflicts of jurisdiction between the State and the Autonomous Communities or amongst the Autonomous Communities themselves’. In terms of procedures, an essential factor here is the nature of the act which has given rise to the jurisdictional conflict. If this act is of legislative rank, it may be referred to the constitutional tribunal for scrutiny only by means of an action for annulment (recurso de inconstitucionalidad), with the claimant relying on lack of competence as the ground for its petition.35 In this respect, the situation in Spain resembles that prevailing in Belgium, where direct challenges to legislative acts on competence grounds can be brought before the court only by means of the general annulment action. Acts of sub-statutory rank adopted by either the State or an autonomous community may be contested by the other level before the constitutional tribunal by means of special procedures for the resolution of competence conflicts. The recurso de inconstitucionalidad and these special actions for conflicts of jurisdiction are linked, in that if a case is initiated using one of the special procedures and the constitutional tribunal considers that the resolution of the jurisdictional dispute is dependent on an examination of the constitutionality of an act of legislative rank, the tribunal must of its own motion convert the procedure into an action for annulment.36 Without wishing to give a complete overview of this latter procedure, some comments are warranted. The standing of the autonomous communities in the context of annulment action has been subjected to a twofold restriction by the drafters of the Organic Law on the Constitutional Tribunal (Ley Orgánica 2/1979 del Tribunal Constitucional).37 First, autonomous communities are not able to lodge an appeal of unconstitutionality against legislative enactments of their own or another autonomous community. Secondly, challenges against national statutes and other acts are admissible only if the autonomous communities can demonstrate that the impugned norm ‘may affect their own area of autonomy’.38 These restrictions were inspired by fears of excessive litigiousness on the part of the autonomous communities against each other as well as against the State, in view of the open-ended approach adopted by the Constitution to the territorial division of competences. The early case law of the Constitutional

34

Belgian Special Law on the Constitutional Court, Art 3(1). Spanish Tribunal Constitucional, No 32/1981, 28 July 1981 (disputaciones catalanas), para 1, confirmed and developed in No 49/1984, 5 April 1984 (películas X), paras 1–2. 36 Ley Orgánica 2/1979 del Tribunal Constitucional, Art 67. 37 Some commentators are of the opinion that this restriction in an organic law of an unconditional right granted by the Constitution is in fact unconstitutional. See the references in A Pascual Medrano, ‘Active Legitimization in Constitutional Proceedings: The Spanish Case’, Report for the 26th International Congress of Comparative Law (Brisbane, 2002), available at , 172. 38 Ley Orgánica 2/1979 del Tribunal Constitucional, Art 32(2). 35

The Court of Justice as a Federal Constitutional Court 93 Tribunal expounded a narrow interpretation of the impact requirement, but later judgments show a more generous approach: [T]he legitimisation of the autonomous communities to lodge an appeal of unconstitutionality is not aimed at vindicating an infringed competence, but at purifying the legal system and, in this sense, this legitimisation extends to all cases where a natural link between state law and the competences of the autonomous communities exists, which, in turn, cannot be interpreted in a restrictive way.39

Furthermore, while lodging a recurso de inconstitucionalidid in general does not result in the suspension of the contested legal norm, challenges by the Government against laws or acts adopted by the autonomous communities automatically suspend the operation of those norms for up to five months.40 When it comes to competence conflicts which have their origins in sub-statutory norms, a distinction is made between ‘positive’ and ‘negative’ conflicts, as in Austria. In the former case, both the State and at least one of the autonomous communities have asserted the authority to deal with the same issue, while in the latter scenario, both governmental levels have declared a lack of competence—an infrequent occurrence. The distinction has a bearing on the rules on locus standi: whereas positive conflicts may be referred to the Constitutional Tribunal by either the central Government or the executive body of the autonomous community, negative conflicts may also be brought by interested natural or legal persons.41 As with the recurso de inconstitucionalidad, the State and the autonomous communities do not enjoy equal procedural rights in the context of conflicts on the allocation of power. While the autonomous communities have standing in the procedure for positive conflicts only ‘where [their] own field of jurisdiction is affected’ by the norm they seek to challenge, the State is always admitted to court without any further qualification. Further, in the event of a positive conflict, if the procedure is initiated by the State and directed against a norm adopted by an autonomous community, the filing of the lawsuit immediately suspends the operation of the contested act.42 The Constitutional Tribunal must, however, decide to continue or terminate the suspension within five months.43 Conversely, if an autonomous community seeks to challenge a State act, it may petition the Constitutional Tribunal for a stay of that act, ‘citing damages impossible or difficult to redress’, but the tribunal has discretion in granting the requested interim relief.44 Lastly, the State may directly seize the Constitutional Tribunal to

39 Spanish Tribunal Constitucional, No 199/1987, 16 December 1987, para 3. See also No 84/1982, 23 December 1982. 40 Spanish Constitution, Art 161(2); Ley Orgánica 2/1979 del Tribunal Constitucional, Art 30. Within these five months, the Constitutional Tribunal must either confirm or lift the suspension. 41 Ley Orgánica 2/1979 del Tribunal Constitucional, Art 60. 42 Ibid Arts 64(2) and 77; Spanish Constitution, Art 161(2) (impugnacíon de disposiciones sin fuerza de Ley y resoluciones de las Comunidadas Autónomas). 43 Ley Orgánica 2/1979 del Tribunal Constitucional, Art 65(2). 44 Ibid Art 64(3).

94 Monica Claes and Maartje de Visser rule on a perceived competence dispute,45 whereas an autonomous community must first request the organ that authored the ultra vires act to reconsider it, and only if that application has not yielded the desired result or has gone unanswered within a set time-limit is it possible to refer the controversy to the Constitutional Tribunal for resolution.46 Article 127 of the Italian Constitution recognises the competence of the Corte costituzionale to adjudicate claims concerning the vertical allocation of competences (conflitti di attribuzioni). The State may contest regional laws as exceeding regional competence, and similarly each of the 20 regions47 and the autonomous provinces Trento and Bolzano may challenge State laws48 or laws adopted by another region for infringing the challenger’s competence. Challenges must be brought within 60 days of the publication of the disputed State or regional law. This procedural framework was introduced in 2001, in the context of a ‘quasifederal’ State reform.49 Until then, the Government enjoyed the competence to question the constitutional legitimacy of regional laws a priori, ie before their enactment, while the regions could challenge state laws only a posteriori.50 The new constitutional text has eliminated the possibility of preventive review and now grants equal procedural rights to the State and the regions. However, in its case law, the Corte costituzionale continues to recognise that there is a difference in the grounds of review that may be pleaded by each of the levels in cases brought under this heading. While the State is able to challenge a regional law for violating any provision of the Constitution, the regions may file petitions against State laws only for infringing the division of competences laid down in the Constitution.51 The Corte costituzionale may, of its own motion, suspend the execution of contested norm if there is a danger of an ‘irreparable prejudice to the public interest or the legal structure of the republic’, or ‘a grave and irreparable prejudice to the rights of citizens’.52 In addition, the State and the regions also have access to the Court to seek to impugn each other’s non-legislative acts that they believe encroach upon their own constitutionally-attributed competences.53

45

Ibid Art 62. Ibid Art 63. 47 Listed in Art 131 of the Italian Constitution. The competences attributed to the regions are set out in Art 117. 48 Or other acts of the State that have the force of law. 49 Constitutional Law No 1 of 22 November 1999; Constitutional Law No 2 of 31 January 2001; Constitutional Law No 3 of 18 October 2001. See also L Del Duca and P Del Duca, ‘An Italian Federalism?—The State, its Institutions and National Culture as Rule of Law Guarantor’ (2006) 54 American Journal of Comparative Law 799; T Groppi and N Scattone, ‘Italy: The Subsidiarity Principle’ (2006) 4 International Journal of Constitutional Law 131. 50 Constitutional Law No 1/1948; Constitutional Law No 87/1953. 51 Corte costituzionale, sentenza No 274/2003, 27 March 2003, confirming the position adopted in sentenza No 30/1959, 30 April 1959. 52 Law No 131/2003, Art 9(4), amending Law No 87/1953, Art 35. 53 Italian Constitution, Art 134; Law No 87/1953, Art 39. There is a time-limit of 60 days after notification of the measure or of the day on which the applicant became aware of its existence. 46

The Court of Justice as a Federal Constitutional Court 95 C. Identification of Some Core Common Features The above sections have yielded a number of differences in the manner in which the constitutional court is set up, may be seized of jurisdiction and decides. Yet there are also a number of common features that may be identified. First, the countries under examination have all decided to entrust a court with the resolution of competence disputes. It is readily understandable that courts have several advantages over other bodies as a forum for dispute resolution: they are (or, at least, are perceived to be) neutral and independent of the litigating parties, and they will decide the matter with reference to legal norms, not resorting to political considerations or the brute force of power.54 As such, the dispute is removed from the political arena and brought into the realm of law. Secondly, the judicial bodies in the Member States have all been conceived as separate constitutional courts, that is, as institutions located outside the ordinary judicial hierarchy. This choice in design does not seem essential to enable the judicial resolution of competence disputes. In the United States and Canada, for instance, the adjudication of such disputes is within the purview of the Supreme Court, located at the apex of the ordinary court system. Thirdly, with the exception of Belgium, the other five jurisdictions all provide for procedures specifically designed for the adjudication of vertical competence disputes. These procedures are often combined with the use of the general action for annulment procedure, where the applicant may assert a violation of the constitutional division of competences as the ground for review. This is particularly true in Spain, where allegations that the legislative acts of the Spanish State or the autonomous communities invade the jurisdictional domain of the other level should be referred to the Spanish Constitutional Tribunal by means of the general annulment action. Fourthly, both the central level and the federated entities have access to the constitutional court to allege an encroachment on their powers by the other level. However, in Spain and Italy, the federal level enjoys decidedly more advantageous procedural rights (as regards the suspensive effect of the appeal and the grounds that may be adduced for review) than the federated entities in procedures for competence infringement brought before the court. There is, thus, a bias in the system in favour of the central level. By contrast, in Austria, Belgium and Germany, that is, in the countries that consider themselves to be fully-fledged federations, all levels have access to the court on equal terms. In addition, in limited circumstances, the Spanish and Austrian rules also grant access to individuals in the context of the specialised competence procedure, namely, in cases of a negative conflict of jurisdiction, ie where both levels decline to act. In Belgium, which relies on the general annulment action to adjudicate

54 M Shapiro, Courts: A Comparative and Political Analysis (Chicago, Ill, University of Chicago Press, 1986).

96 Monica Claes and Maartje de Visser vertical separation of powers disputes, individuals too can go to the constitutional court to complain that a law adopted by one of the State levels violates the constitutional division of competences. More generally, to the extent that competence conflicts may be referred to the constitutional court by means of the annulment action, the normal rules on locus standi associated with these procedures are applicable, which tend to be more generous than those under the specialised competence procedure. The use of the general procedural avenues thus results in a larger set of petitioners that is able to refer violations of the constitutional division of competences to the constitutional court. Recall also that in Germany, the fact of more lenient rules on standing induced State organs to favour the action for annulment for resolving jurisdictional conflicts over the dedicated procedure for institutional dispute resolution. Fifthly, the constitutional court may annul the objectionable acts of either of the two levels, including acts of a legislative nature. These judgments annulling a (legislative) act for infringing the division of competences are recognised as binding on all authorities, at both the central level and at the level of the federated entities. This settlement is accepted in all the Member States mentioned earlier.55 European federal systems share these institutional and procedural arrangements, which seem to form a common core. They may well be essential for federal constitutional courts in European federal systems in order to achieve a smooth functioning thereof, settle conflicts between the federation and the federated entities, and guarantee a working federalism.

D. Prominence and Salience of Federal Issues in the Court’s Docket A final issue that warrants examination is the prominence and salience of vertical competence disputes within the overall docket of the court. Although there are no precise data available for all the courts discussed here covering the same period, the following observations are warranted. In Spain and in particular in Italy, the constitutional courts have been quite actively involved in deciding competence disputes following State reforms in both countries. According to the statistics available on the website of the Spanish Tribunal Constitucional, at the end of 2009, there were 456 pending cases, of which 97 concerned positive conflicts of jurisdiction; 47 annulment actions initiated by the Government against laws adopted by the autonomous communities; and 112 annulment actions filed by the autonomous communities against State laws. As for Italy, the number of State–region actions before the Corte costituzionale under Article 127 of the Constitution rose from 2 per cent in 2002 to 24.41 per cent 55 See Belgian Special Law on the Constitutional Court, Arts 9 and 115; German Law on the Federal Constitutional Court, Art 31 (note that this is the only provision to explicitly state that judgments are ‘binding upon federal and Land constitutional organs’); Austrian Constitution, Art 146(2); Spanish Constitution, Art 164(1); Italian Constitution, Art 136.

The Court of Justice as a Federal Constitutional Court 97 in 2006.56 It has been said that between 2003 and 2006, the Court was more concerned with adjudicating vertical competences disputes than with guaranteeing the protection of fundamental rights.57 But it is not just a matter of quantity. The constitutional courts in both countries have been called on to decide politically controversial conflicts of competence. This has notably been the case as regards the adoption of regional statutes by the lower-level entities which use these to assert a greater degree of autonomy and self-rule than previously.58 Conversely, in Germany the constitutional court is far less frequently seized to adjudicate competence claims, and there have not been any high-profile rulings regarding this issue in recent years. Indeed, Kommers writes that, as regards Germany: In truth, however, most federal–state conflicts never reach the Constitutional Court. They are usually resolved politically through the mechanisms of cooperative federalism and various forms of coordinated activity among the states. Even when governed by different political parties, the federal government and states prefer bargaining over litigation as the usual method of settling their differences.59

Various reasons may be adduced to explain the differences between the positions of the various constitutional courts. There is likely to be an increase in resort to litigation shortly after changes are made to the allocation of competences between the two government levels (be it through constitutional amendment or case law), to re-establish and re-define the precise constitutional position of each level and its competences. This would be followed by periods of relative calm after a new power equilibrium has been established, with both levels more amenable to political rather than judicial dispute resolution. The situation in Italy would seem to corroborate this: there, we observe a marked increase in litigation following the 2001 State reform. Additionally or alternatively, if the constitutional settlement of competence issues is cast in relatively imprecise and open terms, this leaves more room for competing interpretations of the scope of competences of each of the levels and the need for a third-party arbiter to settle these. For instance, and as mentioned earlier, in Spain the demarcation of powers between the State and the autonomous communities was purposively left vague, whereas in Germany the Basic Law is far more precise in this respect.

56

Groppi, ‘The Italian Constitutional Court’, above n 20, 114. Del Duca and Del Duca, above n 49, 799. 58 For Italy, see especially Italian Corte costituzionale, sentenzia No 372/2004, 29 November 2004; sentenzia No 378/2004, 6 December 2004; sentenzia No 379/2004, 29 November 2004. See also G Delledonne and G Martinico, ‘Handle with Care! The Regional Charters and Italian Constitutionalism’s “Grey Zone”’ (2009) 5 European Constitutional Law Review 218. For Spain, see most recently Spanish Tribunal Constitucional, sentencia No 31/2010, 28 June 2010 (on the new statute of autonomy for Catalonia). 59 D Kommers, The Constitutional Jurisprudence of the Federal Republic of Germany, 2nd edn (Durham, NC, Duke University Press, 1997) 87. 57

98 Monica Claes and Maartje de Visser IV. COMPARISON WITH THE CJEU AS A FEDERAL CONSTITUTIONAL COURT

How does the Court of Justice compare to the federal constitutional courts at the national level? Following the structure of the previous section, we shall first discuss the situational logic and composition of the Court of Justice, move on to consider some procedural aspects associated with the resolution of vertical competence disputes and, lastly, briefly address the salience of such cases within the Court’s overall case list.

A. Situational Logic and Composition Similar to the federal constitutional courts at the national level, the CJEU too has a mandate that is broader than only the adjudication of competence conflicts. Several guises of the Court may be distinguished. i. The CJEU as a Federal Constitutional Court The Court of Justice perceives itself as a constitutional court. More precisely, it carries out functions that, in part, resemble those exercised by constitutional courts in national systems.60 Pursuant to Article 19 TEU, the Court has the duty to ensure that the law is observed in the interpretation and application of the Treaties, the Union’s constitutional charter. It acts as the guardian of the EU legal order, a duty which it shares with the national courts.61 As such, the CJEU can control the European legislature by means of the action for annulment, or as a result of preliminary references made by those national courts, questioning the validity of Union legislative measures.62 The CJEU’s fundamental rights case law has been crucial for the development of fundamental rights protection in the context of the EU.63 In fact, it is precisely through the conceptualisation of

60 This is expressed, for instance, in the brochure published on the Court’s website ‘Your Questions on the Court of Justice of the European Communities’, February 2007 edn, 3. The claim is spread not least by the members of the Court themselves, see eg B Vesterdorf, ‘A Constitutional Court for the EU?’ (2006) 4 International Journal of Constitutional Law 607. See also the Report on Certain Aspects of the Application of the Treaty on European Union drafted in preparation for the 1996 IGC preparing the Treaty of Amsterdam and references thereto in M Claes, The National Courts’ Mandate in the European Constitution (Oxford, Hart Publishing, 2006) 401–23. 61 Kadi, above n 12; Opinion 1/09 on the creation of a European and Community Patent Court (ECJ, 8 March 2011). 62 Case 314/85 Foto-Frost v Hauptzollamt Lübeck-Ost [1987] ECR 4199; Case C-354/09 Gaston Schul BV v Staatssecretariat van Financiën (ECJ, 15 July 2010). 63 The case law is well known. See, eg Case 29/69 Eric Stauder v City of Ulm Socialamt [1969] ECR 419; Case 44/79 Hauer v Land Rheinland-Pfalz [1979] ECR 3727; Case C-260/89 ERT v Dimotiki Etairia Pliroforissis and Sotirios Kouvelas [1991] ECR I-2925; Kadi, above n 12. See recently G de Búrca, ‘The Evolution of EU Human Rights Law’ in P Craig and G de Búrca (eds), The Evolution of EU Law (Oxford, Oxford University Press, 2011).

The Court of Justice as a Federal Constitutional Court 99 fundamental rights as general principles of European law (and in the event of clear braches of those rights), that the CJEU has effectively made a stance, as a constitutional court, against the European law-making institutions.64 Lastly, the Court ensures respect for the autonomy of the European legal order.65 Focusing more particularly on the function of the CJEU as a federal constitutional court—acting as the judicial umpire of the vertical separation of powers— one of the heads of review that may be pleaded when challenging an EU legislative act is ‘lack of competence’. This enables the Court judicially to patrol the division of powers between the Union and its Member States on the basis of the Treaties. When it comes to the limits of the Union’s competences as such, and outside the context of fundamental rights or principles, the Court of Justice has generally been very reluctant to constrain the EU law-making institutions to the benefit of the Member States. The CJEU has condoned an expansive reading of the harmonisation competences enjoyed by the EU legislature and the Member States acting therein: while its ruling in Tobacco Advertising66 might have indicated a shift towards stricter judicial control, later case law shows again a more lenient approach.67 Also, the Court has interpreted the rules on the four freedoms very broadly, to the detriment of the legislative autonomy of the Member States.68 More recently, the extension of the Treaty provisions to cover horizontal relationships, in cases such as Viking69 and Laval,70 has brought with it an increased impact of EU law on private conduct, thus circumscribing the area within which national authorities are free to regulate.71 In addition, the CJEU has refrained from breathing constitutional life into the principle of subsidiarity,72 when the Commission,

64 D Wyatt, ‘Is the European Union an Organisation of Limited Powers?’ in A Arnull, C Barnard, M Dougan and E Spaventa (eds), A Constitutional Order of States (Oxford, Hart Publishing 2011) 10–11. See, eg Case 120/86 Mulder v Minister van Landbouw en Visserij [1988] ECR 23210. See also, after the entry into force of the Lisbon Treaty and the Charter of Fundamental Rights, Joined Cases C-92/09 and C-93/09 Volker und Markus Schecke GbR and Hartmut Eifert v Land Hessen (ECJ, 9 November 2010). 65 Opinion 1/91 (EEA Agreement I) [1991] ECR 6079; Opinion 1/09, above n 61. On the latter Opinion, see T Lock, ‘Taking National Courts More Seriously? Comment on Opinion 1/09’ (2011) 36 EL Rev 576. 66 Case C-376/98 Germany v Parliament and Council [2000] ECR I-8419 (‘Tobacco Advertising’). 67 See recently Case C-58/08 Vodafone, O2 et al v Secretary of State (ECJ, 8 June 2010). See also S Weatherill, ‘The Limits of Legislative Harmonization Ten Years after Tobacco Advertising: How the Court’s Case Law Has Become a “Drafting Guide”’ (2011) 12 German Law Journal 827. 68 See, eg S Weatherill, ‘Pre-emption, Harmonisation and the Distribution of Competence to Regulate the Internal Market’ in C Barnard and J Scott (eds), The Law of the Single European Market— Unpacking the Premises (Oxford, Hart Publishing, 2002). 69 Case C-438/05 International Transport Workers’ Federation and Finnish Seamen’s Union v Viking Line ABP and OÜ Viking Line Eesti [2007] ECR I-10779. 70 Case C-341/05 Laval un Partneri Ltd v Svenska Byggnadsarbetareförbundet et al [2007] ECR I-11767. See N Reich, ‘Free Movement v Social Rights in an Enlarged Union: The Laval and Viking Cases before the European Court of Justice’ (2008) 9 German Law Journal 125. 71 Wyatt, above n 64, 17. 72 Ibid 19. For a comparison with the practice of the US Supreme Court, see F Sander, ‘Subsidiarity Infringements before the European Court of Justice: Futile Interference with Politics or a Substantial Step towards EU Federalism?’ (2005) 12 Columbia Journal of European Law 517; G Bermann, ‘Taking

100 Monica Claes and Maartje de Visser the Council and Parliament, and perhaps even the national parliaments, failed to do so.73 ii. The CJEU as a Supreme Court The CJEU further has the task—which it takes very seriously—of ensuring the uniform interpretation and application of Union law across the Member States.74 This is striking when compared to the position of the federal constitutional courts in the countries under study here. There, other highest courts75 are entrusted with the mandate of guaranteeing the uniformity of the interpretation of non-constitutional laws, while the federal constitutional courts have exclusive competence with respect to the constitutionality of legislation. Yet there are federal systems where this is different. Think, for instance, of the United States Supreme Court. The explanation may be found in the manner in which judicial review has originated. In the US, decentralisation is practised, whereas the European countries examined here all adhere to the Kelsenian model of a separate constitutional court. In this respect, Victor Ferreres Comella has noted that while the US Supreme Court is at times able to side-step the constitutional issues by using traditional legal techniques and ordinary legislation to resolve the case, this option is much less available to the federal constitutional courts: they are required to speak.76 One might argue, and rightfully so, that to ensure the uniformity of the law has very little to do with enforcing a constitution. Yet the two cannot be entirely separated, notably when the uniformity of constitutional law itself is at stake. This, it is argued, is exactly what the ECJ is concerned with in many cases, even in those that do not overtly concern issues which in national settings would be considered constitutional.77 So-called constitutional cases and ‘ordinary’ cases are lumped together: there is no special court, and no special procedure to distinguish both types of cases. It might be said that EU law—regardless of the field at issue—is inherently linked to federal constitutional questions, as it always touches upon the relationship between the European and national sphere, albeit sometimes more clearly than on other occasions. The interpretation of ‘ordinary’ non-constitutional EU law Subsidiarity Seriously: Federalism in the European Community and the United States’ (1994) 94 Columbia Law Review 331. 73 See generally A Estella, The EU Principle of Subsidiarity and Its Critique (Oxford, Oxford University Press, 2002). 74 See recently Opinion 1/09, above n 61; Case 235/09 DHL Express France SAS v Chronopost SA (ECJ, 12 April 2011). 75 Eg, a Cour de cassation or Bundesgerichtshof. 76 Ferreres Comella, above n 21, ch 7. 77 This is also related to the fact that the European Treaties, as the Union’s constitution, include provisions that—at the national level—would not be included in the constitution but, rather, be found in ordinary legislation. A broader understanding of the notion ‘Constitution’ concomitantly leads to a broader understanding of the scope of constitutional jurisdiction.

The Court of Justice as a Federal Constitutional Court 101 by the Court may generate a centralising effect, and thus effectively impact on the division of powers between the centre and the parts.78 Consider the following comment by the current president of the Court: This criticism [that the ECJ is driven by a pro-European judicial activism] mainly focuses on the case-law of the Court in the areas of social law and tax law. Although this criticism deserves an elaborated answer, allow me please to deal with it here as briefly as possible: first, the Court has to give effect to a Treaty that is based on the idea of integration and that has integration as its purpose. In interpreting this Treaty, how not to place this integrative approach at the heart of the legal reasoning? Second, interpreting a piece of legislation that lacks clarity or whose interpretation deliberately leaves some freedom of manoeuvre will inescapably lead the Court to fill in the gap.79

A similar phenomenon has been observed in the United States, where federal constitutional rights have functioned as a federalising device,80 and in Germany, where the fundamental rights case law of the Bundesverfassungsgericht has had a unitary effect.81 To be sure, in the context of EU law, this radiating effect does not involve an extension of powers as such, or an additional transfer of competences, but it may increase the reach of the EU within the traditional domain of national law.82 The radiating effect of constitutional case law may also be detected in the context of indirect horizontal effect of fundamental rights, most conspicuously in Germany. In the famous Lüth judgment,83 the Bundesverfassungsgericht held that the value system underlying the constitutional rights in the Basic Law impacted on the entire legal order and all branches of law, thus effectively ‘constitutionalising’ the whole of German law. What these cases have in common is that the case law of the federal constitutional court impacts on the relationship between spheres and areas of law, even without formally changing the division of competences. This is exactly what occurs in the context of the EU as well, when the CJEU interprets EU law and indirectly increases the reach of EU law.

78 K Lenaerts, ‘Fundamental Rights to be Included in a Community Catalogue’ (1991) 16 EL Rev 367; P Eeckhout, ‘The EU Charter of Fundamental Right and the Federal Question’ (2002) 39 CML Rev 945. 79 V Skouris, ‘Self-Conception, Challenges and Perspectives of the EU Courts’ in I Pernice, J Kokott and C Saunders (eds), The Future of the European Judicial System in a Comparative Perspective (Baden-Baden, Nomos, 2006) 25. 80 See also the contribution of Aida Torres Pérez in ch 5 of this volume. 81 See Steiner, above n 16, 15, who writes: ‘The case law of the Federal Constitutional Court on the basic rights thus of necessity has a very powerful, unitary effect. It reduces the political and legislative scope of the Länder.’ 82 An example is the horizontal effect of fundamental freedoms in Laval (above n 70) and Viking (above n 69). The AG’s Opinion in the latter case expressly mentioned the German Lüth judgment. See Viking, above n 69, Opinion of AG Poiares Maduro, fn 38. 83 BVerfG 7, 198, decision of 15 January 1958. See, eg J Bomhoff, ‘Lüth’s 50th Anniversary: Some Comparative Observations on the German Foundations of Judicial Balancing’ (2008) 9 German Law Journal 121.

102 Monica Claes and Maartje de Visser iii. The CJEU as an International Court In addition, it must not be forgotten that the Court of Justice, whatever else it is, is still an international court.84 It was established pursuant to an international treaty, and procedures such as that found in Article 259 TFEU are reminiscent of the inter-State disputes that are the bread and butter of judicial bodies like the International Court of Justice. Perhaps more importantly for the purposes of this contribution, the CJEU lacks the power to annul national laws. All it can doeither in the context of infringement proceedings initiated by the Commission, or in its reply to a national court by means of a preliminary ruling—is to declare that a Member State has violated its European obligations. It is for the State to draw the necessary legislative consequences.85 iv. The CJEU as a Human Rights Court Until recently, an exploration of the various functions engaged in by the CJEU would not have included its characterisation as a human rights court. While the Court did consider such rights in its case law, as mentioned earlier, the protection of fundamental rights would generally be ancillary to the exercise of one of the other heads of jurisdiction, rather than an aim in itself. The grounds for review did not explicitly include violations of fundamental rights, and no provision was made for a special procedure dedicated to the protection of the fundamental rights of individuals, as is the case for instance in Germany or Spain. Yet, more recently, the role of fundamental rights in EU law has changed, and this will also have ramifications for the functioning of the Court of Justice. The entry into force of the Lisbon Treaty heralds a new era in the area of European fundamental rights, with the Charter being truly in force and the EU bound to accede to the European Convention on Human Rights (ECHR). In addition, the EU has been very actively engaged in adopting fundamental rights legislation since the beginning of this millennium, in particular in the field of EU equality law.86 Thus, the EU has developed a full-fledged equality law and conducts an equality policy, promoting equal treatment in a large range of fields, and affecting areas which

84

See Claes, above n 60, 399–400. This is so notwithstanding the possibility for the Court to impose financial penalties for a failure to comply with the demands of EU membership (Art 260 TFEU). 86 See, eg, Council Directive 2000/43/EC implementing the principle of equal treatment between persons irrespective of racial or ethnic origin [2000] OJ L180/22; Council Directive 2000/78/EC establishing a general framework for equal treatment in employment and occupation [2000] OJ L303/16; Commission, ‘Proposal for a Council Directive on implementing the principle of equal treatment between persons irrespective of religion or belief, disability, age or sexual orientation’ COM (2008) 426 final. Not connected to the principle of equality, see, eg, Council Regulation (EC) 1236/2005 concerning trade in certain goods which could be used for capital punishment, torture and other cruel, inhuman or degrading treatment or punishment [2005] OJ L200/1. 85

The Court of Justice as a Federal Constitutional Court 103 until recently had remained purely national.87 The newly-created ‘Directorate General Justice, Fundamental Rights and Citizenship’ of the Commission and the Fundamental Rights Agency, while restricted in scope and powers, form institutional reflections of the Union’s wish to become further engaged in fundamental rights protection. The EU is thus developing a veritable human rights policy. This EU involvement in fundamental rights protection accompanies the expansion of EU law into areas to which such rights are more relevant, such as criminal law, and immigration and asylum law. It may be argued that the EU is gradually turning into a human rights organisation, and that human rights are no longer purely ancillary. To be sure, the TEU, TFEU and the Charter repeat time and again that the Charter and accession shall not ‘extend the competences of the Union’, or ‘extend the field of application of Union law beyond the powers of the Union or establish any new power or task for the Union, or modify powers and tasks as defined in the Treaties’.88 Furthermore, the Charter is binding on the Member States ‘only in so far as they implement EU law’.89 Nevertheless, EU fundamental rights law at the same time necessarily reaches beyond the usual scope of EU law (defined by the powers conferred upon the Union), and beyond the implementation of EU law into the sphere of each Member State, where it coexists with other rights catalogues and their enforcement systems, such as the national Constitution, the ECHR and other international treaties. Article 2 TEU, for instance, imposes a general obligation on Member States to comply with human rights as one of the fundamental values of the Union, and Article 7 TEU provides for sanctions in cases of serious and persistent violations. Yet the issue of EU supervision remains notoriously delicate, with Member States jealously guarding their authority in these areas. The conflict over the French expulsion of Roma in the summer of 2010 and the outcry over the Commission’s intervention, or the reaction of the Hungarian Government to criticisms from European circles regarding the new Constitution and the new Media Law prove the point. Member States simply do not easily accept interference in what they still, after many years of international human rights supervision in Europe, consider purely domestic affairs. Turning back to the EU Charter, it clearly does not, and is not intended to, apply to purely national situations. Pursuant to Article 51(2), the Charter is binding on the Member States only when they are implementing EU law. However, it is as yet

87 European Parliament and Council Directive 2006/54/EC on the implementation of the principle of equal opportunities and equal treatment of men and women in matters of employment and occupation [2006] OJ L204/23; Council Directive 2004/113/EC implementing the principle of equal treatment between men and women in the access to and supply of goods and services [2004] OJ L373/37; Council Directive 2000/43/EC, above n 86; Council Directive 2000/78/EC, above n 86. Consider also the judicial contribution to achieving these goals (eg, Case C-144/04 Werner Mangold v Rüdiger Helm [2005] ECR I-9981; Case C-555/07 Seda Kücükdeveci v Swedex GmbH & Co KG [2010] ECR I-365). 88 Art 6(1) TEU and Art 51(2) Charter, respectively. 89 Art 51(1) Charter.

104 Monica Claes and Maartje de Visser unclear what this means, and what exactly triggers the application of the Charter. Judgments such as Ruiz Zambrano appear to show the Court taking an expansive view of what may constitute such a trigger, which may seem to expand the application of the Charter.90 Surely, neither the Treaties nor the Charter confer upon the Union additional competences to adopt European human rights legislation, so that we cannot speak of a general EU human rights competence. Yet the precise reach of EU fundamental rights law is much less clear. Furthermore, while fundamental rights protection cannot qualify as a policy area in its own right, respect for fundamental rights is a condition of the lawfulness of EU acts and implies the obligation to uphold fundamental rights across the board. Put differently, fundamental rights thus permeate all other policy fields. These developments in the area of fundamental rights will undoubtedly affect the position of the CJEU, including its relationship with national (constitutional) courts. To start with, it might be said that the inclusion of a stronger human rights mandate for the CJEU could further strengthen its characterisation as a constitutional court, given that most national constitutional courts consider fundamental rights protection as one of their core functions. The Charter may provide additional ammunition when it comes to filing successful challenges against acts adopted both by the EU institutions and by the Member States. Particularly salient for present purposes is the relative vagueness in which the Union’s scope for involvement in the fundamental rights field is cast. This reinforces the need for a judicial umpire to demarcate the jurisdictional boundaries of the Union and the national level in this respect. To conclude, the Court of Justice exercises functions and has features that support its characterisation as a federal constitutional court. Adjudicating vertical competence disputes is not the only function performed by it. Moreover, the Court’s other functions and guises impact on the relationship between the European and the national spheres as well. Yet there are also noticeable differences between the CJEU and the national federal constitutional courts. These are a direct result of the fact that the Union is not a State and that the authority of the CJEU—and that of the constitution it seeks to enforce—is at times challenged. v. Composition Turning to the composition of the Court, it is well-known that the CJEU has one judge per Member State.91 As such, it may be said that—in contrast to the situation in the majority of countries discussed in the previous section—the federal composition of the Union is properly reflected in the Court’s make-up. Having said that, the reason for such a composition probably has less to do with federalist thinking than with the Member States’ (original) conception of the CJEU as an 90

Case C-34/09 Gerardo Ruiz Zambrano v Office national de l’emploi (ONEM) (CJEU, 8 March 2011). Art 19(2) TEU. The Civil Service Tribunal is not considered here, as this judicial body does not adjudicate vertical competence disputes. 91

The Court of Justice as a Federal Constitutional Court 105 international court. Moreover, those CJEU judges coming from one of the federal systems discussed here may be drawn from the central level of government (eg, from the federal judiciary). This means that the representation of the federal character of the Member State in question in the CJEU may be partial and exclude the lower echelons of government, such as the Länder in Germany, the Regions and Communities in Belgium, and so forth. Another point worthy of mention is that the design of the rules of appointment at the European level may allow the pendulum of ensuring adequate representation of affected interests to swing too far in the other direction. The terms of CJEU judges are fairly short and subject to renewal.92 It cannot be excluded that a Member State may thus keep an eye on its judge to see whether he or she furthers that State’s interests in judicial decision-making, with possible repercussions when the judge’s term on the bench is about to come to an end.93 This may give the impression that a judge is not sufficiently independent, with concomitant risks for the legitimacy and prestige of the Court as a whole. The possibility of Member States controlling ‘their’ judges would increase significantly if dissenting opinions were to be introduced in the CJEU.94

B. Procedural Issues Regarding Vertical Competence Adjudication What can be said about the procedural rules governing the adjudication of competence conflicts by the Court of Justice? A first point to note is that there is no special procedure available at Union level to assert that one of the levels has overstepped its competences. Yet almost all of the other available procedural avenues included in the Treaties may have some bearing on this issue. In particular the general action for annulment, whereby the applicant may claim a lack of competence on the part of the EU (ie, the federal level), and infringement proceedings initiated by the Commission are commonly used in this respect. As such, the Union legal system closely mirrors the Belgian one, which also makes do without a separate procedure for competence disputes. Secondly, both the EU institutions (in the form of the Commission by means of the infringement procedure) and the Member States have access to the Court of Justice to vindicate the perceived infringement of their competences. One question in this context is whether only the ‘federated entities’ (ie, the Member States)

92

Art 253 TFEU. In favour of separate opinions as a means of strengthening the authority of the CJEU in the eyes of European citizens, see V Perju, ‘Reason and Authority in the European Court of Justice’ (2009) 49 Virginia Journal of International Law 307. 94 Not to mention possible concerns relating to Art 6 ECHR, in light of the anticipated accession of the EU to the ECHR. On accession more generally, see, eg, JP Jacqué, ‘The Accession of the European Union to the European Convention on Human Rights and Fundamental Freedoms’ (2011) 48 CML Rev 995; T Lock, ‘Walking on a Tightrope: The Draft Accession Agreement and the Autonomy of the EU Legal Order’ (2011) 48 CML Rev 1025. 93

106 Monica Claes and Maartje de Visser have a direct right of access to the CJEU, or whether this right may be extended to lower-level echelons (such as the regions).95 It will be interesting to see what (if any) the impact will be of the new formulation of the subsidiarity principle after the Lisbon Treaty, which now explicitly recognises the regional and local tiers of government, perhaps in combination with the right of national parliaments to bring proceedings before the Court (via their Member State) to argue that a Union legislative act infringes the principle of subsidiarity.96 Besides the public actors, private actors may also have a part to play in engaging the CJEU for the purpose of resolving competence disputes. They may do so either by bringing an annulment action if they are able to show ‘direct and individual concern’, or by initiating litigation before the national courts, asserting the invalidity of an EU measure and requesting the use of the preliminary reference procedure. In broadening access to the CJEU beyond the tiers of government, the European system approximates the Belgian and—to a lesser extent—the Spanish rules on locus standi for vertical competence disputes. Thirdly, there is an asymmetry in the system when it comes to the decisional competences of the CJEU, as remarked earlier. While the Court may annul the laws adopted by the Union legislature (ie, the equivalent to the federal level), it lacks the ability to quash laws promulgated by the Member States (ie, equivalent to the federated entities). As such, it could be said that the Union system also incorporates a procedural bias—if this is taken to denote more favourable treatment of one of the tiers of government—but that this bias is in favour of the federated as opposed to the central level, unlike the situation in, for instance, Spain and Italy. Fourthly, arguably more important than the previous point is the issue of final authority. As we have seen, an essential feature of the national federal constitutional courts is that their authority is final and is recognised as such: their judgments on competence issues are binding and must be complied with by all other organs of the State, both at the federal level and at the level of the federated entities. Non-compliance with the decision of the constitutional court is considered a breach of the Constitution itself, and this is accepted as such by the other players in the system. If the political bodies of either level were to disagree with a ruling, their only option to act in accordance with the Constitution would be to amend the Constitution itself, which requires the co-operation of the other level. In contrast, the claim of the CJEU that it alone is the final arbiter of competence claims is not generally recognised by several national constitutional courts, which approach the question of competences from a different perspective, namely, the national (or federated) perspective. Thus, and unlike the situation at the national 95 P Van Nuffel, ‘What’s in a Member State? Central and Decentralized Authorities before the Community Courts’ (2001) 38 CML Rev 871. 96 See Art 5(3) TEU and Art 8 of Protocol No 2 on the application of the principles of subsidiarity and proportionality [2010] OJ C83/206. See also P Van Nuffel, ‘The Protection of Member States’ Regions Through the Subsidiarity Principle’ in C Panara and A De Becker (eds), The Role of the Regions in European Governance (Heidelberg, Springer, 2011); R Schütze, ‘Subsidiarity after Lisbon: Reinforcing the Safeguards of Federalism?’ (2009) 68 Cambridge Law Journal 525.

The Court of Justice as a Federal Constitutional Court 107 level, the European idea of judicial Kompetenz-Kompetenz is heavily contested.97 Unlike the authority of the constitutional courts of the federal Member States, the CJEU’s position is challenged. Under this reading, non-compliance with CJEU judgments would not necessarily constitute a violation of the Constitution, but only of a contested interpretation thereof.

C. Prominence and Salience of Federal Issues Disputes requiring the CJEU to decide on the allocation of competences between the Union and the Member State account for a significant part of the Court’s overall caseload. What is more, many of the rulings dealing with federal issues are also significant from a substantive (as opposed to merely quantitative) point of view. This salience may be readily understood if we consider that the division of authority between the national and the Union level is continuously under development, which may create an on-going need for triadic dispute resolution. In addition, the relevant provisions in the Treaties are not always clear, leaving more room for competing interpretations of the scope of competences of each of the levels which are no longer very amenable to a political solution. As such, the CJEU finds itself in a position comparable to that in which the Italian Corte costituzionale and the Spanish Tribunal Constitucional find themselves at present. It will be interesting to see whether the inclusion of a competence catalogue in Articles 3 to 6 TFEU will result in an increase in disputes concerning the allocation of powers between the Union and the Member States. It may be ventured that, to the extent that this catalogue codifies the extant status quo rather than introduces a considerable shift in competences, there would be no (new) uncertainty or confusion that may incentivise the EU institutions or the Member States to bring more competence matters before the Court. Perhaps the greater precision may even result in a decrease in allocation of powers disputes.

V. CONCLUSION

Despite the many institutional and procedural dissimilarities between the CJEU and traditional federal constitutional courts, which reflect the different situational 97 See J Beneyto and I Pernice (eds), Europe’s Constitutional Challenges in the Light of the Recent Case Law of National Constitutional Courts (Baden-Baden, Nomos, 2011); J Komárek, ‘European Constitutionalism and the European Arrest Warrant: In Search of the Limits of “Contrapunctual Principles” ’ (2007) 44 CML Rev 9; Claes, above n 60, Pt II; S Boom, ‘The European Union after the Maastricht Decision: Is Germany the “Virginia of Europe”?’ Jean Monnet Working Paper no 5/1995, considering the challenge by the German Bundesverfassungsgericht in its Maastricht Urteil to the final authority of the CJEU to draw the boundaries of the Union’s powers in the light of the US experience. In the US, the Supreme Court and the Virginia Court of Appeal were locked in a power struggle on the same issue. Eventually, the Supreme Court won the struggle. It might be said that, in the EU, there now seem to be several ‘Virginias’ besides Germany.

108 Monica Claes and Maartje de Visser logic of the CJEU, there are a number of important parallels too. Crucially, the CJEU perceives itself as a constitutional court. In fact, one might even say that the Court has been on a mission to spread the message of the constitutionalisation of Europe, which is inevitably linked to the constitutional character of the Court itself: the CJEU made itself a constitutional court when conducting the transformation of Europe.98 What is more, the CJEU is seen as a constitutional court by many, not in the least by the national constitutional courts which have accepted it as interlocutor and as an observer in the European Conference of constitutional courts. Moreover, the adjudicatory functions of the CJEU are those which in federal systems are entrusted to the federal constitutional court: it patrols the vertical division of competences, reviews the validity of ‘federal’ law (in this case, of EU law) and protects the integrity of the common legal order. The CJEU thus behaves like a federal constitutional court: its methodology resembles that of constitutional courts, rather than of traditional international courts or non-constitutional courts. Use of general principles, teleological interpretation and concern for the effectiveness of the law are often seen by EU scholars as emblematic of CJEU case law, but they are very common in constitutional courts at the national level as well. Lastly, the CJEU talks like a constitutional court when—admittedly less so than before—it has recourse to constitutional language. In short, applying the ‘duck-test’, the CJEU certainly may qualify as a federal constitutional court. Yet, as said, important differences between the CJEU and federal constitutional courts remain. The CJEU lacks the jurisdiction to annul national law, including for breaching the vertical separation of powers, and it remains dependent on the co-operation of the Member States and their willingness to comply with its judgments. The final authority of the CJEU in matters of competences (ie, its Kompetenz-Kompetenz) remains challenged by several constitutional courts. The same applies to its authority in the context of fundamental rights protection. Paraphrasing Steve Boom,99 it could be argued that Europe has many Virginias, which still do not recognise the Court’s ultimate authority. This has to do with Europe’s style of federalism, which remains, to a large extent, dependent on the voluntary submission of the independent and sovereign Member States to the federal discipline. This, at least, seems to be the conception of the EU which prevails in the Member States and their constitutional courts: at the end of the day, they have retained ultimate sovereignty, and they can decide to withdraw or—less dramatically—to protect their national identity, the core of their Constitution or their inalienable rights as laid down in the national Constitution against the central European level. As long as this remains the case, the EU cannot be considered a mature federation and the CJEU cannot be considered a full-fledged federal constitutional court.

98 99

See more extensively Claes, above n 60. Boom, above n 97.

The Court of Justice as a Federal Constitutional Court 109 What normative lessons may be drawn from this analysis? Most importantly, it is submitted, the European style of federalism, the situational logic of the CJEU and the continuous challenges to its authority as a federal court place an additional burden on the CJEU to earn its legitimacy as a federal constitutional court, and to act like one convincingly, when compared to similar national courts. The European type of federalism thus urges the Court to pay proper attention to the sensitivities of the Member States and their (constitutional) courts, for instance in the field of fundamental rights protection or national identity, and to take seriously its task as guardian of the division of competences in the Treaties. This might require the Court to police the jurisdictional boundaries of the European legislature more strictly, for instance by breathing judicial life into the principle of subsidiarity or by tightening its Tobacco-Advertising line of case law. Lastly, while judicial efforts to engage with the Charter of Fundamental Rights must surely be welcomed, the Court should be aware of the potential centralising tendencies that fundamental rights may exhibit, and exercise due care when applying these rights to situations with a somewhat tenuous link to EU law.

5 The Dual System of Rights Protection in the European Union in Light of US Federalism AIDA TORRES PÉREZ*

In the compound republic of America, the power surrendered by the people is first divided between two distinct governments, and then the portion allotted to each subdivided among distinct and separate departments. Hence a double security arises to the rights of the people. The different governments will control each other, at the same time that each will be controlled by itself. (Madison No 51)1

I. INTRODUCTION

A

S ROBERT C Post put it, ‘federalism is a chameleon-like concept’.2 Thus, discussing ‘federalism as a model’ for the European Union (EU) in the field of fundamental rights requires understanding, first, what we mean by federalism. The federalist system taken as a point of reference is the United States (US). Whereas in the EU the debate about ‘federalism’ generally refers to the empowerment of supranational authorities vis-à-vis the Member States, at present, in the US, the debate about federalism refers to preserving state autonomy for the independent interpretation of state constitutional rights. At the founding, however, ‘federalists’ where those who defended a strong federal government, as opposed to the ‘anti-federalists’, who claimed for safeguarding state autonomy. Thus, depending on the moment to which we are referring, federalism in the US conveys different approaches to the interaction between federal and state authorities, and the US federal system has been evolving over time. * Constitutional Law Professor at Pompeu Fabra University (Barcelona). I thank Elke Cloots for the invitation to participate in the ‘Workshop on Federalism in the European Union’, held at the University of Leuven School of Law (12–13 May 2011), and to all participants for the intense discussion and illuminating comments. 1 C Rossiter (ed), The Federalist Papers (New York, Mentor, 1999) 291. 2 RC Post, ‘Justice Brennan and Federalism’ (1990) 7 Constitutional Commentary 227, 227.

The Dual System of Rights Protection in the European Union 111 To what extent might US federalism be a model for the EU in the field of fundamental rights protection? Although the historical and political circumstances differ greatly, from a conceptual standpoint, the questions that the EU faces are very similar to those that emerged in the US throughout the process of federation-building. Rather than taking the US as a model, or suggesting that the EU should reproduce US answers, we shall examine what lessons may be learnt from looking into the evolution of the US Bill of Rights and the doctrinal debates on the intersection between federalism and rights.3 The model of rights protection in divided-power systems is always debated. This is especially true in the EU, given the elusive character of this supranational form of government.4 Broadly speaking, in divided-power compounds, the model of rights protection is inextricably linked to structural elements.5 As a consequence of this structural dimension of rights catalogues, several questions need to be addressed: Are the constituent states bound by the federal/supranational rights? How are the powers to regulate and enforce fundamental rights to be allocated between the states and the central government? How should interpretive conflicts be resolved? Underlying all these questions there is an under-theorised theme regarding the normative ideal about how to strike the balance between diversity and uniformity. These questions have been faced in the US throughout the process of building the federation. At present, the application of the federal Bill of Rights to the states is taken as given, but this was a highly debated issue for a long time.6 The empowerment of the federal Congress to enforce the Bill of Rights against the states, under section 5 of the Fourteenth Amendment, still gives rise to controversy.7 Moreover, scholars have shown increasing interest in state constitutionalism and the diverging interpretations of parallel federal and state rights. Over time, the structural dimension of the Bill of Rights has brought about considerable case law and a vast literature converging on the topic of federalism and rights.8 Thus, US federalism may well provide illuminating insights for EU dilemmas.

3 E Katz and GA Tarr (eds), Federalism and Rights (Lanham, Rowman & Littlefield Publishers, 1996) ix. Katz and Tarr began their introduction to this volume with the following question: ‘Does federalism promote or undermine rights?’ 4 K Lenaerts, ‘Federalism and Rights in the European Community’ in Katz and Tarr (eds), above n 3, 139. 5 See AR Amar, The Bill of Rights (New Haven, Conn, Yale University Press, 1998) xiii; Lenaerts, above n 4, 140. 6 See Amar, above n 5, 137–294; RC Cortner, The Supreme Court and the Second Bill of Rights. The Fourteenth Amendment and the Nationalization of Civil Liberties (Madison, University of Wisconsin Press, 1981); AE Wilmarth, Jr, ‘The Original Purpose of the Bill of Rights: James Madison and the Founders’ Search for a Workable Balance between Federal and State Power’ (1989) 26 American Criminal Law Review 1261. 7 MW McConell, ‘Institutions and Interpretations: A Critique of City of Boerne v Flores’ (1997) 111 Harvard Law Review 153; EH Caminker, ‘ “Appropriate” Means-Ends Constraints on Section 5 Powers’ (2001) 53 Stanford Law Review 1127. 8 See, eg Katz and Tarr (eds), above n 3; AR Amar, ‘Of Sovereignty and Federalism’ (1987) 96 Yale Law Journal 1425; AR Amar, ‘Foreword: Lord Camden Meets Federalism—Using State Constitutions

112 Aida Torres Pérez In what follows, we shall reflect, from the standpoint of the US federal system, on two controversial issues regarding the relationship between the EU and the Member States in the field of fundamental rights protection. Section II. will focus on the scope of application of EU fundamental rights to the Member States. Section III. will examine the potential conflicts between EU and constitutional rights when they overlap in their scope of application.

II. THE SCOPE OF APPLICATION OF EU FUNDAMENTAL RIGHTS TO THE STATES

The Lisbon Treaty has finally granted legally-binding force to the Charter at the same level of the Treaties (Article 6(1) TEU). As the US experience demonstrates, written declarations of rights might have a centralising force.9 The original purpose of the US Bill of Rights was to limit the action of federal powers. Over time, however, federal rights became binding on state authorities through the so-called ‘incorporation process’.10 Fundamental rights of the EU have also been incorporated into the Member States, but not fully.11 Since this process has already been examined by several authors,12 we shall focus on the latest developments on this issue.

A. The Incorporation of Federal Rights in the US As is well known, the 1787 US Constitution did not include any declaration of rights. Over the constitution-drafting process, and more intensely over ratification, anti-federalists strongly claimed for the inclusion of a declaration of rights.13 In several states, such as Massachusetts, New Hampshire, Virginia and New York,

to Counter Federal Abuses’ (1996) 27 Rutgers Law Journal 845; WJ Brennan, Jr, ‘The Bill of Rights and the States: The Revival of State Constitutions as Guardians of Individual Rights’ (1986) 61 New York University Law Review 536; L Friedman, ‘The Constitutional Value of Dialogue and the New Judicial Federalism’ (2000) 28 Hastings Constitutional Law Quarterly 93; PW Kahn, ‘Interpretation and Authority in State Constitutionalism’ (1993) 106 Harvard Law Review 1147. 9 Lenaerts held that ‘federalism and rights have had a mutually reinforcing impact on each other’ (above n 4, 139). 10 KM Sullivan and G Gunther, Constitutional Law (New York, Foundation Press, 2001) 433–50; Cortner, above n 6. 11 Lenaerts, above n 4, 157–58. 12 See, eg A Knook, ‘The Court, the Charter, and the Vertical Division of Powers in the European Union’ (2005) 42 CML Rev 367; P Eeckhout, ‘The EU Charter of Fundamental Rights and the Federal Question’ (2002) 39 CML Rev 945; A Torres Pérez, ‘La dimensión structural de la Carta de Derechos Fundamentales de la Unión Europea: relaciones verticales y cláusulas horizontales’ (2003) 67 Revista Vasca de Administración Pública 253. 13 HJ Sotring (ed), The Complete Anti-Federalist, vol 1 (Chicago, Ill, The University of Chicago Press, 1981) 64–70; B Schwartz (ed), The Roots of the Bill of Rights (New York, Chelsea House Publishers, 1980).

The Dual System of Rights Protection in the European Union 113 where the battle between federalists and anti-federalists was particularly intense, the ratification of the Constitution was subject to the recommendation of further amendments that would limit the power of the federation.14 Eventually, in 1791, ten new amendments were included: the so-called ‘Bill of Rights’. The debates leading to the amendment of the Constitution and the structure of these provisions clearly show that the Bill of Rights was conceived as an instrument to limit the action of federal institutions.15 The US Supreme Court, in Barron v Baltimore,16 confirmed that the Bill of Rights was only binding upon federal authorities. Akhil R Amar has defended the thesis that the main goal of the Bill of Rights was to protect the autonomy of the states vis-à-vis federal power.17 The original purpose of the US Bill of Rights was essentially transformed after the Civil War and the enactment of the Reconstruction amendments between 1865 and 1870. Throughout time, the due process clause of the Fourteenth Amendment was interpreted to ‘incorporate’ the rights included in the Bill of Rights into the states. The so-called ‘incorporation process’ was steady and slow, and it took place on a case-by-case basis.18 Eventually, both federal and state authorities would be bound by the federal Bill of Rights. In the words of Justice Brennan, incorporation ‘transformed the basic structure of constitutional safeguards for individual political and civil liberties in the nation and profoundly altered the character of our federalism’.19

B. From ‘Half-way’ to ‘Full’ Incorporation in the EU? Given the lack of a written declaration of rights in the EU, after the European Court of Justice (ECJ) proclaimed the primacy of EU law over state law, including constitutional law, it was soon realised that EU institutions were not bound by fundamental rights. Thus, at first, EU fundamental rights crafted by the ECJ case law were addressed to the action of EU institutions. By the end of the 1980s, however, EU fundamental rights were progressively incorporated into the Member States by the ECJ.20 14 B Schwartz, The Great Rights of Mankind. A History of the American Bill of Rights (Madison, WI, Madison House Publishers, 1992) 119–59; R Ketchman (ed) The Anti-Federalist Papers and the Constitutional Convention Debates (New York, Mentor, 1986) 217–26; Knook, above n 12, 374–76. 15 Wilmarth, above n 6; P Finkelman, ‘James Madison and the Bill of Rights: A Reluctant Paternity’ (1990) 9 The Supreme Court Review 301. 16 Barron v Mayor and City of Baltimore, 32 US 243 (1833). 17 Amar, above n 5. 18 Cortner, above n 6; Amar, above n 5. 19 Brennan, above n 8, 543. 20 See generally, B De Witte, ‘The Past and Future Role of the European Court of Justice in the Protection of Human Rights’ in P Alston (ed), The EU and Human Rights (Oxford, Oxford University Press, 1999); J Coppel and A O’Neill, ‘The European Court of Justice: Taking Rights Seriously?’ (1992) 29 CML Rev 669; JHH Weiler and N Lockhart, ‘ “Taking Rights Seriously” Seriously: The European Court and its Fundamental Rights Jurisprudence’ (1995) 32 CML Rev 51; JHH Weiler, ‘Fundamental rights and Fundamental Boundaries: On the Conflict of Standards and Values in the Protection

114 Aida Torres Pérez In contrast to the US, there was no ‘constitutional moment’, like the Reconstruction amendments, justifying such a step. At the same time, incorporation in the EU was ‘half-way’: EU fundamental rights were incorporated into the Member States in so far as they acted within the field of application of EU law.21 Hence, Member States would not be bound by EU rights in ‘purely internal situations’. The line separating state action falling within or outside the scope of application of EU law, however, is not clear-cut. In particular, according to ECJ case law, EU fundamental rights apply to the States in two types of situations: (a) when enforcing regulations22 or implementing directives;23 and (b) when derogating from free movement on the basis of: explicit Treaty provisions,24 overriding requirements admitted by the ECJ case law25 or fundamental rights.26 Still, the ECJ has applied EU fundamental rights to the States in cases not clearly falling within these two categories, in which the States broadly act within the field of application of EU law.27 Having in mind the evolution of the US Bill of Rights, one might ask: To what extent might a binding Charter further promote the incorporation of EU rights into the States? One of the most debated issues over the Charter’s drafting process and later on within the constitutional convention was the scope of application of the Charter to the States.28 The Charter has regulated its scope of application in Article 51(1).29 This provision confirms the ‘half-way’ incorporation of EU rights into the States, but there is still debate regarding its interpretation. The text of Article 51(1) reflects the willingness to restrict the scope of application of EU rights to the States, since it lays down that the Charter is addressed to the Member States ‘only when they are implementing Union Law’ (emphasis added). Accordingly, it has been argued that state measures derogating from free of Human Rights in the European Legal Space’ in JHH Weiler (ed), The Constitution of Europe (Cambridge, Cambridge University Press, 1999). 21 A Torres Pérez, Conflicts of Rights in the European Union. A Theory of Supranational Adjudication (Oxford, Oxford University Press, 2009) 16–26. 22 Case C-5/88 Wachauf [1989] ECR 2609. 23 Joined Cases C-20/00 and C-64/00 Booker Aquaculture [2003] ECR I-7411. 24 Arts 36, 45 and 52 TFEU. See Case C-260/89 ERT [1991] ECR I-2925. 25 Case C-368/95 Familiapress [1997] ECR I-3689. 26 Case C-112/00 Schmidberger [2003] ECR I-5659. See also Eeckhout, above n 12, 945–54; Knook, above n 12, 368–71. 27 Torres Pérez, Conflicts of Rights, above n 21, 18–19. 28 G de Búrca, ‘The Drafting of the European Union Charter of Fundamental Rights’ (2001) 26 EL Rev 126, 136–37; I Pernice, ‘Integrating the Charter of Fundamental Rights into the Constitution of the European Union: Practical and Theoretical Propositions’ (2003) 10 Columbia Journal of European Law 5, 23–25. 29 This article was rewritten several times, first within the drafting Convention and later on by the constitutional Convention. See Eeckhout, above n 12, 954–57.

The Dual System of Rights Protection in the European Union 115 movement (category b) above) should be excluded from the scope of application of the Charter.30 On the other hand, the ‘Explanations relating to the Charter of Fundamental Rights’31 not only fail to confirm this restrictive interpretation, but reproduce the more general form used by the ECJ: ‘[T]he requirement to respect fundamental rights defined in the context of the Union is only binding on the Member States when they act in the scope of Union law’ (emphasis added). In addition, the Explanations expressly quote the ECJ’s judgment in ERT,32 which is the leading case applying EU rights to state action derogating from EU law. Therefore, it does not seem that this article can work to restrict the scope of application of the Charter to the States.33 Indeed, the analysis of the ECJ case law shows a trend to enhance the incorporation of EU rights to the States. i. Non-discrimination on Grounds of Nationality Remarkably enough, the principle of non-discrimination on grounds of nationality has been applied to monitor state action slightly connected with EU law.34 The text of Article 18 TFEU circumscribes the enforcement of the principle of non-discrimination on grounds of nationality to the ‘scope of application of the Treaties’. In practice, the connection of state action with the EU freedoms of movement has been enough to require compliance with the principle of nondiscrimination on grounds of nationality by the States.35 The case of Bickel and Franz36 is very illustrative. Mr Bickel, an Austrian lorry driver, and Mr Franz, a German tourist, were charged with committing two respective crimes while in Italy. Under Italian law, the German-speaking citizens of the Province of Bolzano were entitled to use their own language in relations with the judicial and administrative authorities based in that province. Bickel and Franz requested that the judicial proceedings be conducted in German. The state court referred a preliminary question to the ECJ, asking whether the principle of non-discrimination on the basis of nationality (ex Article 12 EC Treaty) and the right of free movement and residence (ex Article 18 EC Treaty) required that the right to have criminal proceedings conducted in German be extended to non-citizens. 30

F Jacobs, ‘Human Rights in the EU: The Role of the Court of Justice’ (2001) 26 EL Rev 331. Explanations relating to the Charter of Fundamental Rights [2007] OJ C303/17. 32 ERT, above n 24. 33 In J McB v LE, issued after the Charter acquired legally-binding force, the ECJ has restated that ‘according to Article 51(1) of the Charter, its provisions are addressed to the Member States only when they are implementing European Union law’. See Case C-400/10 PPU J McB v LE (ECJ, 5 October 2010), para 51. 34 See the cases commented upon by Eeckhout, above n 12, 959–62; Lenaerts, above n 4, 148–49. 35 JI Ugartemendía and D Sarmiento, ‘Los derechos fundamentales que acompañan al ciudadano europeo en el ejercicio de las libertades de circulación’ in J Alberdi Bidaguren and J Goizueta Vertiz (eds), Algunos retos de la inmigración en el siglo XXI (Barcelona, Marcial Pons, 2008) 117–36. 36 Case C-274/96 Host Otto Bickel and Ulrich Franz [1998] ECR I-7637. 31

116 Aida Torres Pérez The ECJ responded that although the rules of criminal procedure were matters under state power, Community law set certain limits to this power: ‘Such legislative provisions may not discriminate against persons to whom Community law gives the right to equal treatment or restrict the fundamental freedoms guaranteed by Community law’.37 Hence, the ECJ concluded that, since state provisions concerning the language to be used in criminal proceedings might compromise the right of nationals of other Member States to equal treatment in the exercise of their freedom to move and reside in another Member State, those state provisions had to comply with the principle of non-discrimination on the basis of nationality. If the mere fact of moving from one State to another is considered to be enough connection with EU law to enforce the right to non-discrimination on grounds of nationality as regards state action, what about other EU fundamental rights?38 Arguably, in order to secure free movement, ‘moving citizens’ should enjoy a common set of fundamental rights, regardless of whether state authorities act within the field of EU law or not.39 Indeed, the Charter might provide the set of common rights that EU citizens should be able to invoke in all Member States. This would be a way to give content to the concept of European citizenship. The ECJ, however, has not followed this line of argument that would have extended the scope of EU fundamental rights to any moving European citizen vis-à-vis state public authorities. As Eeckhout pointed out, should this extension be admitted, the next question would be: Why should only ‘moving citizens’ be entitled to claim Charter rights before state authorities regardless of the field of application of EU law, as opposed to EU citizens who never moved?40 In other words, once the extension of EU rights to moving citizens were admitted, it would be more and more difficult to deny to the Member States’ own citizens the protection afforded to citizens coming from others in order to avoid situations of reverse discrimination.41 ii. EU Citizenship In this context, we should emphasise the recent ECJ judgment in Ruiz Zambrano,42 which shows the potential extension of EU fundamental rights to ‘purely internal situations’ on the basis of the status of EU citizenship.

37

Ibid para 17. Eeckhout, above n 12, 969–71. 39 Along these lines, AG Jacobs suggested that any European citizen who develops an economic activity in another Member State should be treated in accordance with a common code of fundamental values, in particular those laid down in the European Convention on Human Rights. See Case C-168/91 Konstantinidis [1993] ECR I-1191, Opinion of AG Jacobs, para 46. 40 Eeckhout, above n 12, 972. 41 D Sarmiento, ‘Discriminaciones inversas comunitarias y Constitución española’, available at . 42 Case C-34/09 Ruiz Zambrano v Office national de l’emploi (ECJ, 8 March 2011). 38

The Dual System of Rights Protection in the European Union 117 Briefly, the facts of the case refer to a Colombian national living in Belgium, without the corresponding residence and working permits. While living in Belgium, he had two children who acquired Belgian nationality. In a nutshell, the question before the Court was whether Mr Ruiz Zambrano enjoyed the rights to reside and work in Belgium as a derivation of his children’s citizenship rights, in a context in which free movement was not at stake. In earlier cases, such as Carpenter43 or Zhu and Chen,44 the exercise of free movement45 by EU citizens46 was the link with EU law, tenuous as it might have been, that justified the recognition of the right of residence of third country nationals. The Ruiz Zambrano decision was preceded by a suggesting Opinion by Advocate General Sharpston. Since we cannot develop a full analysis of her Opinion here, we shall focus on her proposal for the definition of the scope of application of EU fundamental rights. Given the difficulties for determining the boundaries of the field of EU law, the recourse to ‘fictitious’ links with EU law and the problems of reverse discrimination, Sharpston argued that EU fundamental rights should apply to the States in all areas under EU powers, whether exclusive or shared, even if such powers had not yet been exercised.47 At the same time, the Advocate General did not recommend the adoption of that interpretation in the specific case. She acknowledged that this interpretation would amount to introducing ‘an overtly federal element into the structure of the EU’s legal and political system’. It would therefore require further evolution in the case law and an ‘unequivocal political statement from the constituent powers of the EU (its Member States), pointing at a new role for fundamental rights in the EU’.48 Sharpston compared the federalising effects of her proposal with the incorporation of the federal Bill of Rights in the US. It should be noted, however, that her proposal would not lead to a ‘full incorporation’. In the US, federal rights apply to the states irrespective of the allocation of powers system. According to Sharpston’s proposal, the incorporation would still be ‘half-way’, although the scope of application of EU fundamental rights to the States would be further extended. As a result, even state action tenuously related to EU law would be bound by EU fundamental rights, if enacted within the field under EU powers. This option,

43 Case C-60/00 Mary Carpenter and Secretary of State for the Home Department [2002] ECR I-6279. 44 Case C-200/02 Kunqian Catherine Zhu and Man Lavette Chen v Secretary of State for the Home Department [2004] ECR I-9925. 45 Free provision of services in Carpenter (above n 43; ex Art 49 TEC) and free movement of people in Zhu and Chen (above n 44; ex Art 18 TEC). 46 The husband and the daughter respectively. 47 Ruiz Zambrano, above n 42, Opinion of AG Sharpston, para 163: ‘The clearest rule would be one that made the availability of EU fundamental rights protection dependent … on the existence and scope of a material EU competence. To put the point another way: the rule would be that, provided that the EU had competence (whether exclusive or shared) in a particular area of law, EU fundamental rights should protect the citizen of the EU even if such competence has not yet been exercised.’ 48 Ibid para 171.

118 Aida Torres Pérez however, would not necessarily resolve situations of reverse discrimination in the remaining areas of residual powers of the States. In contrast to the richness of the Advocate General’s Opinion, the Court’s reasoning was rather laconic. The ECJ did not even mention the Charter, although the question raised by the domestic court included Articles 21, 24 and 34 of the Charter. The ECJ limited its judgment to Article 20 TFEU, and it took a big step to give effect to EU citizenship. First, the ECJ acknowledged that it is for the Member States to determine the requisites for acquiring nationality. To the extent that Mr Ruiz Zambrano’s children had acquired Belgian nationality, Article 20 TFEU granted them EU citizenship, and thus they were entrusted with all rights associated with that status.49 Next, the ECJ held that ‘Article 20 TFEU precludes national measures which have the effect of depriving citizens of the Union of the genuine enjoyment of the substance of the rights conferred by virtue of their status as citizens of the Union’.50 The Court argued that the refusal to grant the right of residence and a working permit to a third country national whose dependent minor children are EU citizens in the Member State in which they reside would have such an effect.51 In other words, if the father, as the caregiver, were not allowed to stay in the country legally, the children would also be forced to leave. Hence, the effectiveness of the rights attached to EU citizenship, mainly the children’s right of residence, involves the conferral of the right of residence on the father. In this case, neither the father nor the children had moved to any other Member State. Thus, the ECJ required the State to recognise the right of residence of a third country national as deriving from his children’s citizenship rights in what seemed to be a ‘purely internal situation’. By disentangling the right of residence from free movement, the ECJ took a daring step towards extending the scope of application of EU rights to the States. The controlling clause of the decision is rather vague, since the ECJ refers generally to the need to secure the ‘genuine enjoyment of the substance of the rights attaching to the status of European Union citizen’.52 Besides the diffuse meaning of ‘genuine enjoyment’, we might wonder what the rights attaching to the status of EU citizenship are (and what their substance is). Only those listed in Article 20 TFEU? Indeed, according to Article 20(2) this is not a closed list, since it is announced using the form ‘inter alia’.53 Hence, as a hypothetical, if the rights included in Article 20 TFEU must be protected by the States disconnected from free movement, why not all the rights included in the Charter, as attached to the status of EU citizenship?

49

Ruiz Zambrano, above n 42, para 40. Ibid para 42. 51 Ibid para 43. 52 Ibid para 45. 53 Art 20(2) TFEU: ‘Citizens of the Union shall enjoy the rights and be subject to the duties provided for in the Treaties. They shall have, inter alia: …’ (emphasis added). 50

The Dual System of Rights Protection in the European Union 119 The main goal of drafting the Charter was to make EU fundamental rights more visible to the citizens.54 The Charter’s Preamble indicates that the Union ‘places the individual at the heart of its activities, by establishing the citizenship of the Union’. It might well be argued that the status of Union citizen includes all the rights enshrined in the Charter and that all should be equally protected. Although the Charter was not mentioned, the right of residence is also protected in Article 45, within Title V (Citizens’ rights). What would justify extending the scope of application to internal situations of the rights included in Title V, but exclude the rest, such as the right to private and family life (Article 7 Charter) or the rights of the child (Article 24 Charter)? In this context, a binding Charter might promote the extension of the scope of application of EU fundamental rights to the Member States. Ruiz Zambrano seemed to slightly open a door for full incorporation. The ECJ, however, qualified the reach of Ruiz Zambrano in McCarthy.55 Mrs McCarthy was a dual Irish–UK national, who was born and always lived in the UK, and who was in receipt of state benefits. She married a Jamaican national who lacked the necessary permit to remain in the UK. Following their marriage, Mrs McCarthy and her husband applied for a residence permit and a residence document under EU law as, respectively, a Union citizen and the spouse of a Union citizen. Their applications were refused and they appealed. Eventually, a preliminary reference was raised before the ECJ. The Court confirmed that Article 3(1) of Directive 2004/38/EC on the right of citizens of the Union and their family members to move and reside freely within the territory of the Member States56 was not applicable to a Union citizen who had never exercised her right of free movement, who always resided in a Member State of which she was a national, and who was also a national of another Member State.57 In these circumstances, the ECJ considered whether Article 21 TFEU was applicable to the case. It is worth focusing on how the ECJ conceives of a ‘purely internal situation’. The Court held that ‘the situation of a Union citizen who, like Mrs McCarthy, has not made use of the right to freedom of movement cannot, for that reason alone, be assimilated to a purely internal situation’.58 Thus, if free movement is not involved, what would be the link with EU law? The answer is implied in the next paragraph (expressly quoting Ruiz Zambrano): ‘EU citizenship,’ and

54 Annex X. European Council Decision on the Drawing up of a Charter of Fundamental Rights of the European Union, Presidency Conclusions, European Council, Cologne, 3–4 June 1999: ‘Protection of fundamental rights is a founding principle of the Union and an indispensable prerequisite for her legitimacy. The obligation of the Union to respect fundamental rights has been confirmed and defined by the jurisprudence of the European Court of Justice. There appears to be a need, at the present stage of the Union’s development, to establish a Charter of fundamental rights in order to make their overriding importance and relevance more visible to the Union’s citizens.’ 55 Case C-434/09 Shirley McCarthy v Secretary of State for the Home Department (ECJ, 5 May 2011). 56 Directive 2004/38/EC on the right of citizens of the Union and their family members to move and reside freely within the territory of the Member States [2004] OJ L158/77. 57 McCarthy, above n 55, paras 39–40. 58 Ibid para 46.

120 Aida Torres Pérez particularly the need to protect the ‘genuine enjoyment of the substance of the rights conferred by virtue of that status’.59 There was no other link with EU law. The ECJ concluded: As a national of at least one Member State, a person such as Mrs McCarthy enjoys the status of a Union citizen under Article 20(1) TFEU and may therefore rely on the rights pertaining to that status, including against his Member State of origin, in particular the right conferred by Article 21 TFEU to move and reside freely within the territory of the Member States.60

Therefore, EU citizenship confers rights on individuals vis-à-vis Member States (including the State of nationality and residence), regardless of the exercise of free movement or any other connection with EU law.61 Until this point, the ECJ followed the reasoning of Ruiz Zambrano. Next, however, the Court distinguished McCarthy from Ruiz Zambrano, thereby limiting the potential reach of the latter: ‘[B]y contrast with the case of Ruiz Zambrano, the national measure at issue in the main proceedings in the present case does not have the effect of obliging Mrs McCarthy to leave the territory of the European Union.’62 And yet, was the situation of Mrs McCarthy that different? Arguably, the denial of the residence permit to her husband would not have the effect of ‘obliging’ Mrs McCarthy to leave the territory of the EU—she could always stay in the UK and have a long-distance relationship with her husband! But would the ‘genuine enjoyment’ of her rights as a Union citizen be preserved if her husband were forced to leave? On the other hand, would Ruiz Zambrano’s children have been ‘obliged’ to leave the territory of the Union? The children could have stayed in Belgium in a foster home. Obviously, this is not the most convenient from the perspective of the children’s best interests. But then, what is at stake is the right to family life and the rights of the child, rather than the right of residence. The Court, however, did not even mention these Charter rights because that would have meant obliterating article 51(1) Charter. The deciding element in Ruiz Zambrano was the unwanted consequences for the dependent children resulting from denying the residence permit to the father.

59 Ibid para 47: ‘Indeed, the Court has stated several times that citizenship of the Union is intended to be the fundamental status of nationals of the Member States (see Case C-34/09 Ruiz Zambrano [2011] ECR I-000, paragraph 41 and case law cited). Furthermore, the Court has held that Article 20 TFEU precludes national measures which have the effect of depriving Union citizens of the genuine enjoyment of the substance of the rights conferred by virtue of that status (see Ruiz Zambrano, paragraph 42).’ 60 McCarthy, above n 55, para 48. 61 This conclusion clashes with previous case law. In Schempp, also quoted in McCarthy, the ECJ had held that ‘citizenship of the Union, established by Article 17 EC, is not intended to extend the material scope of the Treaty to internal situations which have no link with Community law’ (Case C-403/03 Egon Schempp v Finanzamt München V [2005] ECR I-6421, para 20). 62 McCarthy, above n 55, para 50.

The Dual System of Rights Protection in the European Union 121 To conclude, Ruiz Zambrano took a groundbreaking step towards the protection of citizenship rights in ‘no cross-border situations’. The main concern, however, seems to have been the well-being of minor children. Hence, the ECJ formulated a principle with far-fetching implications in a result-orientated decision, which is very problematic for the consistency of case law in the long run. Indeed, in McCarthy, the ECJ aimed at containing those implications, but at the risk of undermining coherence. On the whole, the scope of application of EU fundamental rights to the States refers to a more general issue about the extent to which state autonomy to interpret fundamental rights and develop public policies accordingly should be preserved. To the extent that fundamental rights express the main values of political communities, that States are still a better locus for democratic participation and contestation, and that there is room for ‘reasonable disagreement’ on rights interpretation, there are reasons to secure some degree of state autonomy. In sum, even if—in the long run—the incorporation of Charter rights into the States were extended further, some room for State autonomy and diversity in rights interpretation should be preserved. As will be shown below, in the US, full incorporation has not obliterated the possibility for state diversity in rights’ interpretation.

III. OVERLAPPING EU AND CONSTITUTIONAL RIGHTS

When State authorities act within the field of application of EU law, blurring as it may be, they are bound at the same time by the constitution and the Charter. Conflicts might arise when parallel rights are interpreted differently. The ECJ claims authority to interpret rights that will be used to review state action, sometimes in opposition to the constitutional interpretation of a parallel right. To some extent, state courts are being asked to replace the constitutional interpretation with the one defined by the ECJ. These potential conflicts cannot be solved automatically on the basis of the primacy principle, which in this context would involve the primacy of the Charter over state constitutions. Although domestic courts have admitted the primacy of EU law over ordinary law, constitutional courts have contested the primacy of EU law over the constitution itself or set limits to it.63 In this respect, Article 53 of the Charter lays down: Nothing in this Charter shall be interpreted as restricting or adversely affecting human rights and fundamental freedoms as recognised, in their respective fields of application, … by the Member States’ constitutions.

63 See the national reports in AM Slaughter, A Stone Sweet and JHH Weiler (eds), The European Court and National Courts, Doctrine and Jurisprudence: Legal Change in its Social Context (Oxford, Hart Publishing, 2000); G Martinico and O Pollicino, The National Judicial Treatment of the ECHR and EU Laws (Groningen, Europa Law Publishing, 2010).

122 Aida Torres Pérez Still, the interpretation of this article is disputed, and it is not clear what the consequences are: is the primacy of EU law weakened, or is this article ‘just an inkblot’?64 Given the similarity of this provision with clauses codified in international human rights treaties,65 one might think that it should be interpreted so that the Charter sets a floor of protection that the States may exceed . This interpretation will be evaluated from the standpoint of the US model. Indeed, in the US, federal constitutional rights have been regarded as comprising a floor of protection. When reviewing state legislative or executive action, state courts may interpret state constitutional rights as providing a broader scope of protection than the federal constitution. In the words of Stewart G Pollock, Justice of the Supreme Court of New Jersey, the Bill of Rights in the United States Constitution establishes a floor for basic human liberty. To carry forward the metaphor, the state constitution establishes a ceiling. Although a state may supplement federally granted rights, it may not diminish them through a more restrictive analysis of the state or federal constitution.66

A. The ‘New Judicial Federalism’ and the Federal Floor of Protection During the early 1970s, the so-called ‘new judicial federalism’ emerged in the US, with reference to state courts’ increased reliance on state declarations of rights to review state action. In two famous articles, the US Supreme Court Justice Brennan encouraged state courts to rely on state constitutions civil rights provisions above the federal floor of protection: I believe that the Fourteenth Amendment fully applied the provisions of the Federal Bill of Rights to the States, thereby creating a federal floor of protection and that the Constitution and the Fourteenth Amendment allow diversity only above and beyond this federal constitutional floor … This rebirth of interest in state constitutional law should be greeted with equal enthusiasm by all those who support our federal system.67

Although the supremacy of the federal constitution and federal law over state constitutions is explicitly articulated in the federal constitution (Article VI), this does not mean that the federal Bill of Rights will always apply in case of diverging interpretations of parallel rights. It should be recalled that, originally, the US Bill of Rights was not binding upon the states. Thus, the supremacy clause did not speak to conflicts between federal and state constitutional rights because there was no

64 JB Liisberg, ‘Does the EU Charter of Fundamental Rights Threaten the Supremacy of Community Law? Article 53 of the Charter: A Fountain of Law or Just an Inkblot?’ Jean Monnet Working Paper no 4/01. 65 Eg, European Convention on Human Rights, Art 53. 66 SG Pollock, ‘State Constitutions as Separate Sources of Fundamental Rights’ (1983) 35 Rutgers Law Review 707, 709. 67 Brennan, above n 8, 550; WJ Brennan, ‘State Constitutions and the Protection of Individual Rights’ (1977) 90 Harvard Law Review 489.

The Dual System of Rights Protection in the European Union 123 overlap in their scope of application. After the Civil War and the enactment of the Fourteenth Amendment, federal rights were progressively incorporated into the states by the Supreme Court. State courts were then confronted with the dilemma of applying state or federal constitutional rights when reviewing state measures, particularly if the respective interpretations differed. An absolute interpretation of the supremacy clause would answer that federal constitutional rights should always prevail. Following Justice Brennan, however, it is admitted that should the state interpretation be more protective, state courts would be free to enforce state constitutional rights. In the 1990s, the fact that state courts might interpret state constitutional rights to provide more extensive protection than the US Supreme Court under the federal constitution had been generally accepted, albeit still sometimes controversial.68 From the outset, the legitimacy of the ‘new judicial federalism’ had been contested from the perspective of the relationship between state courts and the US Supreme Court. The question was: To what extent are state courts justified in interpreting state constitutional rights differently from the US Supreme Court’s interpretation of parallel federal constitutional rights?69 The turn towards state constitutions was regarded as result-orientated, after the conservative shift of the Court presided by Justice Burger.70 Critics claimed that state courts should defer to the Supreme Court’s interpretation of analogous provisions.71 Deference to previous US Supreme Court case law was supported on the basis of the value of uniformity and the position of this Court as the highest court in the land.72 Eventually, some even denied the possibility and appropriateness of independent state constitutional interpretation.73 At the same time, the lead of Justice Brennan was followed by several state judges and scholars,74 who claimed for the legitimacy of the ‘new judicial federalism’. Several approaches to the interaction between state courts and the US Supreme Court in interpreting state constitutional rights were formulated, basically: the ‘lockstep’, the ‘supplemental or interstitial’ and the ‘primacy’ approach.75

68 RF Williams, ‘In the Glare of the Supreme Court: Continuing Methodology and Legitimacy Problems in Independent State Constitutional Rights Adjudication’ (1997) 72 Notre Dame Law Review 1015, 1016. 69 A Tarr, Understanding State Constitutions (Princeton, NJ, Princeton University Press, 1998) 175. 70 JA Gardner, ‘The Failed Discourse of State Constitutionalism’ (1992) 90 Michigan Law Review 761, 771–73; Tarr, above n 69, 177–80. 71 Tarr, above n 69, 175. 72 Ibid 175. 73 Gardner, above n 70. 74 Eg Pollock, above n 66, 709; HE Linde, ‘E Pluribus. Constitutional Theory and State Courts’ (1984) 18 Georgia Law Review 165, 179; DE Wilkes, Jr, ‘First Things Last: Amendomania and State Bills of Rights’ (1984) 54 Mississippi Law Journal 223; RF Williams, ‘In the Supreme Court’s Shadow: Legitimacy of State Rejection of Supreme Court Reasoning and Result’ (1984) 35 Supreme Court Law Review 353; Friedman, above n 8. 75 Tarr, above n 69, 180–85.

124 Aida Torres Pérez According to the ‘lockstep’ approach, the state courts’ interpretation should always follow the US Supreme Court case law. This approach banishes legitimacy concerns, but it makes state rights guarantees superfluous.76 According to the ‘supplemental or interstitial’ approach, state courts should first consider fundamental rights under the federal constitution. It would be justified for state courts to deviate from federal case law if certain criteria were fulfilled, such as textual differences, ‘legislative history’ of the state constitutional provision, or differences in federal and state structures.77 State courts would not simply reproduce the approach of the US Supreme Court, but they would be able to develop an independent analysis of the state constitution.78 This approach, however, has been criticised, because in the long run it would hinder the development of independent state constitutional interpretation.79 In addition, it has been argued that Supreme Court decisions cannot have presumptive validity for state constitutional interpretation, since federalism concerns may cause the Supreme Court to limit the scope of federal constitutional rights when applied against the states.80 The ‘primacy’ approach defends independent state interpretation according to state sources.81 According to this approach, state courts should look to the state constitution in the first place, and interpret it taking into account independent state sources, such as the text, history and structure of the document.82 Admittedly, if state constitutional provisions proved less protective than their federal counterparts, federal law should prevail. Few state courts have consistently adopted a primacy approach.83 All these approaches admit that the federal constitution sets a floor of protection and differ on how to proceed regarding the interpretation and application of state constitutions. Broadly, the values of the new judicial federalism are defended from the perspective of the commitment to dual constitutionalism, which originated to provide a ‘double security’ for rights.84 Along these lines, state independent interpretation has been defended from the perspective of checks and balances and the value of dialogue.85

76

Ibid 180–81. Friedman, above n 8, 103; Williams, above n 74, 385–89. 78 Pollock, above n 66, 718–19. 79 Williams, above n 74, 385. 80 Ibid 385–402. 81 Linde, above n 74, 178–79: ‘The right question is not whether a state’s guarantee is the same as or broader than its federal counterpart as interpreted by the Supreme Court. The right question is what the state’s guarantee means and how it applies to the case at hand.’ 82 Ibid 181–93. 83 Tarr, above n 69, 184. 84 Ibid 181. 85 Friedman, above n 8, 97. 77

The Dual System of Rights Protection in the European Union 125 B. Rights Conflicts and Judicial Dialogue in the EU To what extent might the US debate on the ‘new judicial federalism’ illuminate potential conflicts of rights in the EU? At the outset, it should be noted that, in the EU, legitimacy concerns are of a different kind. No one questions the legitimacy of constitutional courts for the independent interpretation of the respective constitution, according to domestic sources. Thus, arguably, in reviewing state action, constitutional courts should be free to interpret rights at a higher level of protection than the ECJ. In other words, the ECJ would set a floor of protection that state courts might exceed. Given the particular dynamics of the EU legal order, however, adopting the ‘floor-of-protection rule’ without further restrictions might undermine the uniformity and efficacy of EU law. It should be noted that, in the EU legal order, the enforcement of EU law within the States is left, to a great extent, to state authorities. Therefore, if state courts were absolutely free, under a more protective constitutional standard, to review and set aside state acts implementing EU law, the uniformity and efficacy of EU law within the territory of each Member State might be undermined. For example, if state courts unilaterllay set aside state legislation implementing the framework decision regulating the European Arrest Warrant86 to ensure constitutional rights protection, the efficacy of this mechanism could be jeopardised. Even though the efficacy argument does not justify the absolute primacy of EU law, it needs to be addressed. The same kind of concern has not arisen in the US, because setting aside state acts under a more protective rights’ interpretation does not necessarily impinge upon the efficacy of federal law. The case of the Hyde Amendment, for example, is very illustrative.87 The Hyde Amendment, enacted by Congress in 1976, banned the use of Medicaid funds for abortion, except when carrying the pregnancy to term would endanger the mother’s life, or when the pregnancy was the result of rape or incest. Several states passed similar statutory provisions. In 1977, the US Supreme Court upheld several state statutes restricting the use of public funds for abortion. Later on, in 1980, the Supreme Court also upheld the federal statutory limitations contained in the Hyde Amendment in Harris v McRae.88 After this decision, supporters of public funding for abortions initiated litigation in state 86 Council Framework Decision 2002/584/JHA on the European arrest warrant and the surrender procedures between Member States [2002] OJ L190/1. For conflicts regarding the EAW, see J Komárek, ‘European Constitutionalism and the European Arrest Warrant: In Search of the Limits of “Contrapunctual Principles” ’ (2007) 44 CML Rev 9; D Sarmiento, ‘European Union: The European Arrest Warrant and the Quest for Constitutional Coherence’ (2008) 6 International Journal of Constitutional Law 171; E Cloots, ‘Germs of Pluralist Judicial Adjudication: Advocaten voor de Wereld and Other References from the Belgian Constitutional Court’ (2010) 47 CML Rev 645; A Torres Pérez, ‘Euroorden y conflictos constitucionales: A propósito de la STC 199/2009’ (2010) 35 Revista Española de Derecho Europeo 441. 87 See Williams, above n 74, 362–74. 88 Harris v McRae, 488 US 297 (1980).

126 Aida Torres Pérez courts and challenged state statutes on state constitutional grounds. Several state courts struck down state funding restrictions on grounds of due process or equal protection, according to the respective state constitution.89 State courts tended to order continuation of state funding under medical assistance programmes, refusing the result and reasoning of the Supreme Court in Harris v McRae. Hence, some state courts granted a higher level of protection than the US Supreme Court to poor pregnant women on the basis of a divergent interpretation of the due process and equal protection clauses. The decision to strike down the respective state legal provisions did not damage the efficacy of federal law. The Hyde Amendment merely banned the use of federal funds for abortion. States were not required to restrict funding, but they were allowed to do it according to the federal interpretation of constitutional rights.90 Thus, state court judgments declaring that such restrictions violated state constitutional rights diverged from the federal interpretation of parallel rights. Yet the efficacy of federal law was not impinged upon, since only state public funds were affected by these state court judgments. Indeed, on occasions, the US Supreme Court has acknowledged the possibility for state courts to give more extensive protection to state constitutional rights, such as free speech. The First Amendment protects free speech only against governmental action. Accordingly, the US Supreme Court had ruled that the owners of shopping centres could restrict or ban speech on their private property.91 In contrast, in Robins v PruneYard Shopping Center,92 the California Supreme Court upheld the right of high school students to distribute pamphlets and collect signatures in a privately-owned shopping centre to seek support for their opposition to a United Nations resolution against ‘Zionism’. The California Supreme Court argued that, in some circumstances, free speech may override the claims of property owners. It pointed out that the activity of the high school students did not interfere with the normal business operations of the mall. The owner of the shopping centre filed an appeal before the US Supreme Court. In PruneYard Shopping Center et al v Robins et al, the US Supreme Court unanimously upheld the California judgment. The US Supreme Court argued that its previous case law did not ‘limit the authority of the State to exercise its police power or its sovereign right to adopt in its own Constitution individual liberties more expansive than those conferred by the Federal Constitution’.93 A general problem for the application of the floor-of-protection rule is that, in many instances, it is not clear what the best level of protection is, particularly when different rights conflict. The PruneYard example shows that the better protection for free speech actually involved a worse protection of the right to

89

For the case of New Jersey, see Pollock, above n 66, 719. Williams, above n 74, 379. 91 GA Tarr, ‘State Constitutionalism and “First Amendment” Rights’ in SH Friedelbaum (ed), Human Rights in the States (Westport, Conn, Greenwood Press, 1988) 36. 92 Robins v PruneYard Shopping Center, 23 Cal 3d 899, 592 P 2d 342, 153 Cal Rptr 854 (1979). 93 PruneYard Shopping Center et al v Robins et al, 447 US 74; Pollock, above n 66, 710. 90

The Dual System of Rights Protection in the European Union 127 property. In other situations, such as in the abortion funding cases, it is clear that the states provided a higher level of protection to pregnant women than did the federal Supreme Court. As indicated above, in the EU, in addition to the difficulties with applying this conflict rule when various rights are at stake, there is a further problem from the perspective of the efficacy and uniform application of EU law. Even when it is clear that the state interpretation is more protective, if state courts were free to make a more protective constitutional right prevail over a parallel EU right in reviewing state action implementing EU law, the efficacy of EU law might be endangered. That is why the floor-of-protection rule would be self-defeating, since it would not account for the interest in the uniform application and efficacy of EU law. Hence, does this mean that Member State courts should renounce the constitutional level of protection when it conflicts with the interpretation of the Court of Justice? Arguably, integration does not automatically justify giving up constitutional protection. Since the constitutional courts’ legitimacy for an independent interpretation of constitutional rights is not questioned, the arguments on the basis of dual constitutionalism and dialogue developed in the US might be specially well-suited to the EU. From the perspective of dual constitutionalism, there are benefits in having national and supranational courts interpreting rights, instead of entrusting this responsibility to an ultimate judicial authority. Courts might have different approaches to the law as a consequence of their respective institutional roles and their position in the legal system.94 One of the main functions of constitutional courts is precisely the protection of rights and the review of legislation in light of fundamental rights. In contrast, the ECJ is a court of general jurisdiction. It was set up to secure that ‘in the implementation of the common market the law is observed’,95 and it might possess a bias favouring the success of the integration process. For instance, EU law’s efficacy or the protection of the EU freedoms of movement might shape the ECJ’s approach to fundamental rights adjudication. The claim here is not that constitutional courts will always afford better protection than the ECJ, but that there are institutional reasons to think that constitutional courts can act as a counterbalance to the ECJ. Indeed, the so-called rebellion of the German and Italian constitutional courts fostered the protection of rights at the supranational level. On the whole, the benefits of courts from different levels of governance protecting rights might be appreciated from the perspective of a ‘dual system of protection’.96 National and supranational participation in the interpretation of rights contributes to limit the excesses and detect the shortcomings of each other, eventually furthering individual protection. 94 95 96

Torres Pérez, above n 21, 118–20. Ex Art 220 TEC (previously, Art 164 EEC Treaty), derogated by the Lisbon Treaty. Torres Pérez, above n 21, 86–89.

128 Aida Torres Pérez Still, there is a potential for conflict as a consequence of overlapping rights that needs to be addressed. Instead of trying to resolve potential conflicts claiming for hierarchy between norms or courts from different systems, we might think of the interaction between national and supranational courts in terms of dialogue. Dialogue is understood here as an exchange of arguments aimed at reaching common understandings regarding the interpretation of fundamental rights.97 Obviously, the ECJ cannot enter into simultaneous dialogue with all state courts until a consensus is reached. Dialogue needs to be conceived as evolving over time, diachronically, rather than synchronically. Judicial dialogue in the EU will develop in a fragmented manner, since the exchange of arguments among national and supranational courts occurs case by case. There is always the opportunity to refine previous interpretations in light of better arguments.98 Thus, dialogue does not provide a solution in advance, but it manages potential conflicts in a process of mutual accommodation over time. Dialogue has been mostly channelled through the preliminary reference procedure. This mechanism, however, has not been used to its fullest potential. The preliminary reference offers state courts the opportunity to engage in robust dialogue with the ECJ, beyond the mere formulation of a question. They might provide information about the interpretation of a specific right and the underlying reasons supporting it. The preliminary reference might be an avenue to signal a potential conflict with constitutional rights. State courts could offer reasons to follow a certain interpretation at the EU level, or claim for deference towards the constitutional interpretation. In the end, this is a way of contributing to define the content of supranational rights that will be applied to review the action of state authorities. Constitutional courts have been reluctant to raise preliminary references.99 This reluctance might be explained in terms of a desire to preserve their autonomy and a resistance to being placed under the ECJ’s interpretive authority. This does not mean, however, that constitutional courts have not entered into dialogue with the ECJ, but they have preferred some forms of indirect or ‘hidden dialogue’.100 For instance, the German federal constitutional court has never raised a preliminary reference, but has engaged in a dialogue with the ECJ, even though the tone has been defiant at times. On the other hand, to foster dialogue, the ECJ should adopt a more discursive style and be more responsive to the claims and arguments by state courts

97

Ibid 109–10. Ibid 110–12. 99 Until now, only the Belgian, Austrian, Lithuanian, Italian and recently Spanish constitutional courts have raised preliminary references to the ECJ. The German constitutional court has declared itself to be willing to do so (BVerfG, 2 BvR 2661/06, 6 July 2010). 100 G Martinico, ‘Judging in the Multilevel Legal Order: Exploring the Techniques of “Hidden Dialogue” ’ (2010) 21 King’s Law Journal 257. 98

The Dual System of Rights Protection in the European Union 129 regarding the interpretation of fundamental rights.101 The ECJ should show how arguments coming from state courts are weighted in order to reach a specific interpretive outcome. The persuasiveness of its decisions is important for the legitimacy of its adjudicative function.102 In addition, the ECJ should consider the possibility of accommodating diversity by deferring the interpretation of fundamental rights to state courts.103 Accommodating a degree of diversity in the interpretation of rights does not always significantly undermine the efficacy of EU law, let alone the success of the entire integration project. Deference could be an instrument for the ECJ to allow for higher constitutional standards of protection, without imposing that standard on the other States, such as in Omega.104 Also, the ECJ might defer the application of the proportionality test to decide whether a restriction of free movement on the basis of rights protection is justified, such as in Familiapress.105

IV. CONCLUDING REMARKS

The US model for the interaction between the federal Government and the states in the field of fundamental rights offers illuminating insights for the EU, but it cannot be transposed automatically. The main lesson from the US may well be the commitment to a dual system of rights protection. The evolution of the US Bill of Rights throughout the process of federation-building shows the ‘federalising force’ of a binding declaration of rights. Fundamental rights in the EU have been incorporated into the States only ‘halfway’, but the process of incorporation has been steadily expanding. A binding Charter may well promote an extension of the scope of application of EU rights to the States. More and more, constitutional and EU rights overlap in their scope of application to the States. For structural reasons, the floor-of-protection rule cannot be applied automatically in the EU to resolve potential conflicts, since it might undermine the efficacy and uniform application of EU law. At the same time, the commitment to European integration does not in itself justify giving up constitutional protection and homogenising fundamental rights. There are reasons to preserve some degree of state autonomy for the interpretation of constitutional rights and to accommodate a certain degree of diversity.

101 M Cartabia, ‘Europe and Rights: Taking Dialogue Seriously’ (2009) 5 European Constitutional Law Review 5, 29–30. 102 JHH Weiler, ‘Epilogue: The Judicial Après Nice’ in G de Búrca and JHH Weiler (eds), The European Court of Justice (New York, Oxford University Press, 2001) 225; LR Helfer and AM Slaughter, ‘Toward a Theory of Effective Supranational Adjudication’ (1997) 107 Yale Law Journal 273, 321–22. 103 Torres Pérez, above n 21, 168–77. 104 Case C-36/02 Omega Spielhallen und Automatenayfstellungs GmbH v Oberbürgermeisterin der Bundesstadt Bonn [2004] ECR I-9609. 105 Familiapress, above n 25.

130 Aida Torres Pérez State autonomy relates to the principle of self-government of democratic polities. By creating and joining the EU, the Member States surrendered their sovereign powers over certain subject matters conferred upon the EU. Nonetheless, they preserved their powers to interpret rights constraining state action. While the protection of rights at the supranational level must be celebrated, this does not mean that constitutional rights are no longer effective. Granting legally-binding force to the EU Charter of Fundamental Rights does not make constitutional rights catalogues superfluous, as the Charter acknowledges in Article 53. Indeed, in the US, full incorporation has not obliterated the possibility for state diversity. In sum, given the EU political structure, the values of a dual protection system and dialogue are reinforced. National and supranational courts should collaborate in giving meaning to fundamental rights, and counterbalance each other in a process of constant mutual accommodation.

6 Federalism and International Relations in the European Union and the United States: A Comparative Outlook GEERT DE BAERE* AND KATHLEEN GUTMAN**

I. INTRODUCTION

A

T PRESENT, THE comparative study of federalism in the European Union (EU) and the United States (US) continues to generate a substantial volume of literature on both sides of the Atlantic, but this is markedly less so when it comes to the field of international (foreign or external) relations.1 As illustrated by past literature, this field highlights the vast differences between the EU and the US that often make it difficult to find much similarity in approach and raise questions as to whether they are too far apart for workable comparison.2 Recently, however, commentators are beginning to pay tribute to the potential comparative insights that can be gained through examination of this field.3 This coincides with the changing landscape regarding federalism and international relations in both legal orders.

* Assistant Professor of International Law and EU Law at the Faculty of Law and Senior Member at the Leuven Centre for Global Governance Studies, University of Leuven. ** Assistant Professor of European Law, University of Antwerp and Affiliated Senior Researcher at the Institute for European Law, University of Leuven. 1 On this chapter’s use of these terms, see section II.B. 2 See, eg, E Stein and L Henkin, ‘Towards a European Foreign Policy? The European Foreign Affairs System from the Perspective of the United States Constitution’ in M Cappelletti, M Seccombe and JHH Weiler (eds), Integration Through Law—Europe and the American Federal Experience, vol 1, book 3 (Berlin and New York, De Gruyter, 1986) 3. 3 See, eg, R Schütze, From Dual to Cooperative Federalism—The Changing Structure of European Law (Oxford, Oxford University Press, 2009) 108–22 and 288–343; D Halberstam, ‘The Foreign Affairs of Federal Systems: A National Perspective on the Benefits of State Participation’ (2001) 46 Villanova Law Review 1015. See also EA Young, ‘Protecting Member State Autonomy in the European Union: Some Cautionary Tales from American Federalism’ (2002) 77 New York University Law Review 1612, 1619, fn 22.

132 Geert De Baere and Kathleen Gutman As alluded to above, a ready explanation for the lack of substantial engagement with this subject stems from the perceived differences in federal structure between the EU and the US in the international relations field. The US is generally considered to exhibit features of a classic federal approach to international relations, in which plenary power or competence4 over international relations is vested in the federal government, not the states. In contrast, the EU is typically framed as a sui generis supranational entity that does not possess many of the capacities traditionally associated with international relations (eg it does not have an army or a navy, and it cannot declare war), and of the competences that it does have in this field, they are, for the most part, of a more limited nature in comparison with the competences held by the Member States, as exemplified by the intergovernmental approach taken to the Common Foreign and Security Policy (CFSP). In other words, at least prima facie, the US appears to be a good example of a federal system5 allowing internal diversity but projecting external unity, whereas the EU allows its internal diversity to be projected outwards in its international relations, thereby evading the confines of a federal system in the classical sense. Yet with ongoing developments taking place in the European and American legal orders, it may be wondered whether these prima facie impressions correspond to current constitutionality reality. In the EU, the entry into force of the Lisbon Treaty6 has changed the constitutional and institutional framework of international relations, repositioning the Union’s role and the scope of its competence in this field vis-à-vis the Member States.7 In the US, the increasing involvement of the states (and other sub-national entities)8 in matters relating to international relations has prompted intense debate concerning the purported federal monopoly over this domain and the extent to which varying types of actions taken by the states can be accommodated within the American constitutional framework.9 As a result, the subject of ‘international

4 In the US, reference is generally made to the ‘powers’ granted to the federal government in line with the wording of the US Constitution, whereas in the EU, reference is more often made to the ‘competences’ conferred on the Union, even though one finds both terms in the Treaties. As such, some commentators use them synonymously as regards the European setting (see A von Bogdandy and J Bast, ‘The Federal Order of Competences’ in Principles of European Constitutional Law, 2nd rev edn (Oxford, Hart Publishing, 2010) 275, 278), whereas others have drawn a theoretical distinction between the two in the discussion of EU external relations (see, eg, Schütze, above n 3, at 298–99 and citations therein). Without discounting the potential relevance of such a distinction, this chapter uses these two terms synonymously in reference to the ‘competences’ or ‘powers’ allocated to the Union and the federal government under the Treaties and the US Constitution, respectively. 5 On this chapter’s use of (con)federation, federal system and federalism, see section II.A. 6 Treaty of Lisbon amending the Treaty on European Union and the Treaty establishing the European Community [2007] OJ C306/1 (‘Lisbon Treaty’); consolidated version of the Treaty on European Union (‘TEU’ or ‘EU Treaty’) and the Treaty on the Functioning of the European Union (‘TFEU’ or ‘FEU Treaty’) [2010] OJ C83/1. 7 See generally section III. 8 See below n 122. 9 See generally section IV.

Federalism and International Relations in the EU and the US 133 relations federalism’ looms large in both the EU and the US, and calls for renewed investigation. The purpose of this chapter is therefore to explore the approach taken to federalism and international relations in the EU and the US, with primary focus placed on evaluating the extent to which comparative inquiry into this subject may be conducted despite differences in federal structure, on the one hand, and the allocation of competences between the central authority (the Union/federal government) and the component entities (the Member States/states) in the international relations field, on the other. To this end, the discussion is divided into three main parts. Section II sets forth the conceptual framework guiding the discussion, elucidating the concepts of federalism and international relations as applied in the EU and the US. Sections III and IV provide an overview of the constitutional framework governing international relations in the EU and the US, respectively, analysing the allocation of competences between the central authority and the component entities in this setting on the basis of the respective ‘constitutional’ texts (the Treaties/US Constitution) and relevant case law of the European Court of Justice10 and the US Supreme Court. To be clear, the analysis presented here does not claim to be comprehensive or exhaustive. Indeed, this chapter readily acknowledges the importance of not reducing federalism to the allocation of competences alone and of placing it in a broader context in order, for example, to examine the limits imposed by the central authority on the exercise of the competences held by the component entities.11 This chapter forms part of ongoing research and merely attempts to establish the foundations for understanding the conceptual and constitutional underpinnings of federalism and international relations in the EU and the US, in order to set the scene for delving further into the existence of tools or mechanisms employed in these two legal orders for managing the challenges posed by ‘international relations federalism’.

10 Under the first para of Art 19(1) TEU, the institution of the Court of Justice of the EU encompasses the Court of Justice, the General Court, and specialised courts (at present, the EU Civil Service Tribunal). For reasons of clarity, this chapter refers to the Court of Justice as the European Court of Justice (‘ECJ’) in the sense of the highest court of this institution. 11 See, eg, K Lenaerts, ‘Federalism and the Rule of Law: Perspectives from the European Court of Justice’ (2010) 33 Fordham International Law Journal 1338, 1339 (referring to the way in which EU law imposes negative limits on the powers retained by the Member States as the ‘pervasive effects of federalism’). In EU external relations, such an effect is, for instance, created by the principle of loyalty and sincere co-operation enshrined in Art 4(3) TEU, which imposes certain limits on the Member States’ exercise of their competence, most notably in mixed agreements (in which the EU and all or some Member States are parties alongside each other) or mixed participation in international organisations. A recent illustration is provided by Case C-246/07 Commission v Sweden (‘PFOS’) [2010] ECR I-3317, on which see G De Baere, ‘“O, Where is Faith? O, Where is Loyalty?” Some Thoughts on the Duty of Loyal Co-operation and the Union’s External Environmental Competences in the light of the PFOS Case’ (2011) 36 EL Rev 405.

134 Geert De Baere and Kathleen Gutman II. CONCEPTUAL FRAMEWORK OF FEDERALISM AND INTERNATIONAL RELATIONS IN THE EU AND THE US

A. Federalism in the EU and the US Notwithstanding the various conceptions of ‘federalism’ put forward in the commentary and jurisprudence,12 this chapter focuses on the concept of ‘federalism’ as the theory of principles for dividing competences between constituent member units and common institutions.13 It regards as a fundamental tenet underlying federalism that power will be divided between a central authority and the component entities of a nation-state or an international organization so as to make each of them responsible for the exercise of their own powers. By doing so, federalism searches for the balance between the desire to create and/or to retain an efficient central authority ... and the concern of the component entities to keep or gain their autonomy so that they can defend their own interests.14

Viewed in this way, federalism is not a form of organising a polity exclusive to nation-States like the US; it exists whenever the ‘search for unity’ is ‘combined with genuine respect for the autonomy and the legitimate interests of the participant entities’, as in the case of the EU.15 This is so, even though the EU and the US constitute different types of federal political systems. The term ‘federal political systems’ may be understood to denote political organisations marked by the combination of shared rule and self-rule, such as constitutionally decentralised unions, federations, confederations and otherwise associated states.16 As a species within the genus of federal systems, a federation such as the US combines constituent units and a central government, ‘each possessing powers delegated to it by the people through a constitution, each 12 In the American setting, see, eg, Bond v United States, 131 S Ct 2355, 2364 (2011), in which the US Supreme Court emphasised that federalism has ‘more than one dynamic’, serving ‘to grant and delimit the prerogatives and responsibilities of the [s]tates and the National Government vis-à-vis one another’, as well as protecting ‘the liberty of the individual from arbitrary power’. In the European setting, see, eg, S Oeter, ‘Federalism and Democracy’ in Principles of European Constitutional Law, above n 4, 55. For broader comparative surveys on the subject, see K Nicolaidis and R Howse (eds), The Federal Vision—Legitimacy and Levels of Governance in the United States and the European Union (Oxford, Oxford University Press, 2001); A Menon and M Schain (eds), Comparative Federalism: The European Union and the United States in Comparative Perspective (Oxford, Oxford University Press, 2006). 13 A Føllesdal, ‘Federalism’ in EN Zalta (ed), The Stanford Encyclopedia of Philosophy, Spring 2010 edn (Stanford, CA, The Metaphysics Research Lab, 2010). 14 K Lenaerts, ‘Federalism: Essential Concepts in Evolution—The Case of the European Union’ (1998) 21 Fordham International Law Journal 746, 748. For similar views, see, eg, D Halberstam, ‘Comparative Federalism and the Role of the Judiciary’ in KE Whittington, RD Kelemen and GA Caldeira (eds), The Oxford Handbook of Law and Politics (Oxford, Oxford University Press, 2008) 142. 15 P Pescatore, ‘Preface’ in T Sandalow and E Stein (eds), Courts and Free Markets (Oxford, Clarendon Press, 1982) ix–x. 16 RL Watts, ‘Federalism, Federal Political Systems, and Federations’ (1998) Annual Review of Political Science 117, 120–21.

Federalism and International Relations in the EU and the US 135 empowered to deal directly with the citizens in the exercise of a significant portion of its legislative, administrative, and taxing powers, and each directly elected by its citizens’.17 Yet ‘confederations’ are typically distinguished from ‘federations’ as a species of federal system in which the central institutions are composed of delegates from the constituent governments and therefore dependent on them. On this basis, the EU Member States considered in the present volume all have some type of a federal political system. In turn, the EU as a federal political system may be described as a hybrid form of a confederation with clearly identifiable features of a federation.18 Consequently, this chapter proceeds on the assumption that the EU occupies a position on the spectrum between a traditional international organisation and a federal State19 and that it is not too sui generis a federal political system to be subjected to comparative legal research with other such systems, including the US.20 Although the EU and the US represent different federal constructs, when a multi-level entity as exemplified by the EU is given the competence to enact measures that are binding on its component entities, ‘federal’ issues relating to the delimitation of competences arise, as they do in the US.

B. Federalism and International Relations in the EU and the US That being said, particular complexities arise when attempting to situate federalism within the field of international relations in the EU and the US. First, there is the matter of terminology. In light of the differences in federal structure between the EU and the US discussed above, it may not be surprising that this plays out in the actual terms used to describe the field of international, foreign or external relations, which may generally be defined as each legal order’s involvement and dealings with other nations and organisations in the international arena that are conducted according to the internal constraints existing

17

Ibid 121. Ibid (noting that historical examples of confederations include Switzerland for most of the period of 1291–1847 and the US during the period of 1776–89). See also L Friedman Goldstein, Constituting Federal Sovereignty (Baltimore, Md, Johns Hopkins University Press, 2001), who compares the EU to the Dutch Union of the 17th century, the USA from the 1787 Constitution to the Civil War, and the first half-century of the modern Swiss federation, beginning in 1848. 19 See, eg, JHH Weiler, The Constitution of Europe: ‘Do the New Clothes Have an Emperor?’ and Other Essays on European Integration (Cambridge, Cambridge University Press, 1999) 131–32; J Crawford, The Creation of States in International Law (Oxford, Oxford University Press, 2006) 495, who refers to the EU as going ‘well beyond the model of an international organisation coordinating areas of State policy’, and as having ‘marked confederal features’. 20 In the same sense, see, eg, K Lenaerts, ‘Constitutionalism and the Many Faces of Federalism’ (1990) 38 American Journal of Comparative Law 205; A Verhoeven, The European Union in Search of a Democratic and Constitutional Theory (The Hague, Kluwer, 2002) xii; R Schütze, ‘On “Federal Ground”: The European Union as an (Inter)national Phenomenon’ (2009) 46 CML Rev 1069. For a powerful criticism of federalism as an appropriate constitutional idea for the EU, see P Eleftheriadis, ‘Federalism and Jurisdiction’, ch 2 of the present volume. 18

136 Geert De Baere and Kathleen Gutman within the legal order concerned, eg the conclusion of international, regional and bilateral treaties, agreements and instruments, as well as the adoption of domestic measures by the central institutions or branches of each legal order that bear upon such relations.21 In the US, terms such as ‘foreign relations’, ‘foreign policy’ or ‘foreign affairs’ are usually employed, although recourse to the ‘international’ and ‘external’ modifiers (eg ‘international affairs’ or ‘external relations’) can be found.22 In the EU, there appears to be a slight preference for ‘external relations’ or ‘external action’,23 but references to ‘foreign relations’ or ‘international relations’ also abound.24 Yet the split in treatment, as a constitutional matter, between the former second pillar of the CFSP and other policy fields related to the EU’s international relations has led some commentators to distinguish the term ‘foreign policy’ as denoting matters specifically concerning the CFSP in the EU setting.25 To avoid confusion in this regard, this chapter generally uses the terms ‘external relations’ within the context of the EU, ‘foreign relations’ or ‘foreign affairs’ within the context of the US, and ‘international relations’ when referring to both systems. Secondly, each brand of European and American federalism has implications for distinguishing the roles carried out by the component entities vis-à-vis the central authority and the vertical allocation of competences between the two. For example,26 in the EU, by virtue of the fact that the Member States possess

21 In the European setting, there is no generally accepted definition of EU external relations. Arts 3(5) and 21(1) TEU refer to the Union’s ‘relations with the wider world’ and ‘action on the international scene’, respectively, and further specify these concepts through the principles that are to guide the Union’s external action. Art 24(1) TEU contains a specific definition of the Union’s CFSP and hence not of ordinary Union external relations: see text accompanying nn 41 and 48 below. In the American setting, see, eg, Restatement (Third) of Foreign Relations Law of the United States § 1 (St Paul, Minn, American Law Institute, 1987), which defines US ‘foreign relations law’ (as dealt with in the Restatement) as consisting of ‘(a) international law as it applies to the United States; and (b) domestic law that has substantial significance for the foreign relations of the United States or has other substantial international consequences’. 22 For examples in US Supreme Court jurisprudence, see text and citations accompanying n 112 below. 23 See, eg, Title V of the TEU, entitled ‘General Provisions on the Union’s External Action and Specific Provisions on the Common Foreign and Security Policy’. 24 See, eg, P Koutrakos, EU International Relations Law (Oxford, Hart Publishing, 2006); M Cremona and B de Witte (eds), EU Foreign Relations Law (Oxford, Hart Publishing, 2008); G De Baere, Constitutional Principles of EU External Relations (Oxford, Oxford University Press, 2008); P Eeckhout, EU External Relations Law, 2nd edn (Oxford, Oxford University Press, 2011). 25 See G de Búrca, ‘EU External Relations: The Governance Mode of Foreign Policy’ in B Van Vooren, S Blockmans and J Wouters (eds), The Legal Dimension of Global Governance: What Role for the EU? (Oxford, Oxford University Press, forthcoming 2012). 26 The implementation by the component entities of international and domestic enactments adopted at the central level concerning international relations in the EU and the US is another vibrant example. In the American setting, the discussion of this matter has been animated by the US Supreme Court’s judgment in Medellín v Texas, 552 US 491 (2008): for discussion of this case and citations to relevant literature, see section IV.B. In the EU, the Kadi cases have exposed the complexities of the relationship between the competence of the Union or the Member States to implement international law obligations, and the impact of that division of competences on the possibility for both levels to comply with those obligations. See Case T-301/01 Yusuf and Al Barakaat International Foundation v Council and

Federalism and International Relations in the EU and the US 137 considerable competences to conduct international relations in their own right, the issues at stake in the European setting of federalism and external relations frequently pertain to whether the Member States are allowed to exercise such competences alongside or in place of the Union in international fora, or are allowed to retain such competences in their relations with third States or international organisations in accordance with EU law. By comparison, in the US, on account of the plenary nature of the federal competences concerning foreign relations, the issues at stake in the American setting of federalism and foreign relations are not so much dealing with the states attempting to displace or sit alongside the federal government in exercising the competences conferred upon it by the US Constitution, say, to make a treaty or to declare war, but instead, typically concern the extent to which a particular form of action taken by the states may be deemed to encroach on or interfere with such competences.27 As a result, when delving into the relevant case law and the respective constitutional arrangements governing international relations in the EU and the US, the analogies may not be exact. Arguably, however, this is not fatal to the comparative analysis undertaken here, since despite differences in federal structure, both legal orders are nonetheless faced with the common problem inherent to federalism concerning how to balance the need to ensure a coherent international relations policy at the central level and the allowance for a certain degree of autonomy of the constituent entities to act in the international relations field in order to protect their own interests. Indeed, the present chapter argues that this is where comparative common ground can be found. Thirdly, following from the foregoing points, this chapter focuses on the quintessentially ‘federal’ issue of the vertical relationship between the respective central authority and the constituent entities in international relations. It does not examine issues relating to the horizontal division of competences between the legislative, executive and judicial institutions or organs at the central level, which is commonly referred to as institutional balance in the EU and the separation of powers in the US.28 Admittedly, it is difficult to separate the two in this

Commission [2005] ECR II-3533; Case T-315/01 Kadi v Council and Commission [2005] ECR II-33649; Joined Cases C-402/05 P and C-415/05 P Kadi and Al Barakaat International Foundation v Council and Commission [2008] ECR I-6351 (appeals against Cases T-301/01 and T-315/01); T-85/09 Kadi v Commission [2010] OJ C317/29; and Joined Cases C-584/10 P, C-593/10 P, C-595/10 P Commission, Council, and UK v Kadi, pending (appeals against Case T-85/09). For an insightful comment drawing parallels between Medellín and the Kadi cases, see G de Búrca, ‘The EU, the European Court of Justice and the International Legal Order after Kadi’ (2009) 51 Harvard International Law Journal 1. 27 See, eg, MA Granne, ‘Two-Dimensional Federalism and Foreign Affairs Preemption’ (2010) 44 Valparaiso University Law Review 863, 869: ‘[N]one seriously contend that the states can speak for the United States in matters of foreign policy. Thus the question becomes: when does state action that purports to merely legislate actually interfere with the federal government’s interest in a uniform expression of foreign policy?’ 28 The notion of separation of powers is certainly not unknown to the EU, but its application poses complexities on account of its roots in the ‘classical functions’ of a nation-State; institutional balance thus reflects the fact that there is not the same kind of strict division between the Union institutions,

138 Geert De Baere and Kathleen Gutman setting, since by virtue of each constitutional framework, issues relating to the vertical and horizontal division of competences are inextricably intertwined: on the vertical axis, competences are generally attributed to a particular institution, which in turn have an impact on the competences of the other institutions on the horizontal level. In the US, as discussed below, there is no general foreign relations competence attributed to the federal government as such; rather, specific competences concerning foreign relations are granted to the legislative, executive and judicial branches of the federal government, which provide the focal point for assessing the extent of the competences exercised by one branch vis-à-vis another, as well as the ambit for state action in this field.29 Similarly, in the EU, the struggle over external relations competences takes place on both the horizontal and vertical axes,30 even if the distinction between the two can sometimes be made more readily. The ECJ took pains to explain this distinction in the Personal Data Protection case,31 holding that the question of the ‘areas of competence’ of the EU presents itself differently depending on whether the competence concerned has already been accorded to the EU in the broad sense or has not yet been accorded to it: In the first hypothesis, it is a question of ruling on the division of areas of competence within the Union and, more particularly, on whether it is appropriate to proceed by way of a [Union legal act] based on the [FEU] Treaty or by way of a [Union legal act] based on the EU Treaty. By contrast, in the second hypothesis, it is a question of ruling on the division of areas of competence between the Union and the Member States and, more particularly, on whether the Union has encroached on the latter’s areas of competence.32

It would appear to be more logical to examine these two hypotheses in reverse order, inquiring, first, whether the Union can act at all, and secondly, if it can, under which precise competence(s) it may do so and through which institutions. In any event, much like in the US, in practical terms, these questions will often be dealt with simultaneously in the EU as well.33 Consequently, while such interplay cannot always be avoided, this chapter must leave aside detailed discussion of the horizontal division of competences in international relations, although the

and depending upon the Treaty provision concerned, the same institution may carry out legislative and executive functions. See generally K Lenaerts and P Van Nuffel, European Union Law, 3rd edn (London, Sweet & Maxwell, 2011) §§ 15-007–15-013, at 635–41. 29

See section IV.A. E Stein, Thoughts from a Bridge: A Retrospective of Writings on New Europe and American Federalism (Ann Arbor, University of Michigan Press, 2000) 259. 31 Case C-301/06 Ireland v European Parliament and Council (‘Personal Data Protection’) [2009] ECR-593. 32 Ibid para 56. 33 See further E Sharpston and G De Baere, ‘The Court of Justice as a Constitutional Adjudicator’ in A Arnull, C Barnard, M Dougan and E Spaventa (eds), A Constitutional Order of States? Essays in EU Law in Honour of Alan Dashwood (Oxford, Hart Publishing, 2011) 128. 30

Federalism and International Relations in the EU and the US 139 analysis undertaken here may help to lay the groundwork for further research on this topic.

III. CONSTITUTIONAL FRAMEWORK OF FEDERALISM AND EXTERNAL RELATIONS IN THE EU

A. Overview of the Allocation of Competences in the Treaties i. The Principle of Conferral One of the most remarkable features of EU external relations is that, unlike the US or indeed most nation-States, when considering a response to an international situation, the EU must always give precedence to considerations of competence over considerations of effectiveness.34 Pursuant to the principle of conferral, which governs the limits of Union competences, the Union is to ‘act only within the limits of the competences conferred upon it by the Member States in the Treaties to attain the objectives set out therein. Competences not conferred upon the Union in the Treaties remain with the Member States.’35 As a constitutional principle, the principle of conferral incorporates the idea, fundamental not only in the law of international organisations, but also in the constitutional law of many federal States, that the Union has only those competences that the Member States have explicitly or impliedly conferred on it in the Treaties. That does not mean, however, that the principle of conferral is there only to safeguard the Member States’ competences, since competences once conferred on the Union cannot be repatriated to the Member States without amending the Treaties. Still, the principle of conferral generally tends to be viewed in terms of protecting the competences of the Member States against encroachment by the Union.36 This implies that the EU is incapable of extending its own competences and that it does not have general law-making capacity: Union action requires one or more legal bases in the Treaties, which determines both the vertical and horizontal division of competences37 and must be based on objective factors that are amenable to judicial review.38

34

De Baere, above n 24, 10 and the literature cited therein. Art 5(1) and (2) TEU. 36 See K Gutman, The Constitutionality of European Contract Law: A Comparative Analysis (Oxford, Oxford University Press, forthcoming 2013), ch 7. 37 See Personal Data Protection, above n 31, para 56. But see Art 352 TFEU (ex Art 308 TEC): ‘If action by the Union should prove necessary, within the framework of the policies defined in the Treaties, to attain one of the objectives set out in the Treaties, and the Treaties have not provided the necessary powers, the Council, acting unanimously on a proposal from the Commission and after obtaining the consent of the European Parliament, shall adopt the appropriate measures.’ (emphasis added) 38 See, eg, Case C-376/10 P Tay Za v Council [2012] ECR I-0000, para 46 and the case law cited there. 35

140 Geert De Baere and Kathleen Gutman Crucially, the principle of conferral applies as much to external relations as to internal policies.39 The consequences of this constitutional characteristic have been exacerbated by the fact that, even after the entry into force of the Lisbon Treaty,40 the EU approaches the subject of external relations from two quite different angles as regards ordinary EU external relations41 and the CFSP (the former first and second pillars, respectively) based on the distinction between two core aspects of external relations: external economic relations sensu lato, such as external trade, development co-operation and emergency aid, which traditionally fell within the former Community, and what are commonly called ‘high politics’ (diplomatic activity and security and defence issues), which fall within the CFSP.42 To point out that the distinction between these two aspects is not always easy to make and that the coherence between the various aspects of EU external relations remains a challenge is stating the obvious.43 The Lisbon Treaty has subjected the CFSP to the overall constitutional framework of the EU, as set out in the EU Treaty.44 The formal abolition of the pillar structure, however, does not imply a complete harmonisation of procedures and an integration of all policies under the former Community, and essentially leaves the former second pillar standing in a modified manner.45 The post-Lisbon EU Treaty takes into account the different characters of different policies and still permits a substantial amount of differentiation as to how the Union’s institutions

39 Opinion 2/94 Accession by the Communities to the Convention for the Protection of Human Rights and Fundamental Freedoms [1996] ECR I-1759, para 24. See also Opinion 2/00 Cartagena Protocol [2001] ECR I-9713, para 5: ‘The choice of the appropriate legal basis has constitutional significance. Since the [Union] has conferred powers only, it must tie the [international agreement in question] to a Treaty provision which empowers it to approve such a measure. To proceed on an incorrect legal basis is therefore liable to invalidate the act concluding the agreement and so vitiate the [Union]’s consent to be bound by the agreement it has signed.’ See further De Baere, above n 24, 9–11. 40 Above n 6. 41 G De Baere, ‘The Basics of EU External Relations Law: An Overview of the Post-Lisbon Constitutional Framework for Developing the External Dimensions of EU Asylum and Migration Policy’ in M Maes, M-C Foblets, and Ph De Bruycker (eds), External Dimensions of EU Migration and Asylum Law and Policy/Dimensions Externes du Droit et de la Politique d’Immigration et d’Asile de l’UE (Brussels, Bruylant, 2011) 121. 42 A Dashwood, M Dougan, B Rodger, E Spaventa and D Wyatt, Wyatt and Dashwood’s European Union Law, 6th edn (Oxford, Hart Publishing, 2011) 13. See also De Baere, above n 24, 219. 43 On coherence and consistency in EU external relations, see, eg, C Hillion, ‘Tous pour un, un pour tous! Coherence in the External Relations of the European Union’ in M Cremona (ed), Developments in EU External Relation Law (Oxford, Oxford University Press, 2008) 10; and De Baere, above n 24, 250–64 and 267–74. 44 It has also introduced a number of institutional innovations intended to bridge the cleft between ordinary EU external relations and the CFSP, notably the establishment of the office of the High Representative of the Union for Foreign Affairs and Security Policy. The latter conducts the CFSP as a Council functionary, but is at the same time one of the Vice-Presidents of the Commission and is therefore charged with ensuring the consistency of the Union’s external action (Art 18(2) and (4) TEU). The High Representative is assisted in that regard by the European External Action Service (EEAS) (Art 27(3) TEU). See further B Van Vooren, ‘A Legal-Institutional Perspective on the European External Action Service’ (2011) 48 CML Rev 475. 45 Perhaps more like a separate room formed by the few walls left standing after having knocked down all the other ones to create the new loft-like Union: De Baere, above n 41, 123.

Federalism and International Relations in the EU and the US 141 are involved in law-making. The 2007 Intergovernmental Conference, which led to the adoption of the Lisbon Treaty, decided to drop the idea of one single Constitutional Treaty and to keep the EU Treaty and the FEU Treaty as two distinct Treaties, with the CFSP and the European Neighbourhood Policy as the only substantive policies in the EU Treaty. The Lisbon Treaty therefore establishes a single legal order for the Union, but with a more markedly separate sub-order for the CFSP.46 Moreover, the present Chapter 2 of Title V of the EU Treaty remains characterised by an absence of the technique of detailed and specific attribution of competences,47 otherwise applicable to the ordinary Union framework under the FEU Treaty. Instead, the allocation of competences in the CFSP consists of the general grant of competence in Article 24(1) TEU,48 which covers ‘all areas of foreign policy and all questions relating to the Union’s security, including the progressive framing of a common defence policy that might lead to a common defence’. Article 42 TEU supplements Article 24(1) TEU by stating that the common security and defence policy (CSDP) is to be an integral part of the CFSP and provide the Union with an operational capacity drawing on civilian and military assets. The CSDP is to include the ‘progressive framing of a common Union defence policy’, which according to Article 24(1) TEU might and according to Article 42(2) TEU will ‘lead to a common defence’, that is, ‘when the European Council, acting unanimously, so decides’.49 Should the European Council take a decision to that effect, it will recommend to the Member States that they adopt such a decision in accordance with their respective constitutional requirements.50 The question arises whether the principle of conferral actually applies to the CFSP. Given that, under the previous Treaty framework, the principle of conferral

46

See further De Baere, above n 24, 209–13. That technique is the corollary of the principle of conferral: cf A Dashwood, ‘The Relationship Between the Member States and the European Union/European Community’ (2004) 41 CML Rev 357 et seq. 48 Ex Art 11(1) TEU. 49 In the Maastricht Treaty, this was phrased as ‘the eventual framing of a common defence policy, which might in time lead to a common defence’: Art B, second indent and Art J.4(1). Clearly, not much legal significance should be attached to the phrase ‘will lead to a common defence’, which should be understood as an aspirational statement of purely political nature. See De Baere, above n 24, 108–09. 50 This avoids having to go through the normal procedure for amendment of the Treaty (Art 48 TEU). The reasons for this are symbolic rather than practical, as the approval by the Member States in accordance with their own constitutional requirements is likely to be as cumbersome as an amendment of the Treaty. See also the Decision of the Heads of State or Government of the 27 Member States of the EU, Meeting within the European Council, on the Concerns of the Irish People on the Treaty of Lisbon (Annex 1 to the Presidency Conclusions of the European Council held at Brussels on 18 and 19 June 2009). Section C of this Decision on security and defence reaffirms that any ‘decision to move to a common defence will require a unanimous decision of the European Council. It would be a matter for the Member States, including Ireland, to decide, in accordance with the provisions of the Treaty of Lisbon and with their respective constitutional requirements, whether or not to adopt a common defence.’ The European Council also felt the need to specify the following: ‘The Treaty of Lisbon does not provide for the creation of a European army or for conscription to any military formation.’ 47

142 Geert De Baere and Kathleen Gutman was not enshrined in the former EU Treaty51 but in the first paragraph of then Article 5 TEC, some doubted whether the principle applied to the CFSP.52 Article 5(1) and (2) TEU, as introduced by the Lisbon Treaty in Title I (‘Common Provisions’) of the EU Treaty, has now removed all doubt in that regard. Remarkably, however, the Lisbon Treaty has reinforced the contrast between attribution in ordinary EU external relations and in the CFSP by deleting the specific CFSP objectives listed in ex Article 11(1) TEU, thus making the attribution within that field even less detailed and specific. Instead, Article 21 TEU now contains the overall objectives of EU external relations. Article 21(1) TEU provides that the Union’s action on the international scene is to be ‘guided by the principles which have inspired its own creation, development and enlargement, and which it seeks to advance in the wider world’,53 while Article 21(2) TEU further elaborates these objectives in a rather more comprehensive list. Detailed and specific attribution of competences is an important aspect of the principle of conferral as it operated under the Community and now operates under the FEU Treaty, and its absence in Chapter 2 of Title V of the EU Treaty indicates that the principle applies in a different manner in the CFSP legal order. The Lisbon Treaty has reinforced the contrast between ordinary Union external relations and the CFSP in that respect. In what follows, this chapter examines the issues of competence involved in determining the possibility for the EU to act internationally: the existence of EU external competence, the nature of such competence and the consequences of that particular nature. In that regard, it is important to note that competence questions continue to be at the heart of EU external relations law, and that the intensity and complexity of these questions is still increasing.54 The following outline illustrates why that is the case.

51 Ex Art 5 TEU could, however, be seen as an expression of the principle of conferral: see the Opinion of AG Mengozzi in Case C-354/04 P Gestoras Pro Amnistía, JM Olano Olano, J Zelarain Errasti v Council [2007] ECR I-1579 and Case C-355/04 P SEGI, A Zubimendi Izaga, A Galarraga v Council [2007] ECR I-1657, point 167. 52 See, eg, R Gosalbo Bono, ‘Some Reflections on the CFSP Legal Order’ (2006) 34 CML Rev 358–59. However, while not mentioned explicitly in the pre-Lisbon EU Treaty, the principle of conferral was, and still, is a general organising principle of the constitutional structure of the EU, as well as a principle of the law of international organisations (Jurisdiction of the European Commission of the Danube Between Galatz and Braila [1927] PCIJ Series B—No 14, 64; cf D Sarooshi, International Organizations and Their Exercise of Sovereign Powers (Oxford, Oxford University Press, 2005)). Thus, while not being very detailed or exact, the CFSP objectives in ex Art 11(1) TEU determined the Union’s competences to the extent that any CFSP action had to be connected to those objectives in accordance with the principle of conferral. 53 That is to say, ‘democracy, the rule of law, the universality and indivisibility of human rights and fundamental freedoms, respect for human dignity, the principles of equality and solidarity, and respect for the principles of the United Nations Charter and international law’. 54 P Eeckhout, ‘General Report’ in XL Xenopoulos (ed), External Relations of the EU and the Member States: Competence, Mixed Agreements, International Responsibility, and Effects of International Law (Limassol, FIDE, 2006) 275–76. See further P Koutrakos, ‘Legal Basis and Delimitation of Competences in EU External Relations’ in Cremona and de Witte (eds), above n 24, 171–98.

Federalism and International Relations in the EU and the US 143 ii. Existence of EU External Competences The former EC Treaty55 did not contain a general legal basis for external relations. Article 281 TEC explicitly conferred legal personality on the Community,56 but the general capacity derived from that Article did not constitute an independent legal basis for the former Community’s adoption of international agreements. However, as a legal person, the Community had the capacity to exercise rights in international legal transactions and enter into obligations over the entire field of its objectives.57 Unlike the Community, the pre-Lisbon Union had not been explicitly endowed with legal personality.58 Nevertheless, former Article 24 TEU provided the Union with a procedural framework for making international agreements within the spheres of the former second and third pillars concerning the CFSP and Police and Judicial Cooperation in Criminal Matters (PJCCM), respectively.59 With the entry into force of the Lisbon Treaty, Article 47 TEU explicitly confirms the Union’s legal personality, thereby removing any lingering doubt in that regard. Furthermore, Article 216(1) TFEU affirms the general capacity of the Union to conclude international agreements. That provision stipulates that the Union may conclude an agreement with one or more third countries or international organisations: (a) where the Treaties so provide; or (b) where the conclusion of an agreement is ‘necessary in order to achieve, within the framework of the Union’s policies, one of the objectives referred to in the Treaties, or is provided for in a legally binding Union act or is likely to affect common rules or alter their scope’. That phrase is intended as a codification of the Court of Justice’s case law on implied competences, and will be examined more closely in the next section. Indeed, the Treaty provisions regulating external relations have always been both spread over the entire Treaty and incomplete. Attempts to address the lack of explicit legal bases for external action have been made predominantly in two ways: first, explicit legal bases for external competences were added to the Treaties in subsequent 55

Consolidated Version of the Treaty Establishing the European Community [2006] OJ C321E/37. See also the identical Art 184 EAEC Treaty (Consolidated Version of the Treaty establishing the European Atomic Energy Community [2010] OJ C84/1). 57 Case 22/70 Commission v Council (‘ERTA’) [1971] ECR 263, paras 13–14; and Joined Cases 3, 4, and 6/76 Cornelis Kramer and Others (‘Kramer’) [1976] ECR 1279, paras 17–18. 58 Many commentators have come to the conclusion that, even before the Lisbon Treaty, the EU did have international legal personality (eg I Brownlie, Principles of Public International Law, 7th edn (Oxford, Oxford University Press, 2008) 678), and this was confirmed in practice. For one of the last examples under the previous Treaty framework, see Council Decision 2010/53/CFSP of 30 November 2009 concerning the conclusion of the Agreement between Australia and the European Union on the security of classified information [2010] OJ L26/30. See further De Baere, above n 24, 143–46. 59 See further RA Wessel, ‘The EU as a party to international agreements: shared competences, mixed responsibilities’ in A Dashwood and M Maresceau (eds), Law and Practice of EU External Relations (Cambridge, Cambridge University Press, 2008) 152–87. 56

144 Geert De Baere and Kathleen Gutman amendments;60 and, secondly, the Court of Justice’s interpretation of the existing Treaty provisions so as to allow the Union to develop a viable external relations policy,61 resulting in what is mostly referred to as ‘implied external competences’. Notably, however, the nature of the attribution as regards the CFSP in Article 24(1) TEU would appear to be so broad that an unself-conscious application of the doctrine of implied competences implying all the competences needed for an effective CFSP would lead to an extensive grant of foreign policy competences going far beyond what the EU Treaty intends. Applying the doctrine of implied competences to the CFSP therefore requires caution and restraint, even more so because the ECJ for the most part lacks jurisdiction with respect to the provisions in the EU Treaty relating to the CFSP and with respect to acts adopted on the basis of those provisions.62 iii. The Nature of EU External Competences One of the more significant novelties introduced by the Lisbon Treaty is Title I of Part One of the FEU Treaty entitled ‘Categories and Areas of Union Competence’. Three of those categories as listed in Article 2 TFEU are most relevant for the Union’s external relations and will be explored further here,63 ie exclusive competences,64 shared competences,65 and the competence to define and implement a CFSP, including the progressive framing of a CSDP.66 Article 2(1) TFEU provides that when the Treaties confer on the Union exclusive competence in a specific area, only the Union may legislate and adopt legallybinding acts. The Member States are able to do so themselves only if they have been so empowered by the Union or for the implementation of Union acts. The main principles on when the Union is exclusively competent have now been laid down in Article 3 TFEU, the first paragraph of which lists the five explicitly attributed exclusive competences of the Union. That category is rather small and comprises the customs union, ie the original core of the Community, the establishing of the

60

For an overview as regards the EC Treaty, see De Baere, above n 24, 11–16. As regards the important role of the ECJ in (what were then) Community external relations, see De Baere, above n 24, 93–98; and on its virtual absence in the CFSP (except for its role under Art 40 TEU (ex Art 47 TEU) and, since the entry into force of the Lisbon Treaty, Art 275 TFEU), ibid, 176–91. 62 There are certain exceptions carved out in the Treaties conferring jurisdiction on the Court of Justice of the EU to monitor compliance with Art 40 TEU and to rule on proceedings, brought in accordance with the conditions laid down in Art 263, fourth para, TFEU, reviewing the legality of decisions providing for restrictive measures against natural or legal persons: Art 24(1), second subpara, TEU; Art 275, second para, TFEU. 63 The Union also has the competence to provide arrangements within which the Member States are to co-ordinate their economic and employment policies (Art 2(3) TFEU) and the competence to carry out actions to support, coordinate or supplement the actions of the Member States (Art 2(5) TFEU), which will not be considered further in the present chapter. 64 Art 2(1) TFEU. 65 Art 2(2) TFEU. 66 Art 2(4) TFEU. 61

Federalism and International Relations in the EU and the US 145 competition rules necessary for the functioning of the internal market, monetary policy for the Member States whose currency is the euro, the conservation of marine biological resources under the common fisheries policy, and the common commercial policy. Given that they have important external relations aspects, the external dimensions of these policies belong to the exclusive competence of the Union. In addition, the second paragraph of Article 3 TFEU provides for the Union to have exclusive competence ‘for the conclusion of an international agreement when its conclusion is provided for in a legislative act of the Union or is necessary to enable the Union to exercise its internal competence, or in so far as its conclusion may affect common rules or alter their scope’.67 While it attempts to codify the case law of the Court of Justice on exclusive competences, the second paragraph is of little use in clarifying ex ante the extent of the Union’s exclusive external competences. The criteria listed therein appear to be neither entirely clear nor sufficiently nuanced, and hence in need of further judicial clarification.68 The Court’s existing case law on the nature of EU external competences, which will be examined in the section below, will need to be taken into account. However, in contrast to the constituent states of the US, the Member States retain capacity to conduct international relations under both national and international law. Exclusive Union competences simply require them not to act autonomously. This is evident from the text of Article 2(1) TFEU, which provides for the possibility for the Union to empower the Member States to act in an area in which the Union enjoys exclusive competence. The Member States may therefore be authorised by EU law to act jointly on the international plane, even within exclusive external Union competences.69 As mentioned above, Article 216(1) TFEU grants the Union the competence to conclude an agreement with one or more third countries or international organisations where the conclusion of an agreement is ‘provided for in a legally binding Union act’, that is to say, in a regulation, a directive or a decision.70 Does the difference in language between Articles 3(2) (‘provided for in a legislative act of the Union’) and 216(1) TFEU (‘provided for in a legally binding Union act’) imply that the exclusive nature of the external competence depends on the nature of the procedure by which the internal act granting that competence was adopted? That would seem to be the implication of the wording of those two provisions:71 if the possibility to conclude an international agreement is provided for in a legally-binding Union act, the EU acquires competence to conclude that agreement on the basis of Article 216(1) TFEU. If that same possibility is provided 67

Art 3(2) TFEU. Compare the Treaty establishing a Constitution for Europe [2004] OJ C310/1: P Craig, EU Administrative Law (Oxford, Oxford University Press, 2006) 418. 69 See further De Baere, above n 24, 59–61. 70 Art 288 TFEU. 71 Perhaps the distinction in wording was introduced to make it clear that this principle cannot give rise to an exclusive competence within the CFSP, given that the second paragraph of Art 24 TEU and Art 31(1) TEU exclude the adoption of legislative acts. 68

146 Geert De Baere and Kathleen Gutman for in a legislative act, the Union acquires exclusive competence to conclude that agreement. That said, it remains the case that even a legislative act cannot grant the Union an exclusive external competence in those areas for which Article 4(3) and (4) TFEU explicitly provide that the Member States cannot be prevented from acting internationally. Indeed, it is important to note that the EU’s competence in external matters should be presumed to be non-exclusive, unless there are clear indications to the contrary.72 Within the category of non-exclusive competences, two sub-categories may be distinguished: first, ‘shared competences’, as provided for by Articles 2(2) and 4 TFEU (sometimes referred to as ‘concurrent’ competences), may be exercised by the Member States to the extent that the Union has not exercised, or has decided to cease exercising, its competence; and, secondly, ‘parallel competences’ (a term not used in the Treaties) leave both the Union and the Member States competent to act internationally without one being able definitively to prevent the other from acting. Some provisions of the TFEU explicitly provide for such competences. An important example73 is Article 211 TFEU74 on development co-operation, which states that the Union and the Member States, each within their respective spheres of competence, are to co-operate with third countries and with the competent international organisations. Pursuant to Article 209(2) TFEU, the arrangements for Union co-operation may be the subject of agreements between the Union and the third parties concerned, which are negotiated and concluded in accordance with Article 218 TFEU.75 However, Article 209(2) TFEU ends with the explicit assurance that the external competence described in that Article ‘shall be without prejudice to Member States’ competence to negotiate in international bodies and to conclude international agreements’. Furthermore, Article 4(4) TFEU provides for the exercise of the Union’s competence in the areas of development co-operation and humanitarian aid not to result in Member States being prevented from exercising theirs, thus creating parallel competences in those areas. What type of competence is the CFSP? Might it be argued that Article 4(1) TFEU (stating that ‘The Union shall share competence with the Member States where the Treaties confer on it a competence which does not relate to the areas referred to in Articles 3 and 6’) implies that the CFSP is a shared competence as well? Though a literal reading of Article 4(1) TFEU would seem to have that consequence, in view of Declarations Nos 13 and 14, annexed to the Lisbon Final Act,76 it appears highly 72 See, in that sense, the Opinion of AG Jacobs in Case C-316/91 European Parliament v Council [1994] ECR I-625, point 40. 73 See also Art 219(4) TFEU (ex Art 111(5) TEC); Art 191(4) TFEU (ex Art 174(4) TEC). 74 Ex Art 181 TEC. 75 Ex Art 300 TEC. 76 [2010] OJ C83/343. These two declarations emphasise that the provisions in the TEU covering the CFSP, including the creation of the office of High Representative and the establishment of an EEAS, do not affect the responsibilities of the Member States, as they currently exist, for the formulation and conduct of their foreign policy, their national representation in third countries and international organisations, or the existing legal basis, responsibilities and powers of each Member State in relation to the formulation and conduct of its foreign policy, its national diplomatic service,

Federalism and International Relations in the EU and the US 147 unlikely that the Member States had the intention of subjecting the CFSP to the corollary of shared competences, namely, that the Member States can exercise their competence ‘to the extent that the Union has not exercised its competence’ or ‘has decided to cease exercising its competence’.77 Moreover, if the CFSP was intended to be a ‘shared competence’, why then create a separate category of CFSP competence in Article 2(4) TFEU instead of listing it among shared competences in Article 4(2) TFEU and adding a clause to the effect that ‘the exercise of that competence shall not result in Member States being prevented from exercising theirs’, as was done with regard to research, technological development, space, development co-operation and humanitarian aid in Article 4(3) and (4) TFEU?78

B. Case Law of the European Court of Justice In view of the complex issues concerning the allocation of competences, it is not surprising that the case law of the ECJ plays an important role in guarding the federal balance in EU external relations. The present section examines the Court’s case law on the existence and the nature of EU external competences, as well as the consequences of the exclusive or non-exclusive nature of those competences. i. Case Law on the Existence of EU External Competences Famously expounded by Chief Justice Marshall of the US Supreme Court in M’Culloch v Maryland,79 the doctrine of implied competences is a well-known principle of municipal constitutional law and of the law of international institutions.80 Within the EU, implied competences have been relied on mostly with regard to external relations. The Court of Justice recognised implied external competences for the first time in ERTA,81 against the advice of Advocate General Dutheillet de Lamothe, who had argued that [n]o matter what legal basis the Court finds for it, recognition of the Community’s authority in external matters for negotiating and concluding the AETR concedes by

relations with third countries and participation in international organisations, including a Member State’s membership of the UN Security Council. Moreover, the provisions concerning the CFSP are explicitly said not to give new powers to the Commission to initiate decisions or to increase the role of the European Parliament. 77

Art 2(2) TFEU. See further De Baere, above n 24, 110–11. 79 M’Culloch v Maryland, 17 US 316, 421 (1819). 80 See Reparation for Injuries suffered in the Service of the United Nations, Advisory Opinion [1949] ICJ Reports 174, 182. 81 Above n 57, para 17. The acronyms ERTA/AETR stand for European Agreement Concerning the Work of Crews of Vehicles Engaged in International Road Transport/Accord européen relatif au travail des équipages des véhicules eff ectuant des transports internationaux par route. 78

148 Geert De Baere and Kathleen Gutman implication that the Community authorities exercise, in addition to the powers expressly conferred upon them by the Treaty, those implied powers whereby the Supreme Court of the United States supplements the powers of the federal bodies in relation to those of the [federated]82 States.83

The Court of Justice now recognises implied external competences mainly on the basis of two principles. First, the ERTA principle follows the logic of the principle of primacy84 and was first expounded in the judgment in ERTA: the Member States are not allowed to act internationally in a way that would affect existing EU law, because the situation cannot be remedied merely by disapplying the infringing national rule. The Member States’ competence is thus excluded, which necessitates the existence of EU competences to compensate for the Member States’ inability to act.85 That principle is now codified in Article 216(1) TFEU: The Union may conclude an agreement with one or more third countries or international organisations … where the conclusion of an agreement … is likely to affect common rules or alter their scope.

Secondly, the complementarity principle was spelled out in Opinion 1/76 Draft Agreement establishing a European laying-up fund for inland waterway vessels:86 whenever EU law has conferred internal competences on the institutions to attain a specific objective, the Union may enter into the international commitments necessary for attainment of that objective even in the absence of an express provision to that effect.87 Internal Union competences are supported by the corresponding external competences only when the latter are truly ‘implicit’ in the former. This is the case when the internal Union competences cannot reasonably be expected to be exercised effectively without the possibility for the Union to 82 The original French version of the AG Opinion uses ‘États fédérés’, which appears to have been mistranslated in English as ‘confederated States’. 83 Opinion in ERTA, above n 57, 293. 84 First affirmed by the ECJ in Costa (Case 6/64 Costa v ENEL [1964] ECR 585), the principle of supremacy (or primacy) of Union law generally denotes that a conflict between a provision of national law and a provision of EU law is to be resolved by a national court or other national authority applying Union law, if necessary by refusing to apply the conflicting national provision (see, to that effect, eg, Case C-314/08 Filipiak [2009] ECR I-11049, para 82). It is not enshrined in the Treaties, as is the Supremacy Clause in the US Constitution (see below n 105). A codification was proposed in Art I-6 of the Treaty establishing a Constitution for Europe (‘The Constitution and law adopted by the institutions of the Union in exercising competences conferred on it shall have primacy over the law of the Member States’), but the 2007 IGC decided to omit that provision from the Lisbon Treaty. Nevertheless, in Declaration 17, annexed to the Final Act of the 2007 IGC [2010] C83/344, the IGC recalled that ‘in accordance with well settled case law of the Court of Justice of the European Union, the Treaties and the law adopted by the Union on the basis of the Treaties have primacy over the law of Member States, under the conditions laid down by the said case law’. The IGC further annexed an Opinion of the Council Legal Service confirming as much. As to whether the principle of primacy applies to the CFSP, see De Baere, above n 24, 201–12. 85 Cf A Dashwood and J Heliskoski, ‘The Classic Authorities Revisited’ in A Dashwood and C Hillion (eds), The General Law of EC External Relations (London, Sweet & Maxwell, 2000) 6. 86 Opinion 1/76 Draft Agreement establishing a European laying-up fund for inland waterway vessels [1977] ECR 741. 87 Ibid para 3; Opinion 2/94, above n 39, para 26.

Federalism and International Relations in the EU and the US 149 enter into international agreements with third countries on the same subjectmatter. That principle is now codified in Article 216(1) TFEU: The Union may conclude an agreement with one or more third countries or international organisations … where the conclusion of an agreement is necessary in order to achieve, within the framework of the Union’s policies, one of the objectives referred to in the Treaties …

Article 216(1) TFEU lists a third possibility, namely, that the Union ‘may conclude an agreement with one or more third countries or international organisations … where the conclusion of an agreement … is provided for in a legally binding Union act’, which was examined above. ii. Case Law on the Nature of EU External Competences A number of instances of when the Union acquires an exclusive external competence may broadly be distinguished in the case law. First, a competence may become exclusive through the exercise of an EU internal competence. That eventuality is called ‘the ERTA-doctrine’ or ‘ERTA exclusivity’. This is now reflected in Article 3(2) TFEU, which provides that the Union will have exclusive competence for the conclusion of an international agreement ‘insofar as its conclusion may affect common rules or alter their scope’. First recognised by the Court in the ERTA case and further refined inter alia in the Open Skies cases, this manner of acquiring exclusive external competences is the result of the fact that Member States are not to enter into international obligations outside the framework of the Union institutions if these obligations fall within the scope of the common rules, or within an area which is already largely covered by such rules, even if there is no contradiction between those commitments and the common rules.88 If the Union has achieved complete harmonisation in a given area, it acquires an exclusive external competence in that area, even in the absence of any express provision authorising its institutions to negotiate with non-Member States. This is so because the common rules thus adopted could be affected within the meaning of the ERTA principle if the Member States retained freedom to negotiate with non-Member States.89 Both explicit and implied external competences may become exclusive through the exercise by the Union of its competences.90

88 Case C-467/98 Commission v Denmark [2002] ECR I-9519, para 82; Case C-476/98 Commission v Germany [2002] ECR I–9855, para 108. 89 Commission v Denmark, above n 88, para 84; Commission v Germany, above n 88, para 110. 90 By analogy with US constitutional doctrine (see the discussion of the pre-emption doctrine in section IV.B. below), this is sometimes referred to as ‘pre-emption’. However, the ECJ has not adopted this term, and there is no academic consensus on its usage. Nevertheless, the term has on occasion been used by an Advocate General: see, eg, the Opinion of AG Ruiz-Jarabo Colomer in Case C-478/07 Bude˘jovický Budvar [2009] ECR I-7721, point 93.

150 Geert De Baere and Kathleen Gutman Secondly, in Opinion 1/94 Competence of the Community to conclude international agreements concerning services and the protection of intellectual property, the Court held that [w]henever the [Union] has included in its internal legislative acts provisions relating to the treatment of nationals of non-member countries or expressly conferred on its institutions powers to negotiate with non-member countries, it acquires exclusive external competence in the spheres covered by those acts.91

The Court seemed to hold that this automatically implied exclusive competence for the Community. The Lisbon Treaty has now codified the Court’s case law in this respect in Article 3(2) TFEU, which provides that the Union will have exclusive competence for the conclusion of an international agreement ‘when its conclusion is provided for in a legislative act of the Union’.92 As mentioned above, that statement does appear to merit some nuance in the light of, inter alia, Article 4(3) and (4) TFEU. Thirdly, the Court’s case law also provides for the possibility of exclusivity to arise out of the fact that the internal and external aspects of a policy area can be exercised effectively only if they are exercised together. In that respect, the Court appears to regard the distinction between two types of envisaged international agreements to be crucially important. On the one hand, there are the rare factual constellations in which the objectives of EU competences cannot be achieved without including third countries, through international agreements, within the binding legal framework regulating the situation. The Union acquires exclusive implied external competences in the event of such an inextricable link.93 The Court’s case law has now been codified in Article 3(2) TFEU, which provides that the Union will have exclusive competence for the conclusion of an international agreement when this is ‘necessary to enable the Union to exercise its internal competence’. On the other hand, there exist many situations in which the Union wishes to further, on the external front, internal legislative goals, the optimal use of which presupposes an external complement but which could nonetheless arguably have been sufficiently attained with internal legislation only. The Union does not acquire exclusive external competence in those situations. This is what Opinion 2/92 Competence of the Community or one of its institutions to participate 91 Opinion 1/94 Competence of the Community to conclude international agreements concerning services and the protection of intellectual property [1994] ECR I-5267, para 95; see also Commission v Denmark, above n 88, para 83; Commission v Germany, above n 88, para 109. 92 Pursuant to Art 289(3) TFEU, ‘[l]egal acts adopted by legislative procedure shall constitute legislative acts’. On the distinction between ‘legislative acts’ and ‘regulatory acts’, which is of importance for the standing of natural and legal persons to bring an action for annulment under Art 263, fourth para, TFEU, see the order of the General Court of 6 September 2011 in Case T-18/10 Inuit Tapiriit Kanatami and Others v European Parliament and Council [2011] ECR II-0000, para 56 (appeal pending before the ECJ: Case C-583/11 P); and the judgment of the General Court of 25 October 2011 in Case T-262/10 Microban International and Microban (Europe) v Commission [2011] ECR II-0000, para 21. 93 Though the Court omitted a reference to the need for such an inextricable link in Opinion 1/03 Competence of the Community to conclude the new Lugano Convention on jurisdiction and the recognition and enforcement of judgments in civil and commercial matters [2006] ECR I-1145, para 115.

Federalism and International Relations in the EU and the US 151 in the Third Revised Decision of the OECD on national treatment seemed to say,94 and it appears to correspond to Article 216(1) TFEU. However, neither Article 3(2) TFEU nor Article 216(1) TFEU sufficiently reflects the complexity of the case law on which it is based, especially with regard to the need for an ‘inextricable link’ in order for ‘necessity’ to give rise to an exclusive Union competence. The Court will have to interpret what is to be understood under ‘necessary’ as it had to before the codification, and it will presumably rely on its earlier case law to do so. It follows from the Court’s case law that the exclusivity of the Union’s external competence has two main consequences, which are really two sides of the same coin.95 First, when the external competence in a certain area is exclusive, the Member States, as the Court put it in ERTA, ‘no longer have the right, acting individually or even collectively, to undertake obligations with third countries’.96 The Court also referred to what is now Article 4(3) TEU, enshrining the principle of loyalty and sincere co-operation, and concluded that it would be impossible for the Member States operating outside the institutional framework of the Union to assume responsibilities that might affect or alter the scope of Union rules that have been promulgated for the attainment of Treaty objectives.97 Secondly, the Court pointed out in Ruling 1/78 Draft Convention of the International Atomic Energy Agency on the Physical Protection of Nuclear Materials, Facilities and Transports that when external Union competence is exclusive, ‘the Member States, whether acting individually or collectively, are no longer able to

94 Opinion 2/92 Competence of the Community or one of its institutions to participate in the Third Revised Decision of the OECD on national treatment [1995] ECR I-521, para 32. 95 Exclusivity also implies the inapplicability of the principle of subsidiarity. Art 5(3), first subpara, TEU provides: ‘Under the principle of subsidiarity, in areas which do not fall within its exclusive competence, the Union shall act only if and in so far as the objectives of the proposed action cannot be sufficiently achieved by the Member States, either at central level or at regional and local level, but can rather, by reason of the scale or effects of the proposed action, be better achieved at Union level’. But see Case T-253/02 Ayadi v Council [2006] ECR II-2139, paras 105-114, where the General Court held that the principle of subsidiarity could not ‘be relied on in the sphere of application of Arts 60 TEC and 301 TEC [now Arts 75 and 215 TFEU], even on the assumption that it does not fall within the exclusive competence of the Community’ (para 108), adding that ‘even assuming that the principle of subsidiarity finds application in circumstances such as those of this case, it is plain that the uniform implementation in the Member States of Security Council resolutions, which are binding on all members of the United Nations without distinction, can be better achieved at Community level than at national level’ (para 112). For comparative reflections with the US, see, eg, G Bermann, ‘Taking Subsidiarity Seriously: Federalism in the European Community and the United States’ (1994) 2 Columbia Law Review 331; Nicolaidis and Howse, above n 12; A Mills, ‘Federalism in the European Union and the United States: Subsidiarity, Private Law, and the Conflict of Laws’ (2010) 32 University of Pennsylvania Journal of International Law 369. For theoretical reflections on subsidiarity in the EU and in other settings, see, eg, A Føllesdal, ‘Subsidiarity and democratic deliberation’ in EO Eriksen and JE Fossum (eds), Democracy in the European Union. Integration through deliberation? (Abingdon, Routledge, 2000) 85–110; A Herwig, ‘Federalism, the EU and International law—On the Possible (and Necessary) Role of Subsidiarity in Legitimate Multilevel Trade Governance’, ch 3 of the present volume. 96 ERTA, above n 57, para 17. 97 Ibid para 22; see further Opinion 1/75 Draft Understanding on a Local Cost Standard drawn up under the auspices of the OECD [1975] ECR 1355, 1364.

152 Geert De Baere and Kathleen Gutman impose on the [Union] obligations which impose conditions on the exercise of prerogatives which thenceforth belong to the [Union] and which therefore no longer fall within the field of national sovereignty’.98 In other words, the Member States must not attempt to constrain the Union’s exercise of its exclusive competence and must co-operate loyally with it in order to facilitate such an exercise.99 Conversely, in addition to the shared and parallel competences provided for in the Treaties,100 non-exclusive external competence has also been recognised to exist by the Court on the basis of internal Union minimum standards: the concept of minimum requirements involves the Union harmonising a certain policy area on the basis of minimum standards, while leaving the Member States free to adopt more stringent measures. As the Court held in Opinion 1/03 Competence of the Community to conclude the new Lugano Convention on jurisdiction and the recognition and enforcement of judgments in civil and commercial matters, the fact that both the Union rules and the international agreement in question lay down minimum standards ‘may justify the conclusion that the [Union] rules are not affected, even if the [Union] rules and the provisions of the agreement cover the same area’.101 The requirement that both the Union rules and the international agreement in question lay down minimum standards is necessary in order not to inhibit the development of Union law. If an international agreement lays down an absolute standard, and the Union subsequently decides to raise its minimum standards above the absolute standard of the agreement, a conflict may arise, with inevitable consequences for the international responsibility of the Union. The foregoing discussion illustrates the inherent complexity of the EU’s competences in international relations as evidenced by the intricate relations between the Union and the Member States and the clear emphasis on non-exclusive international competences.

98 Ruling 1/78 Draft Convention of the International Atomic Energy Agency on the Physical Protection of Nuclear Materials, Facilities and Transports [1978] ECR 2151, para 32. 99 Cf ibid paras 33 and 22; Kramer, above n 57, paras 44–45. 100 It has also been pointed out that a Union competence may be non-exclusive simply in situations where international agreements cover areas that will naturally fall both within the competences of the Union and the Member States without the competence of one excluding the competence of the other. An example is the Vienna Convention on the Law of Treaties Between States and International Organizations or Between International Organizations (1986) 25 International Legal Materials 543. Both the Union and the Member States should in theory be able to accede to such an international agreement without in any way impinging on the competence of the other: cf I MacLeod, ID Hendry and S Hyett, The External Relations of the European Communities: A Manual of Law and Practice (Oxford, Clarendon Press, 1996) 66, fn 157. 101 Above n 93, paras 123 and 127. See also Opinion 2/91 Convention No 170 of the International Labour Organization concerning safety in the use of chemicals at work [1993] ECR I-1061, para 18. The Court further clarified the impact of minimum standards in PFOS, above n 11, para 102. Sweden had put forward the argument that a proposal to list a substance in the Annex to an international convention which was binding on the Union was equivalent to a national measure that was more stringent than a minimum Union measure and was permitted by then Art 176 TEC (now Art 193 TFEU). The Court disagreed. The Union could be bound by an amendment to an Annex to such a convention while it is not bound by such a national measure: see further De Baere, above n 11.

Federalism and International Relations in the EU and the US 153 IV. CONSTITUTIONAL FRAMEWORK OF FEDERALISM AND FOREIGN RELATIONS IN THE US

A. Overview of the Allocation of Competences in the US Constitution Similar to the Treaty framework governing the EU, the US Constitution does not have one ‘go-to’ provision that delineates the foreign relations powers of the federal government. Instead, the US Constitution grants a number of enumerated powers to each of the three branches of the federal government and contains some other provisions relating to the foreign relations field. Generally, attention rests on five main sets of provisions. First, under Article I, section 8, Congress is granted various powers concerning foreign relations, eg the power to regulate commerce with foreign nations, to declare war, and to provide for the army and the navy.102 Secondly, under Article II, the President is also granted various powers concerning foreign relations, eg he is the Commander in Chief of the army and the navy, and is granted the power (with the advice and consent of the Senate) to make treaties and to appoint ambassadors, ministers and consuls.103 Thirdly, under Article III, the powers of the federal judiciary include the adjudication of several kinds of cases implicating foreign relations, eg cases affecting ambassadors, ministers and consuls, and controversies between a state, or the citizens thereof, and foreign States, citizens or subjects.104 Fourthly, under the Supremacy Clause, treaties and other federal enactments regarding foreign relations are the ‘supreme Law of the Land’ and thus trump any conflicting state enactment.105 Lastly, under Article I, section 10, the US Constitution forbids the states from assuming considerable foreign relations powers, and places various limitations on them, eg providing that no state may enter into any treaty, alliance or confederation, or without the consent of Congress enter into any agreement or compact with a foreign power.106 Noticeably, looking closely at the enumerated powers granted to each branch of the federal government in the US Constitution, many issues are not dealt with expressly. To take a classic example, the fact that the President is granted the power to make a treaty (with the advice and consent of the Senate) does not speak to the question whether he has the power to terminate or suspend such a treaty, or to

102 See US Constitution, Art I, s 8, particularly cls 1–5, 10–16. Moreover, under the so-called ‘Necessary and Proper Clause’, enshrined in cl 18 of the same section, Congress is granted the power ‘[t]o make all Laws which shall be necessary and proper for carrying into Execution the foregoing Powers, and all other Powers vested by this Constitution in the Government of the United States, or in any Department or Officer thereof.’ 103 See ibid, Art II, particularly s 1, cl 1; s 2, cls 1–2; s 3. 104 See ibid, Art III, particularly s 2, cls 1–2. 105 Ibid, Art VI, cl 2, provides: ‘This Constitution, and the Laws of the United States which shall be made in Pursuance thereof; and all Treaties made, or which shall be made, under the Authority of the United States, shall be the supreme Law of the Land; and the Judges of every State shall be bound thereby, any Thing in the Constitution or Laws of any State to the Contrary notwithstanding.’ 106 See ibid, Art I, s 10, cls 1–3.

154 Geert De Baere and Kathleen Gutman make other kinds of international agreements with or without the involvement of Congress.107 As a result, various methods have been used to fill the gaps in the US Constitution, eg implied powers have been read into the express powers set forth in the US Constitution, or several powers taken together have been interpreted to lead to others, although ultimately, there have been difficulties in finding a generally accepted strategy by which to account for all of the foreign relations powers wielded by the federal government.108 Nevertheless, the predominant view is that the federal government is considered to be attributed plenary and exclusive power over foreign relations, meaning that it has been vested with comprehensive power to conduct foreign relations to the exclusion of the states.109 This view is based on the various provisions of the US Constitution mentioned above, as well as language running through a number of Federalist papers110 and seminal cases of the US Supreme Court, such as United States v Belmont,111 proclaiming that ‘complete power over international affairs is in the national government and is not and cannot be subject to any curtailment or interference on the part of the several states … [I]n respect of our foreign relations generally, state lines disappear.’112 In fact, the interpretation of the various provisions enshrined in the US Constitution has led to the recognition of what has been called ‘foreign affairs

107

See Restatement (Third) of Foreign Relations Law, above n 21, §§ 303, 339. L Henkin, Foreign Affairs and the United States Constitution, 2nd edn (Oxford, Clarendon Press, 1996) 13–15. 109 See, eg, ibid 14: ‘No explanation has been universally accepted, and no proposed principle of constitutional construction has supplied what is missing to universal satisfaction. Different doctrines that have been suggested have different legal and political consequences. Under all of them, however, foreign affairs remains exclusively national.’ See also M Schaefer, ‘Constraints on State-Level Foreign Policy: (Re)Justifying, Refining and Distinguishing the Dormant Foreign Affairs Doctrine’ (2011) 41 Seton Hall Law Review 201, 222 (‘no one seriously questions that the federal government’s foreign affairs powers are plenary; thus, any alleged gaps in the textual allocations of powers to federal actors are not considered gaps in substance’). Indeed, there is usually no dispute in the commentary regarding the plenary nature of federal power over foreign relations, in the sense that the federal government has comprehensive power to conduct foreign relations without interference or limitation by the states, whereas the exclusive nature of such power in the absence of a controlling federal foreign relations enactment has been challenged, thereby explaining why a distinction is drawn between the two: JL Goldsmith, ‘Federal Courts, Foreign Affairs, and Federalism’ (1997) 83 Virginia Law Review 1617, 1619–20; see further below n 126. 110 See, eg, EH Scott (ed), The Federalist and Other Constitutional Papers (Chicago, Ill, Scott, Foresman & Co, 1898), Federalist No 42, at 232 (‘If we are to be one nation in any respect, it clearly ought to be in respect of other nations.’) (James Madison); Federalist No 80, at 435 (‘[T]he peace of the whole, ought not to be left at the disposal of a part.’) (Alexander Hamilton). For reference to others, see, eg, Hines v Davidowitz, 312 US 52, 62 fn 9 (1941). 111 United States v Belmont, 301 US 324 (1937). 112 Ibid 331 (citations omitted). See also, eg, United States v Pink, 315 US 203, 233 (1942) (‘Power over external relations is not shared by the [s]tates; it is vested in the national government exclusively.’); Hines v Davidowitz, 312 US 52, 63 (1941) (‘The Federal Government, representing as it does the collective interests of the … states, is entrusted with full and exclusive responsibility for the conduct of affairs with foreign sovereignties … Our system of government is such that the interest of the cities, counties and states, no less than the interest of the people of the whole nation, imperatively requires that federal power in the field affecting foreign relations be left entirely free from local interference.’) 108

Federalism and International Relations in the EU and the US 155 exceptionalism’,113 denoting the idea that federal regulation of foreign affairs is subject to a different, and generally more relaxed, set of constitutional constraints than that of domestic or internal affairs.114 For example, in Missouri v Holland,115 the US Supreme Court held that the treaty-making power,116 by which the President makes treaties with the advice and consent of the Senate, is not limited to matters falling within Congress’s enumerated powers in the US Constitution.117 Likewise, in United States v Curtiss-Wright Export Corp,118 the US Supreme Court, in a well-known opinion delivered by Justice Sutherland, declared that the broad statement that the federal government can exercise only those powers specifically enumerated in the US Constitution, or implied therefrom, is categorically true only with respect to internal affairs, and the federal government’s powers over foreign affairs do not depend on affirmative grants in the US Constitution: ‘The powers to declare and wage war, to conclude peace, to make treaties, to maintain diplomatic relations with other sovereignties, if they had never been mentioned in the Constitution, would have vested in the federal government as necessary concomitants of nationality.’119 At the same time, however, it is widely recognised that foreign relations are not ‘wholly insulated from the states’ and that they do take actions that affect foreign relations in various ways.120 In recent decades, fuelled by such factors as globalisation,121 the involvement of the states (as well as other sub-national

113 To be sure, the term ‘exceptionalism’ has been used to mean other things in the American setting, often relating to matters concerning the US position vis-à-vis international law: see J Resnik, ‘Law’s Migration: American Exceptionalism, Silent Dialogues, and Federalism’s Multiple Ports of Entry’ (2006) 115 Yale Law Journal 1564, 1582–84 and citations therein. 114 CA Bradley, ‘The Treaty Power and American Federalism’ (1998) 97 Michigan Law Review 390, 391, 461. See further, eg, CA Bradley, ‘International Delegations, The Structural Constitution, and Non-Self-Execution’ (2003) 55 Stanford Law Review 1557, 1582–86; EA Young, ‘Dual Federalism, Concurrent Jurisdiction, and the Foreign Affairs Exception’ (2001) 69 George Washington Law Review 139. Yet as evidenced by the above cited publications, the recognition of such differential treatment in foreign affairs has been accompanied by calls for its reassessment. Indeed, opposition to ‘foreign relations exceptionalism’ has been considered the ‘key animating principle’ of the ‘revisionist’ scholarship: J Ku, ‘The State of New York Does Exist: How the States Control Compliance with International Law’ (2004) 82 North Carolina Law Review 457, 470, fn 63. See further below n 126. 115 Missouri v Holland, 252 US 416 (1920). 116 US Constitution, Art II, s 2, cl 2. 117 Missouri v Holland, above n 115, at 432–35. That said, this case has been subject to critical inquiry by scholars: see below n 126. 118 United States v Curtiss-Wright Export Corp, 299 US 304 (1936). 119 Ibid 315–18. Justice Sutherland’s so-called ‘extraconstitutional’ reading has been vigorously challenged in the commentary: see, eg, MD Ramsey, ‘The Myth of Extraconstitutional Foreign Affairs Power’ (2000) 42 William and Mary Law Review 379 and citations therein. For a broader discussion, see, eg, SH Cleveland, ‘Powers Inherent in Sovereignty: Indians, Aliens, Territories, and the Nineteenth Century Origins of Plenary Power over Foreign Affairs’ (2002) 81 Texas Law Review 1; GE White, ‘The Transformation of the Constitutional Regime of Foreign Relations’ (1999) 85 Virginia Law Review 1. 120 Henkin, above n 108, 150. 121 As regards this term and its potential implications in this setting, see, eg, PJ Spiro, ‘Globalization and the (Foreign Affairs) Constitution’ (2002) 63 Ohio State Law Journal 649.

156 Geert De Baere and Kathleen Gutman actors)122 in matters affecting foreign relations has intensified. Not only are the states assuming a more prominent role in their own right on the international stage, but also, on account of the interaction between the ‘foreign’ and the ‘domestic’ spheres, traditional activities of the states increasingly impact foreign relations.123 Examples range from state laws incorporating provisions of international conventions not yet ratified by the federal government, to various kinds of agreements entered into by the states with foreign governments in pursuit of economic, political or social objectives.124 These developments have produced a rising volume of literature directed at issues relating to ‘foreign affairs federalism’,125 prompting re-examination of traditional assumptions about the scope of federal competence in foreign relations on numerous fronts126 and enlivening discussions as to the extent to which numerous kinds of actions taken by the states in matters concerning foreign relations are permitted under the constraints 122 In the US, the discussion of federalism and foreign relations includes actions taken by the states, as well as other sub-national or sub-federal actors, eg city and county municipalities, either alone or as part of a variety of networks: see, eg, J Resnik, ‘The Internationalism of American Federalism: Missouri and Holland’ (2008) 73 Missouri Law Review 1105 and citations therein. For brevity’s sake, however, this chapter generally refers to the states. In the EU, the discussion of federalism and external relations encompasses inquiry into actions taken by the Member States, as well as sub-national or devolved authorities within the Member States and even overseas territories linked thereto. See, eg, S Blockmans, ‘Between the Devil and the Deep Blue Sea? Conflicts in External Action Pursued by OCTs and the EU’ in D Kochenov (ed), EU Law of the Overseas—Outermost Regions, Associated Overseas Countries and Territories, Territories Sui Generis (The Hague, Kluwer, 2011) 307. 123 N Robinson, ‘Citizens Not Subjects: US Foreign Relations Law and the Decentralization of Foreign Policy’ (2007) 40 Akron Law Review 647, 648. 124 See, eg, Resnik, above n 122; DB Hollis, ‘Unpacking the Compact Clause’ (2010) 88 Texas Law Review 741. State activities concerning climate change have warranted particular attention: see, eg, D Vogel and JFM Swinnen (eds), Transatlantic Regulatory Cooperation—The Shifting Roles of the EU, the US and California (Cheltenham and Northampton, Edward Elgar, 2011), particularly ‘Part I: Federalism and Cooperation at the National and International Levels’; Symposium, ‘Federalism and Climate Change: The Role of the States in a Future Federal Regime’ (2008) 50 Arizona Law Review 673; Note, ‘Foreign Affairs Preemption and State Regulation of Greenhouse Gas Emissions’ (2006) 119 Harvard Law Review 1877. 125 For other factors relevant in this context, such as the US Supreme Court’s revival of federalism restrictions in the domestic sphere, see, eg, CA Bradley, ‘A New American Foreign Affairs Law?’ (1999) 70 University of Colorado Law Review 1089. 126 Those challenging the exclusive nature of federal competence in foreign relations are sometimes portrayed as taking a ‘revisionist’ position, as compared to a ‘nationalist’ or ‘orthodox’ view: see, eg, Goldsmith, above n 109; Ku, above n 114, 465–76; JJ Pascoe, ‘Time for a New Approach? Federalism and Foreign Affairs after Crosby v National Foreign Trade Council’ (2002) 35 Vanderbilt Journal of Transnational Law 291, 303–08; see also MD Ramsey, ‘Review Essay: Textbook Revisionism’ (2003) 43 Virginia Journal of International Law 1111, 1116–19 (exploring foreign relations ‘revisionism’). Such discussions are often situated within the context of particular doctrines or powers. For instance, as regards the treaty-making power, see, eg, Symposium, ‘Return to Missouri v Holland: Federalism and International Law’ (2008) 73 Missouri Law Review 921; DB Hollis, ‘Executive Federalism: Forging New Federalist Constraints on the Treaty Power’ (2006) 79 Southern California Law Review 1327, 1330–52 (providing a summary of the arguments of the so-called ‘nationalist’ and ‘new federalist’ camps). Another example may be found within the context of the doctrine of American ‘federal common law’ (see further citations below n 129) with regard to the status of customary international law: for a recent summary of the various positions and relevant literature, see CM Vázquez, ‘Customary International Law as US Law: A Critique of the Revisionist and Intermediate Positions and a Defense of the Modern Position’ (2011) 86 Notre Dame Law Review 1495.

Federalism and International Relations in the EU and the US 157 imposed by the American constitutional framework,127 which in turn invites inquiry into the case law of the US Supreme Court on such matters.

B. Case Law of the US Supreme Court Given that, as noted above, the US Constitution grants a number of enumerated powers concerning foreign relations to each of the legislative, executive and judicial branches of the federal government, there is a considerable body of case law of the US Supreme Court, not to mention the lower federal courts and the state courts, relevant to the discussion of the federal–state relationship in foreign relations, since depending upon the particular provision concerned, such case law arises in a variety of contexts128 and dovetails with so many doctrines,129 the analysis of which far exceeds the scope of this chapter. Therefore, for the purposes of the present discussion, focus is placed on certain landmark cases of the US Supreme Court that are of crucial importance for highlighting the limitations or constraints placed on state action in matters affecting foreign relations and the ambit for state autonomy in this setting. At the forefront of the constraints placed on the states in the field of foreign relations stands the dormant foreign affairs doctrine.130 This doctrine prohibits state action in foreign relations in the absence of a federal enactment in the field, and thus is rooted in the exclusive nature of federal competence over foreign relations, the idea being that since such competence is exclusive, state intrusions in foreign relations are prohibited even where the federal government has not acted (the power lays ‘dormant’).131 This doctrine was established in Zschernig v Miller,132 127 For typologies of various forms of state action implicating foreign relations, see, eg, Granne, above n 27, 872–76; Robinson, above n 123, 687–713. 128 Eg a notable area of jurisprudence concerns the dormant foreign commerce clause: see, eg, AJ Colangelo, ‘The Foreign Commerce Clause’ (2010) 96 Virginia Law Review 949; LM Wilson, ‘The Fate of the Dormant Foreign Commerce Clause after Garamendi and Crosby’ (2007) 107 Columbia Law Review 746; Schaefer, above n 109. As illustrated by some of the cases discussed below, this clause may constitute an independent ground (in addition to, inter alia, the dormant foreign affairs and pre-emption doctrines) by which to challenge state laws affecting foreign relations: see below nn 148 and 157. 129 Eg the doctrine of American ‘federal common law’: see generally RH Fallon, Jr, JF Manning, DJ Meltzer and DL Shapiro, Hart and Wechsler’s The Federal Courts and the Federal System, 6th edn (New York, Reuters/Foundation Press, 2009) ch 7, particularly 665–79. For a comparative analysis, see K Lenaerts and K Gutman, ‘“Federal Common Law” in the European Union: A Comparative Perspective from the United States’ (2006) 54 American Journal of Comparative Law 1. 130 The dormant foreign affairs doctrine (or power, sometimes unaccompanied by the ‘dormant’ label) has also been referred to as ‘dormant foreign affairs preemption’ by commentators and members of the US Supreme Court (see, eg, text accompanying n 169 below). This is not done here in order to avoid confusion between this doctrine and the preemption doctrine, even if as discussed below, they have become intertwined through the course of US Supreme Court jurisprudence. 131 Schaefer, above n 109, 204, 225. 132 Zschernig v Miller, 389 US 429 (1968). The judgment was delivered by eight, as opposed to nine, justices (Justice Marshall did not take part in the case), and of those, it claimed an overriding majority of seven justices, with Justice White dissenting. In addition to Justice Stewart (joined by Justice Brennan), Justice Harlan also issued a concurring opinion, though he parted ways, inter alia, with the

158 Geert De Baere and Kathleen Gutman decided by the US Supreme Court in 1968. Situated within the context of the Cold War, this case concerned a challenge to the constitutionality of an Oregon inheritance reciprocity statute generally requiring foreign persons inheriting property in that state to prove that there was a reciprocal right to inherit property in their country and that the property would not be confiscated.133 The Supreme Court had previously upheld a similar statute in Clark v Allen,134 but the problem here, in its view, was the way in which the Oregon statute was being applied, since it had led to inquiries by the state probate courts into the actual administration of foreign law and the credibility of foreign diplomatic statements,135 which had a ‘direct impact upon foreign relations’ and illustrated ‘the dangers which are involved if each [s]tate, speaking through its probate courts, is permitted to establish its own foreign policy’.136 Thus, despite assertions by the federal government to the contrary,137 the Supreme Court held that the Oregon statute was ‘an intrusion by the [s]tate into the field of foreign affairs which the Constitution entrusts to the President and the Congress’.138 In doing so, it emphasised that although the states have traditionally regulated matters concerning the descent and distribution of estates, ‘those regulations must give way if they impair the effective exercise of the Nation’s foreign policy’.139 Thereafter, this judgment eventually found its way into a related line of cases concerning the pre-emption doctrine. This doctrine flows from a positive federal act in combination with the Supremacy Clause,140 providing that the US Constitution, the federal laws adopted pursuant thereto and treaties made under the authority of the US are the ‘supreme Law of the Land’ and override any conflicting state enactment.141 Under this doctrine, state action may be pre-empted expressly or implicitly by a particular federal act, as well as in circumstances in which such act has ‘occupied the field’ or the state action concerned is deemed to conflict directly with such act or otherwise constitutes an obstacle to the achievement of the federal objectives embodied therein.142

majority’s decision concerning the dormant foreign affairs doctrine: see ibid 457–62 (Harlan J, concurring). His reasoning found its way into recent case law: see text accompanying nn 159–63 below. 133

Ibid 430–31 (majority opinion). Clark v Allen, 331 US 503 (1947). Justice Douglas delivered the opinion of the Supreme Court in this case, as well as in Zschernig. There were no dissenting opinions, although Justice Rutledge concurred in part. 135 Zschernig, above n 132, at 435. For examples, see ibid 436–40. 136 Ibid 441. 137 Ibid 434. 138 Ibid 432. 139 Ibid 440. 140 LH Tribe, American Constitutional Law, 3rd edn (New York, Foundation Press, 2000) vol 1, § 6-28, at 1172. 141 See above n 105. 142 Tribe, above n 140, § 6-28, at 1176–77; § 6-31, 1204–05. As highlighted by the aforementioned commentator and the US Supreme Court, however, the categories of express, implied, field and conflict pre-emption are not discrete, and field pre-emption may be considered to fall into the others: see Crosby v National Foreign Trade Council, 530 US 363, 373 fn 6 (2000) and citations therein. 134

Federalism and International Relations in the EU and the US 159 Decided in 2000, Crosby v National Foreign Trade Council143 confronted the Supreme Court with a challenge to the constitutionality of the Massachusetts Burma Law, a procurement statute barring state entities from buying goods or services from companies doing business with Burma (now Myanmar), save for some limited exceptions.144 Massachusetts was not alone; as noted in the Supreme Court’s opinion, at least 19 other state and local governments had adopted similar laws.145 The ‘rub’ in the case was that a few months after the Massachusetts statute was enacted, Congress passed a federal statute imposing a set of mandatory sanctions on Burma and authorising the President to impose further sanctions subject to certain conditions, thereby directing him to work to develop a comprehensive, multilateral strategy to bring democracy and improve human rights practices in Burma.146 In a unanimous judgment of all nine justices,147 the Supreme Court held that the Massachusetts statute was pre-empted under the Supremacy Clause because it stood as an obstacle to the accomplishment of Congress’s objectives under the federal statute.148 As fleshed out in separate sections of the judgment, the Supreme Court found that the state statute undermined the intended purpose and ‘natural effect’ of at least three provisions of the federal statute: its delegation of effective discretion to the President to control economic sanctions against Burma, its limitation of sanctions solely to US persons and new investment, and its directive to the President to proceed diplomatically in developing a comprehensive, multilateral strategy toward Burma.149 In its view, by imposing a different state system of economic pressure against the Burmese regime, ‘the President has less to offer and less economic and diplomatic leverage as a consequence’,150 and the statute conflicted with the federal scheme in a number of ways by penalising individuals and conduct that Congress explicitly exempted or excluded from sanctions.151

143 Crosby, above n 142. For detailed discussion, see, eg, Symposium, ‘New Voices on the New Federalism’ (2001) 46 Villanova Law Review 907; RJ Delahunty, ‘Federalism Beyond the Water’s Edge: State Procurement Sanctions and Foreign Affairs’ (2001) 37 Stanford Journal of International Law 1; J Goldsmith, ‘Statutory Foreign Affairs Preemption’ (2000) Supreme Court Review 175; Pascoe, above n 126; M Tushnet, ‘Globalization and Federalism in a Post-Printz World’ (2000) 36 Tulsa Law Journal 11. 144 See Crosby, above n 142, at 366–68. 145 See ibid 372, fn 5. 146 See ibid 368–70. 147 The opinion was delivered by Justice Souter; there was a concurring opinion by Justice Scalia (joined by Justice Thomas), which dealt with the majority’s use of legislative history: ibid 388–91 (Scalia J, concurring). 148 Ibid 373. Notably, the Supreme Court sidestepped issues concerning the dormant foreign affairs doctrine and the dormant foreign commerce clause, which had been dealt with in the proceedings before the lower federal courts, preferring to deal only with the pre-emption claim. It also sidestepped the issue of whether a presumption against pre-emption should apply in this context, though noting that even assuming such a presumption was appropriate, the statute was nonetheless pre-empted. See ibid 374, fn 8. On this latter point, see further Wyeth v Levine, 555 US 555, 624 fn 14 (2009) (Alito J, dissenting). 149 Crosby, above n 142, at 373. 150 Ibid 376–77. 151 Ibid 378.

160 Geert De Baere and Kathleen Gutman Moreover, the Supreme Court considered that the statute hindered the President’s capacity for effective diplomacy, underlining that ‘[i]t is not merely that the differences between the state and federal [a]cts in scope and type of sanctions threaten to complicate discussions; they compromise the very capacity of the President to speak for the Nation with one voice in dealing with other governments.’152 A few years later, in 2003, the Supreme Court delivered its judgment in American Insurance Association v Garamendi,153 concerning a constitutional challenge to the California Holocaust Victim Insurance Relief Act (HVIRA), which required insurance companies doing business in the state to disclose information about insurance policies sold in Europe during the Holocaust era.154 The California statute was not adopted in a vacuum. International efforts were already underway to deal with claims stemming from the confiscation of Jewish life insurance policies issued before and during World War Two, which led to the adoption of executive agreements between the President and several European countries to set up a foundation that was intended to be the exclusive forum for victims to receive compensation and, as regards insurance claims, to work with a voluntary organisation, the International Commission on Holocaust Era Insurance Claims, responsible for providing information and settling such claims.155 Still, the California legislature found the adoption of the statute necessary to protect the claims and interests of its residents, as did other states.156 Compared to Crosby, this time it was a ‘close call’, but by a narrow 5:4 majority, the Supreme Court held that the California statute was pre-empted on account of its interference with the federal government’s conduct of foreign relations.157 Since there was no express pre-emption clause in the executive agreements concerned, the pre-emption claim rested on the asserted interference with the foreign policy those agreements embodied, which turned the Supreme Court’s attention to Zschernig.158 It compared the majority’s opinion in that case with Justice Harlan’s concurrence, which disagreed with such reasoning but would have found 152 Ibid 381. In this regard, the Supreme Court cited protests by various US allies and trading partners against the Massachusetts statute to support the fact that it threatened the President’s powers to speak and to bargain effectively with other nations, which included submissions by the EU to the federal government and the lodging of complaints by the EU and Japan with the World Trade Organization: see ibid 382–86. 153 American Insurance Association v Garamendi, 539 US 396 (2003). For detailed discussion, see, eg, JB Crace, Jr, ‘Gara-Mending the Doctrine of Foreign Affairs Preemption’ (2004) 90 Cornell Law Review 203; BP Denning and MD Ramsey, ‘American Insurance Association v Garamendi and Executive Preemption in Foreign Affairs’ (2004) 46 William and Mary Law Review 825; Note, ‘Foreign Affairs Preemption and State Regulation of Greenhouse Gas Emissions’, above n 124. 154 See Garamendi, above n 153, at 409–11. 155 See ibid 401–08. 156 Ibid 411, 413, fn 6. 157 Ibid 401. Justice Souter delivered the majority opinion (joined by then Chief Justice Rehnquist, as well as Justices O’Connor, Kennedy and Breyer). Although the grant of certiorari encompassed three questions (whether the HVIRA intrudes on the federal foreign affairs power, violates the foreign commerce clause and exceeds the state’s ‘legislative jurisdiction’), the Supreme Court chose to deal only with the first: see ibid 413, fn 7. 158 Ibid 417.

Federalism and International Relations in the EU and the US 161 pre-emption in a case of ‘conflicting federal policy’.159 This led the Supreme Court to surmise that ‘[i]t is a fair question whether respect for the executive foreign relations power requires a categorical choice between the contrasting theories of field and conflict preemption evident in the Zschernig opinions’.160 It decided, however, that there was no need to answer this question because, even on Justice Harlan’s view, the likelihood that a state law will produce ‘something more than incidental effect in conflict with express foreign policy’ of the federal government would require pre-emption of that law.161 Nevertheless, in line with his view, the Supreme Court stated that ‘it would be reasonable to consider the strength of the state interest, judged by standards of traditional practice, when deciding how serious a conflict must be shown before declaring the state law preempted’.162 Judged by these standards, it ruled that there was a sufficiently clear conflict to justify pre-emption here.163 In particular, the Supreme Court emphasised that the situation created by the California statute resembled the Massachusetts statute at issue in Crosby ‘in the way HVIRA threatens to frustrate the operation of the particular mechanism the President has chosen’.164 Furthermore, the statute’s economic compulsion to make public disclosure of far more information about far more policies than the International Commission rules undercut the President’s diplomatic discretion and the choice he had made in exercising it, thereby compromising his ‘very capacity to speak for the Nation with one voice’ in dealing with other countries in order to resolve claims against European companies arising out of World War Two.165 It then pointed to the ‘weakness of the [s]tate’s interest, against the backdrop of traditional state legislative subject matter, in regulating disclosure of European Holocaust-era insurance policies in the manner of HVIRA’.166 159

Ibid 418–19. Ibid 420. In an accompanying footnote, the Supreme Court posited that these two positions could be considered complementary: ‘If a [s]tate were simply to take a position on a matter of foreign policy with no serious claim to be addressing a traditional state responsibility, field preemption might be the appropriate doctrine, whether the National Government had acted and, if it had, without reference to the degree of any conflict, the principle having been established that the Constitution entrusts foreign policy exclusively to the National Government. Where, however, a [s]tate has acted within what Justice Harlan called its “traditional competence”, but in a way that affects foreign relations, it might make good sense to require a conflict of a clarity or substantiality that would vary with the strength of the traditional importance of the state concern asserted. Whether the strength of the federal foreign policy interest should itself be weighed is, of course, a further question.’ (ibid 420, fn 11 (citations omitted)) 161 Ibid 419–20. 162 Ibid 420. 163 Ibid. 164 Ibid 423–25. 165 Ibid 424 (quoting Crosby, above n 142, at 381). The Supreme Court admitted that in the present case the President was acting without any congressional authority and thus did not have the ‘plenitude of Executive authority’ that controlled the issue of pre-emption in Crosby; however, in Crosby, ‘we were careful to note that the President possesses considerable independent constitutional authority to act on the behalf of the United States on international issues, and conflict with the exercise of that authority is a comparably good reason to find preemption of state law’ (ibid 424, fn 14 (citations omitted)). 166 Ibid 425–27. 160

162 Geert De Baere and Kathleen Gutman In well-cited language, the Supreme Court paid heed to the fact that California’s strategy of using an ‘iron fist’ might work better than the President’s choice of ‘kid gloves’ in addressing the matter concerned, but reiterated that the question relevant to pre-emption in this case was conflict, and the evidence amply showed that the California statute stood in the way of the President’s diplomatic objectives.167 By comparison, in her dissenting opinion, Justice Ginsburg argued that although the federal approach differed from California’s, there was no executive agreement or other formal expression of disapproval for state disclosure laws like the one at issue, and therefore, absent a clear statement to this effect, the state statute should be left intact.168 In doing so, she challenged Zschernig’s application to the present proceedings: We have not relied on Zschernig since it was decided, and I would not resurrect that decision here. The notion of ‘dormant foreign affairs preemption’ with which Zschernig is associated resonates most audibly when state action ‘reflect[s] a state policy critical of foreign governments and involve[s] “sitting in judgment” on them.’ The HVIRA entails no such state action or policy. It takes no position on any contemporary foreign government and requires no assessment of any existing foreign regime. It is directed solely at private insurers doing business in California, and it requires them solely to disclose information in their or their affiliates’ possession or control. I would not extend Zschernig into this dissimilar domain.169

In Justice Ginsburg’s view, upholding the HVIRA ‘would not compromise the President’s ability to speak with one voice for the Nation’, and foreign affairs pre-emption should be reserved for circumstances in which the President, acting under statutory or constitutional authority, has spoken clearly to the issue at hand.170 Recently, the Supreme Court’s 2008 judgment in Medellín v Texas171 brought to the fore the ‘flip side’ of ‘foreign relations federalism’ in connection with the ambit and consequences of state autonomy in this field. By way of brief background,172 in the wake of brutal crimes committed by several gang members in Texas, Medellín, a Mexican national residing in the US, was arrested and confessed to

167

Ibid 427 (quoting Crosby, above n 142, at 386). Ibid 430 (Ginsburg J, dissenting). 169 Ibid 439–40 (citations omitted). In an accompanying footnote, Justice Ginsburg also criticised the majority’s reliance on Crosby, a statutory pre-emption case that ‘provides little support for pre-empting a state law by inferring preclusive foreign policy objectives from precatory language in executive agreements’ (ibid 440, fn 4). 170 Ibid 442. 171 Medellín v Texas, 552 US 491 (2008). For a selection of literature devoted to this case and its broader background, see Agora, Medellín (2008) 102 American Journal of International Law 529; Symposium, Medellín v Texas (2008) 31 Suffolk Transnational Law Review 209; Colloquy, (2010) 44 Valparaiso University Law Review 759; T Cruz, ‘Defending US Sovereignty, Separation of Powers, and Federalism in Medellín v Texas’ (2010) 33 Harvard Journal of Law and Public Policy 25; EA Young, ‘Treaties as “Part of Our Law” ’ (2009) 88 Texas Law Review 91. 172 See Medellín, above n 171, at 497–504. 168

Federalism and International Relations in the EU and the US 163 such crimes, but was not advised of his right to notify the Mexican consulate of his detention in accordance with the Vienna Convention on Consular Relations (‘Vienna Convention’).173 He was convicted of capital murder and sentenced to death by a Texas state court. In his subsequent attempts to challenge his death sentence, he raised a claim based on the Vienna Convention, but under a Texas procedural default rule, such claim was precluded because he had failed to raise it at trial or on direct review. In the meantime, the International Court of Justice (ICJ) delivered its judgment in Avena,174 ruling that the US violated the Vienna Convention by failing properly to inform several Mexican nationals awaiting execution, including Medellín, of their right to speak with a consular official after being detained, and that such nationals were entitled to review and reconsideration of their state court convictions and sentences. Following Avena, then President Bush issued a Memorandum to the Attorney General, stating that the US ‘will discharge its international obligations’ under Avena ‘by having [s]tate courts give effect to the decision in accordance with general principles of comity in cases filed by the 51 Mexican nationals addressed in that decision’.175 Consequently, this case confronted the US Supreme Court with two main questions concerning whether the Avena decision was directly enforceable as domestic law in state and federal courts, and whether the Presidential Memorandum independently required the states to provide review and reconsideration of the claims of the Mexican nationals named in Avena without regard to state procedural default rules.176 In an opinion delivered by Chief Justice Roberts,177 the Supreme Court held that neither the Avena judgment nor the Presidential Memorandum constituted directly enforceable federal law that pre-empted the state rules concerned.178 First, although the Avena judgment created an international law obligation on the part of the US, it did not have automatic legal effect such that it applied of its own force in state and federal courts.179 In this regard, the Supreme Court rejected arguments that several relevant treaty sources (ie the Optional Protocol to the Vienna Convention,180 the UN Charter and the ICJ Statute) supplied the ‘relevant obligation’ to give the Avena judgment binding effect in domestic courts, 173 In 1969, the US ratified the Vienna Convention on Consular Relations, 24 April 1963, [1970] 21 United States Treaties and Other International Agreements (UST) 77, and the Optional Protocol Concerning the Compulsory Settlement of Disputes to the Vienna Convention (‘Optional Protocol’), 24 April 1963, [1970] 21 UST 325. By ratifying the Optional Protocol, the US consented to the specific jurisdiction of the ICJ with respect to claims arising out of the Vienna Convention; however, in 2005 (after the ICJ’s judgment in Avena: see text accompanying n 174 below), the US gave notice of withdrawal from the Optional Protocol: Medellín, above n 171, at 500. 174 Avena and Other Mexican Nationals (Mexico v United States of America), Judgment [2004] ICJ Reports 12. 175 Medellín, above n 171, at 503. 176 Ibid 498. 177 The opinion garnered a 6:3 majority, with Justice Stevens concurring and Justice Breyer (joined by Justices Souter and Ginsburg) dissenting. 178 Medellín, above n 171, at 498–99. 179 See ibid 504–22. 180 See above n 173.

164 Geert De Baere and Kathleen Gutman finding that none of these sources created binding federal law in the absence of implementing legislation.181 Secondly, the Avena judgment was not binding on state courts by virtue of the Presidential Memorandum.182 The Supreme Court rejected arguments that, inter alia, the Memorandum was a valid exercise of the President’s foreign affairs authority to resolve claims disputes with foreign nations, underlining that the President’s narrow and strictly limited authority to settle international claims disputes pursuant to an executive agreement, as illustrated by cases including Garamendi, could not be stretched so far as to support the Memorandum.183 In closing, it stressed that the Memorandum constituted an ‘unprecedented action’ that ‘reaches deep into the heart of the [s]tate’s police powers and compels state courts to reopen final criminal judgments and set aside neutrally applicable state laws’.184 To be sure, as noted by one commentator, Medellín ‘implicated virtually every conceivable axis of the structural limitations on government’.185 Still, federalism permeated this case. At base, the Supreme Court’s judgment signified deference to state (criminal procedural) law in the absence of implementing legislation, but with an awkward result for US foreign relations. Indeed, both Justice Steven’s concurrence and Justice Breyer’s dissent emphasised the extent to which Texas’s non-compliance with the ICJ’s judgment would jeopardise American dealings with foreign governments.186 As matters stand now, it remains an open question whether other cases allowing for state autonomy in the face of federal power over foreign relations will follow.187 Taken together, the foregoing case law attests to ‘the difficulty of finding the proper constitutional balance among state and federal’ powers in foreign relations.188 On the one hand, Crosby and Garamendi constitute salient examples of state legislation struck down on the ground that it was found to interfere directly with federal objectives in foreign relations.189 On the other hand, the course of this case law may arguably be read to allow for a certain degree of leeway for action taken by the states in matters affecting foreign relations. Notably, in

181

Medellín, above n 171, at 506–14. See ibid 523–32. Ibid 530–32. The Supreme Court also rejected arguments that the relevant treaties gave the President the authority to implement the Avena judgment and that Congress had acquiesced in the exercise of such authority: see ibid 526–30. Moreover, it swiftly dismissed the assertion that the Memorandum was a valid exercise of the President’s ‘take care’ power: see ibid 532. 184 Ibid 532. Medellín’s subsequent attempts to stay execution were unsuccessful: see Medellín v Texas, 554 US 759 (2008); Ex parte Medellín, 280 SW 3d 854 (Tex Crim App, 2008), cert denied, 555 US 922 (2008). Ultimately, he was executed. 185 Cruz, above n 171, 25. 186 See Medellín, above n 171, at 536–37 (Stevens J, concurring) and at 565–66 (Breyer J, dissenting). 187 Eg a key issue left open by this case concerns the application of the anti-commandeering doctrine in the context of foreign relations: see, eg, C Jackson, ‘The Anti-Commandeering Doctrine and Foreign Policy Federalism—The Missing Issue in Medellín v Texas’ (2008) 31 Suffolk Transnational Law Review 335. 188 Medellín, above n 171, at 565 (Breyer J, dissenting). 189 See Chamber of Commerce of the United States v Whiting, 131 S Ct 1968, 1983 (2011). 182 183

Federalism and International Relations in the EU and the US 165 Garamendi, despite the disagreement between the majority and the dissent, both appeared to be united in scaling back to some extent the dormant foreign affairs doctrine and the federal pre-emption of state action in this context, as evidenced by the majority’s framing of Zschernig as pre-emption requiring the showing of a clear conflict with federal foreign relations policy in this case and its introduction of a balancing test taking into account the strength of the state interest as part of the assessment, as well as by the dissent’s emphasis on the limited application of Zschernig and the need for a clear statement from the federal level to justify the pre-emption of state legislation. Medellín also tightened the reins on the scope of executive foreign relations power capable of overriding action taken by the states, especially taking its treatment of Garamendi into account.190 Certainly, the issues presented in Medellín differed from those in Zschernig, Crosby and Garamendi, and these cases may be viewed from a number of angles. Nevertheless, akin to the ECJ jurisprudence in this setting, the cases discussed above help to illuminate the US Supreme Court’s important role in ‘filling out’ the constitutional dimensions of the interaction between the federal and state levels in foreign relations.

V. CONCLUSION

In view of the foregoing analysis, the approach taken to federalism and international relations in the EU and the US reveals some remarkable parallels. First, the examination of the conceptual framework underlying this subject indicates that the EU and the US are not too distinct for meaningful comparison. Although the EU and the US constitute different kinds of federal systems, at least in functional terms, they are both grappling with the common challenge underlying federalism of balancing the need to ensure a coherent international relations policy at the central level and the allowance for some degree of autonomy for the component states to act in this field in order to protect their own interests. Secondly, inquiry into the constitutional frameworks governing international relations in the EU and the US highlights that parsing out the allocation of competences between the central authority and the component entities in each legal order is more nuanced and complex than would appear on the face of the respective ‘constitutional’ texts alone. Importantly, although the EU and the US have different starting points when it comes to the vertical division of competences in international relations, both legal orders evidence a subtle shift or accommodation in the opposite direction. In the EU, the Member States in principle occupy a primary role in external relations, as opposed to the Union. In view of the developments heralded by the Lisbon Treaty and relevant case law of the ECJ, however, there appears to be some movement towards the Union level, with respect to the accretion of the Union’s external relations competences 190 See, eg, Schaefer, above n 109, 299–300; JK Schrag, ‘A Federal Framework for Regulating the Growing International Presence of the Several States’ (2009) 48 Washburn Law Journal 425, 437.

166 Geert De Baere and Kathleen Gutman and the emphasis on shared competences with a role for both the Union and the Member States to play. In the US, the federal government in principle occupies the primary role in foreign relations, as opposed to the states. Even so, the states’ increasing engagement with matters concerning foreign relations combined with the approach taken by the US Supreme Court in some recent case law have weakened assumptions of a blanket federal monopoly over all matters implicating foreign relations. Accordingly, in the EU and the US, the point of departure in this setting is no longer solely the exclusive competence of either level, but instead, issues bearing on the relationship between the central and constituent state levels have to be taken into account. In this way, when it comes to federalism and international relations in the EU and the US, now more than ever, this subject beckons fruitful comparative reflection, and it is the hope of the present authors that this chapter may lay the groundwork for more to come.

7 European Ties that Bind: Political or Cultural? HELDER DE SCHUTTER*

I. INTRODUCTION

I

N THE PAST 20 years, many political philosophers have come up with normative analyses of the relationship between the political community and nationality. Two opposing schools have emerged out of this debate: liberal nationalism and constitutional patriotism. Whereas liberal nationalism identifies the national culture as the central source of unity for a political community, constitutional patriotism argues that the source of unity should not be pre-political and national or ethnic, but political and constitutional: State unity should be seen as based on shared reference to liberal-democratic principles embodied in the State’s constitution. It is typically argued by both friends and enemies of liberal nationalism that this doctrine is incompatible with democracy and citizenship beyond the domestic State such as exercised by the European Union (EU): because liberal nationalists believe that democratic units can and should only be national units, democracy beyond the nation-State is regarded as impossible and unattractive. Constitutional patriots, in contrast, are able to embrace EU citizenship. The list of features defining what binds Europeans together drawn up by Jürgen Habermas and Jacques Derrida1 in their opinion piece on the grounds of European identity, for example, includes no cultural or national references, but only political principles and practices, such as social justice, secularism and political institutions: the list is a prime example of constitutional patriotism. It would seem, indeed, that support for European citizenship cannot be sustained by reference to non-political sources of

* Assistant Professor of Social and Political Philosophy at the Institute of Philosophy, Katholieke Universiteit Leuven. I owe many thanks to Lynn Dobson, Florence Delmotte, Andreas Føllesdal and Raf Geenens for the valuable comments and questions I received from them on an earlier draft of this paper. 1 J Habermas and J Derrida, ‘February 15, or What Binds Europeans Together: A Plea for a Common Foreign Policy, Beginning in the Core of Europe’ (2003) 10 Constellations 291.

170 Helder De Schutter unity such as shared national-cultural features, which remain the privilege of the Member States of the EU, not of the EU itself.2 In this chapter, I argue that a case can be made, however, for working out a way of understanding unity within the EU which appeals to more than purely political principles. I argue that the arguments brought forward to argue against liberal nationalism are often sound, but not sufficient to urge us to switch to a constitutional patriotic understanding of unity. Instead, it is possible to envision a pluralistic understanding of national culture, for both domestic and supranational spheres. Our understanding of European unity can then include features only liberal nationalists would typically want to talk about in domestic settings, such as shared cultural identity features, and we can wholeheartedly engage in European nation-building in order to create the solidarity and cohesion needed for social justice to be successful, for democracy to be deliberative, and for trust in institutions to emerge and be sustained. In short: this is an argument for making a liberal nationalist case for European democracy, and runs counter both to European constitutional patriots and to many liberal nationalists who declare supranational citizenship undesirable. My argument may be summarised as follows: (a) (b) (c) (d)

Orthodox liberal nationalism is incompatible with European citizenship. Constitutional patriotism argues that liberal nationalism is exclusionary. But constitutional patriotism itself is also exclusionary. Although liberal nationalism’s liberalism partially deflects the exclusion critique, the exclusion critique still bites. (e) If we pluralise the conception of national identity at the heart of liberal nationalism, thereby transforming it into national pluralism, we can successfully deflect the exclusion critique in a way which keeps it distinct from the constitutional patriotism alternative. (f) National pluralism is no longer incompatible with European citizenship. (g) National pluralism can successfully ground the pursuit of European ties that bind.

2 Indeed, if you are a normative political theorist favourably inclined towards the EU, you are extremely likely to be also a constitutional patriot. This joint intellectual commitment (pro EU and pro constitutional patriotism) has recently been advocated by European constitutional patriots like Andreas Føllesdal, Jürgen Habermas, Justine Lacroix, Jan-Werner Müller and Patricia Nanz. See A Føllesdal, ‘The Sort of Nationalism and Patriotism that Europe Needs’ in H Syse and G Reichberg (eds), Ethics, Nationalism, and Just War: Medieval and Contemporary Perspectives (Washington, DC, Catholic University of America Press, 2007) 267–89; Habermas and Derrida, above n 1; J Lacroix, ‘Does Europe Need Common Values? Habermas vs Habermas’ (2009) 8 European Journal of Political Theory 141; JW Müller, Constitutional Patriotism (Princeton, NJ, Princeton University Press, 2007); P Nanz, Europolis. Constitutional Patriotism beyond the Nation-State (Manchester, Manchester University Press, 2006).

European Ties that Bind: Political or Cultural? 171 II. DEFINING THE TERMS

In line with a practice shared in the literature, I shall understand the term ‘nation’ in a sociological sense as referring to a group sharing collective identity features such as a common language, history, set of symbols and other cultural characteristics, whose members mutually recognise each other as belonging to it, although there may be more shared national characteristics and a nation need not share all of these characteristics. In this sense, a nation is roughly synonymous to a ‘culture’ or a ‘people’. This accords with how most scholars in this field understand ‘nations’.3 It differs, however, from some conceptions which impute nationhood only to States (and that consider ‘nations’ and ‘States’ as synonyms). So in my use of the term, one State may contain more than one nation: Quebec and Flanders may be understood as nations, different from the States (Canada and Belgium) to which they belong, although nothing prevents States such as Canada and Belgium from also providing a certain shared cultural context, as a result of which they too might be called nations. The Flemish citizens of Belgium might thus be part of both the Flemish and the Belgian nations. What are constitutional patriotism and liberal nationalism? Both are theories of cohesion: they are models to conceive unity in a political community. Constitutional patriotism is the view that the central source of unity for a contemporary political community should be located in shared allegiance to the political principles on which State authority and policies are based and which are embodied in a constitution. Liberal nationalism, in contrast, stipulates that the unity of the political community should be grounded in a shared national culture, as nationality not only realises people’s identity interest, but also provides the trust, stability and willingness to reciprocate offers to the community which are highly needed in liberal democratic political communities. Let us look at both in more detail, beginning with liberal nationalism. In general, liberal nationalism’s main feature is that ‘it fosters national ideals without losing sight of other human values against which national ideals ought to be weighed’.4 Thus understood, it is primarily a very moderate nationalism, a nationalism stripped of the illiberal tendencies it has acquired in the course of its history. More specifically, liberal nationalism has, as its name suggests, a liberal and a nationalist component. Its basic aim is to show that both components may be reconciled, that there is a version of nationalism which is not inherently illiberal and that liberalism does not have to be inherently anti-nationalist. The liberal characteristic of liberal nationalism comprises the familiar liberal principles. These impose constraints on nationalist aims and policies. In 3 W Kymlicka, Multicultural Citizenship: A Liberal Theory of Minority Rights (Oxford, Oxford University Press, 1996) 18; D Miller, On Nationality (Oxford, Oxford University Press, 1995) 18–19. 4 Y Tamir, Liberal Nationalism (Princeton, NJ, Princeton University Press 1995) 79.

172 Helder De Schutter particular, they stipulate that individual rights should not be overridden by group rights, and that nationalism has to be ‘moderate’. Liberal nationalist policies are not aggressive against other national identities and typically expose a ‘thin’ or open conception of national identity. This thin view ensures that in principle everyone can join the nation. Liberal nationalism is thus very different from racebased and ethnic or descent-based forms of nationalism. It is essentially cultural and linguistic in nature. The essence of the nationalism of liberal nationalism is the belief that national identities matter, that the State can and should protect and promote them, and, most importantly, that it is desirable that the boundaries of political units coincide with national (and linguistic) units.5 The particular marriage between liberalism and nationalism goes deeper, however, than a mere reconciliation of both ideals. Liberal nationalism also holds that its nationalism follows directly from its liberalism: there are liberal reasons for making nations and political units (ultimately: States) coincide. These reasons all have to do with the furthering of important goods which are both liberal democratic in nature and which can best be provided in national units. The three most crucial goods are identity, social justice and deliberative democracy. The first is the good of national identity. Liberal nationalists believe that national identity is valuable and that it should be politically protected. Different accounts are given for why national identity matters. On the very influential account of Will Kymlicka, for instance, national identity matters because it is a precondition of the liberal ideal of autonomy. Kymlicka argues that, to be autonomous beings, individuals need a national/cultural/linguistic background (which he terms a ‘societal culture’) that makes available options to choose from: ‘Put simply, freedom involves making choices amongst various options, and our societal culture not only provides these options, but also makes them meaningful to us.’6 Our nation thus offers us a ‘context of choice’. Note that this first reason for making States and nations coincide—because it is likely to satisfy people’s interest in identity—does not necessarily lead to secession into fully independent separate nation-States. Although some liberal nationalists do believe full self-determination (and secession) to be the best possible outcome, both those who do and those who do not explicitly acknowledge the legitimacy of multinational States and adhere to a federal solution for realising multinational justice in specific contexts.7 More specifically, they often endorse federalism because they consider it to be a mechanism for recognising claims to 5 Versions of this normative recommendation to (re)draw political boundaries around national units are defended by Chaim Gans, Will Kymlicka, David Miller, Margaret Moore and Yael Tamir. See C Gans, The Limits of Nationalism (Cambridge, Cambridge University Press, 2003); Kymlicka, above n 3; Miller, above n 3; M Moore, ‘Normative Justification for Liberal Nationalism: Justice, Democracy and National Identity’ (2001) 7 Nations and Nationalism 1; Tamir, above n 4. 6 Kymlicka, above n 3, 83. 7 For the first view, see Miller, above n 3, 82. For the second view, see Kymlicka, above n 3; Tamir, above n 4.

European Ties that Bind: Political or Cultural? 173 self-determination. Multinational federalism is then supported as a way of giving national groups the ability to exercise territorial self-determination.8 The second good is that of social (or distributive) justice.9 Redistributive policies presuppose that people are willing to make sacrifices for ‘anonymous others whom we do not know, will probably never meet, and whose ethnic descent, religion and way of life differs from our own’.10 This kind of sacrifice requires a high level of trust—trust that sacrifices will be reciprocated. The principle of nationality (which sets out to draw political boundaries around national boundaries) and (liberal) nation-building, according to liberal nationalists, can create such a common bond between individuals which will increase their willingness to make those sacrifices. The third argument is that shared nationality is conducive to deliberative democracy. This is the case not only because deliberative forms of democracy presuppose that people can understand each other and a shared nationality (including a shared language) facilitates mutual understanding, but also because the success of democracy hinges to a large extent on the belief (or the trust) that others are genuinely willing to consider your own views, and that they will in the future also be willing to moderate their claims in order to reach common ground or consensus if you do so now.11 Today, liberal nationalism’s main contender is constitutional patriotism. Constitutional patriotism ‘designates the idea that political attachment ought to center on the norms, the values and, more indirectly, the procedures of a liberal democratic constitution’.12 It is typically argued for (and also emerged historically) as an alternative to nationalism, which explicitly organises the political community around a shared national culture. Political communities characterised by diversity of a multinational or a multicultural nature, however, do not have a homogeneously shared national culture. In such cases, constitutional patriots insist that we should move the central source of unity to something all the citizens may be expected to share. This new form of attachment would then be political, and not pre-political, as national identities are. The idea, then, is to attempt to create patriotism around the State’s central political principles embodied in a constitution. The idea has especially been designed and invoked for understanding unity within multinational States like Canada, Spain or Belgium, for multicultural

8 See W Kymlicka, Politics in the Vernacular: Nationalism, Multiculturalism, and Citizenship (Oxford, Oxford University Press, 2001) 91–119; W Kymlicka, ‘Territorial Boundaries: A Liberal Egalitarian Perspective’ in D Miller and SH Hashmi (eds), Boundaries and Justice: Diverse Ethical Perspectives (Princeton, NJ, Princeton University Press, 2001) 249–75. 9 See Miller, above n 3, 83–85 and 98; Kymlicka, Politics in the Vernacular, above n 8, 225–26; Tamir, above n 4, 117–21. 10 Kymlicka, ‘Territorial Boundaries’, above n 8, 225. 11 Kymlicka, Politics in the Vernacular, above n 8, 226–27; Kymlicka, ‘Territorial Boundaries’, above n 8, 266; Miller, above n 3, 96–98. 12 Müller, above n 2, 1.

174 Helder De Schutter societies, for societies emerging from pervasive injustice or civil war, as well as for supranational political entities like the EU.13 Not just any constitution will do for the defenders of constitutional patriotism. They have a constitution in mind the content of which is formed by (a contextualisation of) liberal-democratic principles. As such, they do not fundamentally differ from liberal nationalism, which is equally grounded in liberal democracy. So there is no relevant difference in terms of the basic values guiding the constitution on both models. The relevant difference rather lies in the sources of unity thought appropriate for the State: national culture versus the (contextualised) liberal-democratic principles themselves. In the past 20 or so years, these two models have emerged as the central poles of the discussion over how to conceive unity for liberal-democratic States. A large part of the debate has focused so far on the objections formulated to the constitutional patriotic paradigm by liberal nationalists. The latter suspect that constitutional patriotism is too thin and abstract to generate unity,14 that it cannot ‘explain why the boundaries of the political community should fall here rather than there’15 and that Habermas’s correlated idea of separating the country’s political culture from the majority culture is wishful thinking, because it is simply not possible to have culturally neutral policies.16 In response, constitutional patriots have tried to respond to these charges, stressing that individuals without pre-political ties who co-operate in shared institutions can develop enough unity and solidarity, that constitutional patriotism is not a theory of drawing boundaries and that a constitutional political culture will still retain a distinctively national character.17 While several constitutional patriots have thus tried to reformulate the theory such that it can more robustly withstand the liberal nationalist objections, far less energy has been invested by liberal nationalists to develop a response to the central constitutional patriotic charges. So far the onus of proof has rested with the constitutional patriots, not with the liberal nationalists. This is understandable in so far that constitutional patriotism was specifically developed by Habermas as an alternative to nationalism. Yet this silence on the part of liberal nationalists is regrettable, as it may appear that there is no satisfactory answer to the central constitutional patriotic challenge. In what follows, I shall focus on the response the liberal nationalist paradigm can give to the constitutional patriot’s central exclusion critique. In doing so, I shall argue that orthodox liberal nationalism is not able successfully to withstand 13 JW Müller and KL Scheppele, ‘Constitutional Patriotism: An Introduction’ (2008) 6 International Journal of Constitutional Law 67. 14 Kymlicka, above n 3, 200; Miller, above n 3, 162–65. 15 Miller, above n 3, 163. 16 J Habermas, Die postnationale Konstellation (Frankfurt am Main, Suhrkamp, 1998) 114; Kymlicka, above n 3, 111. 17 C Cronin, ‘Democracy and Collective Identity: In Defence of Constitutional Patriotism’ (2003) 11 European Journal of Philosophy 1; Lacroix, above n 2; Müller, above n 2.

European Ties that Bind: Political or Cultural? 175 this critique. To make liberal nationalism’s main premises stronger in deflecting it, I shall reformulate the conception of nationhood at the heart of the liberal nationalist paradigm, so that theorists sympathetic to protection of national cultures may legitimately continue to defend their views.

III. THE EXCLUSION CRITIQUE: WORKING BOTH WAYS

What is the constitutional patriotic critique of liberal nationalism? Constitutional patriots argue that the nationalist paradigm can no longer legitimately work in today’s world and that still relying on it is dangerous. For instance, having identified liberal nationalism as the main domestic rival to constitutional patriotism, Jan-Werner Müller says that ‘[u]nlike most of its philosophical competitors the theory of constitutional patriotism tries to come to terms with the complex nature of contemporary societies’.18 Ciaran Cronin states that nationalist interpretations of citizenship ‘imply a systematic privileging of one’s own form of life simply because it is one’s own’; and whereas nationalism has led to ‘gross injustices against groups stigmatised as aliens or enemies’, constitutional patriotism respects ‘the equal right to coexist of all cultural forms of life that are not predicated on the destruction or denigration of others’.19 This is also clear in Habermas’s understanding of the basic rationale for constitutional patriotism. Today we live, says Habermas, in ‘pluralistic societies that are moving further and further away from the model of a nation-state based on a culturally homogeneous population’.20 In such a society not shared nationality but the democratic process itself should create social integration or unity: In a society characterized by cultural and religious pluralism, this task [creating social integration] cannot be displaced from the level of political will-formation and public communication onto the seemingly natural substrate of a supposedly homogeneous nation. The latter would merely serve as a facade for a hegemonic majority culture.21

In these critiques, the problem with liberal nationalism is characterised as follows. Today societies are no longer united over a shared national culture. In a multicultural, multinational and supranational age,22 seeking unity on the shared platform of a national culture is tantamount to oppression or exclusion of those who do not share the culture, thereby subjecting the latter to the status of secondclass citizens. This is unjust: in a liberal democratic community, all citizens enjoy 18

Müller, above n 2, 141. Cronin, above n 17, 20–21. 20 J Habermas, Die Einbeziehung des Anderen: Studien zur politischen Theorie (Frankfurt am Main, Suhrkamp, 1999) 117. 21 Ibid. 22 I understand ‘multicultural States’ as States with a significant number of recent immigrants or descendants of recent immigrants. I understand ‘multinational States’ as States where more than one national identity is present, such as Belgium, Canada and Spain. ‘Supranationalism’ refers to the fact that States increasingly take part in supranational decision-making bodies, such as the EU. 19

176 Helder De Schutter equal legal status; no citizen should be devalued just because he or she does not share the same national culture as the majority. In short, the objection states that, by grounding politics in specific nations, liberal nationalism excludes minorities who do not share the national or cultural identity that is endorsed by the political regime. The constitutional patriotic alternative, then, seeks to avoid relegating a significant number of citizens to second-class status by transferring the source of unity. In this conception, the members of the political community, irrespective of whether they adhere to the majority nation and speak the majority language or not, come together as free and equal citizens and constitute the nation (or the State) themselves. In working out the conception of justice, we take the interests of each citizen equally into account. How can liberal nationalists react to this objection? As I mentioned above, few liberal nationalists have explicitly sought to guard nationalism against the constitutional patriotic challenge (in part because constitutional patriotism was developed as an alternative to nationalism, not vice versa). But liberal nationalists are of course liberal, and they have been at pains to come up with a nationalism purified of illiberal tendencies. This rejection of illiberal nationalist practices already gives some ammunition to liberal nationalists to deflect the constitutional patriotic challenge. Already as liberals, liberal nationalists are able partially to respond to the challenge, in the following ways. First, liberals liberal nationalists give absolute priority to individual rights and argue that these should always take priority over collective rights. What they defend is not group rights but group-differentiated rights: groups as such have no rights, but individuals do have rights and should be granted the right to live within their own self-governing national groups.23 Secondly, liberal nationalists also believe that nationalism should be liberal in the sense of ‘moderate’: liberal nationalistic politics are not aggressive against other national identities and typically expose a ‘thin’ or open conception of national identity. This thin conception ensures that in principle everyone can join the nation. Liberal nationalism is in this sense very different from race-based and ethnic or decent-based forms of nationalism. It is essentially cultural and linguistic in nature. Thirdly, liberal nationalists (typically) endorse minority claims to selfgovernment. In majority–minority disputes, they side with the national minorities’ demands for self-government rights. Thus, instead of rejecting, liberal nationalists like Chaim Gans, Will Kymlicka and Yael Tamir defend self-government rights for national minorities such as the Catalans, the Quebecois, the Flemish, the Kurds and the Puerto Ricans. In this way they strongly oppose a ‘nationalistic’ rejection of the claims of national minorities. This concern about the rights of minorities is also reflected in the fact that liberal nationalists are often also multiculturalists

23

Kymlicka, above n 3, 45–48.

European Ties that Bind: Political or Cultural? 177 who are sympathetic to minority rights demands of non-national minorities like immigrant groups.24 The fourth reason why liberal nationalism’s liberalism can partly withstand the oppression and exclusion claim is that the liberal nationalist argument against national neutrality does not (necessarily) amount to a criticism of the liberal ideal of remaining neutral with regard to conceptions of the good life. Contemporary nations are typically pluralistic, and liberal nationalists do not wish politically to secure and protect the current values, character or content of a particular national community. What they want is that governments pay attention to national membership and allow every national minority to maintain its own distinct political context. In doing so they are committed to remaining neutral to conceptions of the good life.25 So these responses, which in large part work as liberal-democratic filters for claims made on the basis of nationhood, do constitute important buffers against the charge of oppression and exclusion. If ‘[n]ationalist conceptions of patriotism are incompatible with individual autonomy insofar as they authorise the state to make categorical demands on its subjects without regard to their individual needs and interests’,26 then liberal nationalism argues that it is precisely individual needs and interests (such as autonomy) which form the basis for protecting citizens’ national identity. And if ‘nationalism is beset by an inherent ambivalence that fosters discrimination in the domestic sphere and chauvinistic self-assertion in international relations’,27 then liberal nationalists reply that discrimination is ruled out by neutrality over the good, by the rejection of rights pertaining to groups as such and by the principle that minority nations are equally protected, and that international aggression is ruled out because liberal nationalistic policies respect other national identities (domestic or abroad). Nonetheless, I do not believe that these liberal answers are satisfactory. It is true that a large part of the constitutional patriotic challenge may be rebutted by most of the liberal nationalist accounts we have today. But they thereby neglect a crucial issue: the fact that a one-to-one correspondence between national cultures and political communities is often not sustainable in today’s world. The main problem with liberal nationalism as it stands is that it both empirically and normatively relies on a misguided picture of national affiliation. If we were to live in a mosaic world, where stable boundaries marked off sharply delineated and transparent cultural and linguistic surfaces, and where individuals who lived on such surfaces were all monolingual, monocultural and mononational, then liberal nationalism might have a point. But we do not live in such a world. Our world is not a

24 This holds primarily for Will Kymlicka, Yael Tamir and Joseph Raz. See Kymlicka, above n 3; Tamir, above n 4; J Raz, ‘Multiculturalism: A Liberal Perspective’ (1994) 41 Dissent 67. 25 David Miller appears to be an exception here. He seems to believe that national identities can hold a shared conception of the good life. See Miller, above n 3, ch 2. 26 Cronin, above n 17, 10. 27 Ibid 2.

178 Helder De Schutter mosaic of territorially distinct ‘national monisms’. It is full of heterogeneous and culturally intermixed situations where there is no clear-cut distinction between territorially-concentrated cultural blocs. Individuals in this world are very often not monocultural (or monolingual): their identity may be derived from nations larger (or smaller) than linguistically and territorially distinct nations (such as from the Canadian nation instead of from the Quebec nation, or from the European instead of the German nation); they may be members of more than one national community at once (such as of the Canadian and of the Quebec nation simultaneously); and different individuals of the same ‘nation’ may be internally divided with regard to which context is the most relevant one (42.2 per cent of Flemish respondents in a 1999 survey answered that they self-identify as much with Flanders as with Belgium, 22.4 per cent self-identified more as Flemish than as Belgian;28 surprisingly similar figures are found in Spain and Canada).29 These figures and examples show that the problem is not simply that there is a tendency for liberal nationalism to equate States with nations in their descriptive analysis of the world. It is quite possible to realise that there are many nations without States, who find themselves in multinational States, and to modify liberal nationalism accordingly. But doing that is only part of the job required to come to terms with the fact that we do not live in national monisms, for the people living in such multinational States cannot be adequately characterised as individuals with one national identity who happen to find themselves in a larger State wherein other nations are also embedded. If individuals’ identities may be derived from the State-wide identity, if individuals may be members of more than one national identity within such multinational States, and if groups may be internally divided with regard to which national identity context is the most relevant one, in short if groups have nested and conflicting identities, then we cannot just replace the equation of State and nation with one between the sub-State political community (Flanders, Quebec, Catalonia) and the nation. Thus, while the nationalistic case might have some value for domestic nationStates like Iceland or Hungary, which have remained virtually mononational, and even if it may have value for such multinational States where the sub-State nation is composed of individuals who by and large all adhere only to the sub-State national identity and not to the State-wide identity, it cannot provide a solution for the majority of the world’s political communities. What this means is that the nationalistic solution which upholds a one-political-community/one-nation ideal, whenever instantiated, will give rise to an injustice that does indeed invoke the old liberal concern about minorities. Given the fact that we are confronted not only with a plurality of conceptions of the good life (a fact most liberal

28 B Maddens, J Billiet and R Beerten, ‘De (sub)nationale identiteit en de houding tegenover vreemdelingen in Vlaanderen en Wallonië’ in L Vos and K Deprez (eds), Nationalisme in België. Identiteiten in Beweging. 1780–2000 (Antwerp, Houtekiet, 1999). 29 L Moreno, The Federalization of Spain (London, Frank Cass, 2001) 112–26; M Mendelsohn, ‘Measuring National Identity and Patterns of Attachment: Quebec and Nationalist Mobilization’ (2002) 8 Nationalism and Ethnic Politics 72.

European Ties that Bind: Political or Cultural? 179 nationalists appropriately acknowledge) but also with a plurality of conceptions of national membership, picking out one nation and endowing it with (near-) Statehood unduly excludes minority conceptions of national membership. So, while much of the constitutional patriotic critique has been fended off, it does bite in the end. I shall set out to develop a conception of nationhood that tries to overcome the exclusion problem and is able to respond to the critique in a more fundamental way. But before doing so, it is important to remark that there are reasons to think that constitutional patriotism itself may be liable to the same exclusion charge it levels at liberal nationalism. This may happen in several ways. First, creating patriotism around the State-wide public culture in a multinational State may often unavoidably play into the hands of the majority nation. In States like Canada and the United Kingdom, for instance, the State-wide public culture will often be infused with the majority culture, and even if the State could succeed in uncoupling the political culture from the majority culture, the choice to instil patriotism around that political culture instead of sub-State political units (such as Quebec and Scotland) may imply a preference for the majority’s point of view over that of minorities. Secondly, to the extent that constitutional patriots realise (like Habermas does, for instance)30 that the political culture cannot be entirely uncoupled from cultural influences, that universal liberal-democratic principles will be ‘nationalised’ and will bear the marks of their instantiation in time and place, that automatically implies the possibility that those who do not share that cultural reference scheme might feel excluded. Some constitutional patriots even cultivate this unavoidable permeation of culture and politics. Brian Barry, for instance, argues that it is an appropriate objective of public policy in a liberal democratic state to facilitate the achievement of a state of affairs in which all immigrants—or at least their descendants— become assimilated to the national identity of the country in which they have settled.31

And he argues that in many cases we ought to teach children of minority nations (such as Wales) only the majority language at school. And lastly, the political principles themselves which are to provide the source of unity may be divisive. Even if policies guided by the ideal of constitutional patriotism succeeded in deriving unity purely from shared allegiance to political and constitutional principles, these principles themselves might equally be exclusive. For instance, Justine Lacroix, a constitutional patriot, criticises the list drawn up by Habermas and Derrida for still being too exclusive: One could here recall some of the criteria defined by Habermas—‘the social privatization of faith’, ‘an ethos of struggle for more social justice’, etc—as potential justifications that may be used to deny any European identity to some peoples.32

30

Habermas, above n 16, 215–20. B Barry, Culture and Equality: An Egalitarian Critique of Multiculturalism (Cambridge, Polity Press, 2001) 72. 32 Lacroix, above n 2, 150. 31

180 Helder De Schutter Indeed, there is no reason to think that shared allegiance to political principles is not exclusive. So it appears that both liberal nationalism and constitutional patriotism are incompatible with the central motivation behind the constitutional patriotic model: both exclude. In response to the constitutional patriotic problem of not living up to its own premises, several authors have tried to reformulate constitutional patriotism so as to make it more compatible with its central motivation.33 Anna Stilz, for example, undoes constitutional patriotism from naive understandings about uncoupling the political culture from the majority culture. She agrees with this uncoupling ambition, but argues that so far constitutional patriots have not come up with a successful account of doing so. Her project is to come up with an internal reworking of constitutional patriotism so as to carry out the uncoupling ambition in a more robust way. In the remainder of this chapter, I do for liberal nationalism what Stilz does for constitutional patriotism. My aim is not to criticise or improve constitutional patriotism, but rather to show that there is an alternative based on nationalist premises which does respond to the exclusion critique articulated by constitutional patriotism and yet is distinct from the constitutional patriotic conclusions. The remainder of this contribution is dedicated to such a reworking of liberal nationalism.

IV. NATIONAL PLURALISM: DOMESTIC AND EUROPEAN

What follows is an attempt to respond to the constitutional patriotic challenge by pluralising liberal nationalism. I call the resulting theory ‘national pluralism’. The core idea of national pluralism is that of culturally permeated policies which attempt to be inclusive and give some form of equal consideration to the different cultural frames of reference present within the polity. The national pluralism view accepts the fact, emphasised by all liberal nationalists, that liberal democratic States cannot remain neutral with regard to national membership. In the way a State draws its internal boundaries, in the language policies pursued by the State, in the choice of educational curriculum for State schools, political communities unavoidably engage in national and cultural policies: no polity can remain culturally neutral. In the words of Will Kymlicka, [i]t is quite possible for a state not to have an established church. But the state cannot help but give at least partial establishment to a culture when it decides which language is to be used in public schooling, or in the provision of state services. The state can (and should) replace religious oaths in courts with secular oaths, but it cannot replace the use of English in courts with no language.34 33 See, eg Lacroix, above n 2; A Stilz, ‘Civic Nationalism and Language Policy’ (2009) 37 Philosophy and Public Affairs 257. 34 Kymlicka, above n 3, 111.

European Ties that Bind: Political or Cultural? 181 National pluralism also shares with liberal nationalists the belief that national contexts deserve political recognition, and that this is so for reasons related to democracy, trust and social justice, and, most crucially, because national identities are important goods for individuals who derive value from living in a context in which their national culture is recognised. At the same time, national pluralism shares with constitutional patriotism the belief that ways of conceptualising liberal democracies that avoid privileging certain national memberships over others are very desirable. National pluralists believe this is possible by giving equal recognition to different national communities.35 Let me explain this. The nationalistic alternative to the idea of remaining politically neutral toward cultures consists of granting each nation its own political and territorial sphere. In that sense we may understand liberal nationalism as a form of national monism: every polity recognises only one national culture. But this may be too quick. It is one thing to say that a strict separation between political and national communities is unworkable, but quite another to conclude that the relation between them has to be one-to-one. Liberal nationalists suggest that liberal nationalism is the only credible alternative to the mistaken belief in cultural neutrality. But they thereby oversimplify the issue, since we can think of forms of policy in which a political community’s cultural support is not limited to only one nation, in which a form of national pluralism is pursued. The State cannot avoid taking culturally-permeated decisions regarding the language of public schooling and public services, but it can allow for bilingual schools or different schools operating in different languages within one and the same political community. In short, we cannot ‘replace the use of English in courts with no language’, but we can recognise French alongside English in courts. So it is possible to grant rights to national recognition while at the same time respecting culturally heterogeneous and hybrid situations, because granting rights to a separate territorially and politically autonomous nation is not the only means to protect national choice contexts. There are at least two other ways in which this is possible. The first is perhaps most suitable for constellations in which the nationalist picture is inapplicable because a given territory contains more than one national community that is not territorially separate. In cities and regions like the Basque Country, Brussels, Catalonia, Kosovo and Romanian regions with a majority of citizens with a Hungarian nationality, there is no strict coincidence of territory and nation. In such cases we might consider recognising and supporting not just one but two (or more) national choice contexts, by giving language rights or other forms of support to both nations. In such cases we might additionally consider supporting and recognising forms of overlap between national choice contexts by 35 For similar understandings of equal recognition or treatment, see JH Carens, Culture, Citizenship, and Community: A Contextual Exploration of Justice as Evenhandedness (Oxford, Oxford University Press, 2000), on even-handedness; and A Patten, ‘Liberal Neutrality and Language Policy’ (2003) 31 Philosophy and Public Affairs 356, on liberal neutrality.

182 Helder De Schutter endorsing a binational regime. Such a policy would seek to stimulate participation across the boundaries of strictly monoculturally understood nations while nonetheless recognising national particularities. It does not aim for a dubious separation between the private and the culturally neutral public, but upholds a pluralism in which two or more languages and cultures have a place. Bilingual and bicultural schools, media, police services, hospitals, political parties and clubs are paradigm examples of how participation may go hand in hand with cultural recognition. No doubt this policy will often be highly difficult if not impossible to implement, but it seems to be the only possible just solution, and it avoids many of the serious and violent conflicts over territories that would accompany any solution based on the nationalistic model. A second form in which different nations may be supported on one territorial unit is by implementing forms of federalism. In constellations where many members of a particular nation, say Flanders, experience dual identities—that is, they derive their identity both from a sub-State nation (Flanders) and from the Statewide nation (Belgium)—a federal political system may be a solution that does justice to both identities and recognises them politically. Federal political systems divide powers between a federal government and two or more subunits, such that each level has sovereign authority over certain issues. Note that the national pluralist grounding of federalism differs from the nationalistic understanding of federalism. It is interesting to see that liberal nationalists have trouble justifying the existence of federal States. Any political theory of multinational federalism must include at least two parts: (a) an argument that shows why the federal State should accommodate national identities by granting them powers of self-government; and (b) an argument that shows why a federal State is preferable to secession. Liberal nationalists are very good at providing an answer to the first question, but in doing so make an argument in favour of the second step, if not impossible, then at least very difficult. At best, federalism for national groups can in their view be only a second-best solution. This is different from the national pluralism model, which can easily accept the existence of dual or multiple identities. Apart from dual identities, what might also motivate us to turn to federal solutions is the fact that even within nations such as Catalonia, Flanders or Quebec, there is a considerable level of divergence with regard to which nationality people refer to as the context of their freedom and identity: whereas some will refer uniquely to Spain as their nationality, others will mention both Spanish and Catalan affiliations, and still others will refer uniquely to Catalonia. In this case, Catalan secession would be unjust for the first two groups, whereas retaining (or re-implementing) a unitary non-federal Spanish State would be unjust for the last two groups. A federal regime that gives fair political expression to both the sub-State and the State-wide nations by granting them forms of self-government would therefore be best.

European Ties that Bind: Political or Cultural? 183 Note that both alternatives to the ‘each political community protects one and only one nation’ view—recognising more than one nation within one territorial unit, and implementing federal structures protecting both the sub-State and the State-wide nations—are still compatible with re-installing a version of national monism at the overarching State level. I shall elaborate on this below in the context of the EU. For now it suffices to see that we can support a State-wide national identity context (one compatible with sub-State nation-building), not just because of the importance of the State-wide national identities of those who have them, but also to furnish the federal State itself with enough national ties to bind. Perhaps it might be objected that the form of recognition that follows from the national pluralism model will be less extensive than what the nationalistic model might have to offer to nations. This is true, since the national pluralism model cannot offer every nation a separate political and territorial unit where only its own national culture is recognised. But this can hardly be an objection since it results from a concern about justice. Whenever there is a diversity of national and linguistic identities, one nation can no longer legitimately claim to be the one and only official nation of a specific territorial and political unit. The alternative that the national pluralism model offers for such constellations is to recognise multiple national cultures within one territory. That answer is superior in terms of justice. And it is superior in terms of the specific type of justice in which liberal nationalism is interested: national identity justice, or justice as far as protecting national identity contexts is concerned. National pluralism, then, while different from orthodox liberal nationalism, is a truer development of liberal nationalism’s core idea. I think national pluralism is normatively superior to both orthodox liberal nationalism and constitutional patriotism. But I also think it offers a better description of what is actually going on in many multinational States. Liberal nationalists describe such States as political communities without shared nationality—they argue that the national-cultural ties are to be found among the sub-State units, such as Quebec or Flanders, not Canada or Belgium. And constitutional patriots would see such a State as not honouring pre-political commitments at all. But to think that there are no national-cultural ties at the multinational level, or that multinational States do not engage in nation-building is misguided. Multinational States engage in nation-building just as much as sub-State nations or independent nation-States: they have official holidays, they honour political heroes, their representatives give political speeches, they tend to filter the news in State-wide terms, they have boundaries which inevitably foster a State-wide identity, they typically embrace their internal linguistic diversity as something to be proud about and to foster, they typically hold the foreign policy competence and thereby engender a feeling of reaching out as a nation to other nations, and so on. And citizens do tend to experience nationality also at the multinational level, as the numbers cited above suggest—in that sense the term ‘multinational State’, suggesting the State consists of several nations but is itself no longer a nation, is a

184 Helder De Schutter misnomer. Rather than there being no shared ties of nationality at the State level, such ties are pursued and citizens experience them in ways which are not fundamentally different from those ties at sub-State levels. National pluralism seems a more apt way of describing nationalism and nation-building in ‘multinational States’ than the two competitive theories do. Thus far I have focused mainly on the domestic dimension of this pluralisation of liberal nationalism.36 National pluralism may also be constructed for a European way of distributing national recognition and reaching unity. In fact there are three ways of thinking about liberal forms of nationhood in conjunction with the EU. The first is rejection or scepticism. Indeed, most liberal nationalists are sceptical about the prospects of European citizenship and the emergence of a European identity. David Miller, for instance, thinks that the necessarily national dimension of democracy and citizenship implies that the prospects for European citizenship are for the time being dim. Along with other nationalists he believes that setting up a European level of decision-making would require a shared European identity and nationality, which they do not see emerging now or in the near future.37 This is then an area where constitutional patriotism decisively differs from liberal nationalism. Miller agrees that in contrast to liberal nationalism, constitutional patriotism would ‘allow an easy extension of political identity to European or some other level, since if one can be patriotic about the constitution of a particular state, one can presumably also be patriotic about a federal constitution such as the EU might one day come to have’.38 The second way is to embrace a European liberal nationalism: this envisions a European nation in the way that liberal nationalists do for domestic political communities. This solution again seeks to re-install a pure one-to-one relationship between the political and the national community without recognising any sub-State national identities, and granting all the national recognition to the one European nation.39 The third way is to extend national pluralism onto the European platform. What would this look like? As in Habermas’s theory, the national pluralism model’s view of supranational forms of liberal democracy is not essentially different from its view of the domestic project. This is not a surprise given that the domestic form of this model already abandons the ‘one nation one political community’ paradigm and affirms nationally heterogeneous political communities. Thus, policies can be endorsed that recognise and grant rights to more than one national identity.

36 A fuller account of domestic national pluralism may be found in H De Schutter, ‘Nations beyond Nationalism’ (2007) 50 Inquiry: An Interdisciplinary Journal of Philosophy 378. 37 Miller, above n 3, 160–63; Kymlicka, ‘Territorial Boundaries’, above n 8, 323–26. 38 Miller, above n 3, 163. 39 This is not the view that sees a European nation based on, for instance, Christian values. I am focusing throughout this chapter on liberal understandings of nationhood in conjunction with the EU. Just as liberal nationalists refuse to talk about religious forms of unity at the domestic level, so would a European liberal nationalism not focus on religious bonds.

European Ties that Bind: Political or Cultural? 185 This can be done in a non-exclusive manner in two conceptually distinct ways: to create overarching ‘national’ unity, and to respect sub-EU national identities. The first is to conceive a European-wide national identity. This involves working out an understanding of what Europeans share and can share in terms of identity and nationality. But this can be successful only in connection with the second: to justify and corroborate an ethos at the EU which supports the idea that the sub-EU national identities of States and sub-State regions matter and deserve respect. This respect policy may itself become a reason for attachment to and a principle of unity of the EU, in a way similar to how, for instance, multiculturalism is built into Canadian identity. It might be pursued in general policy domains, in the domain of language policy and in the protection given to national minorities, but it might also be given institutional expression through federalism which grants large self-determining power to sub-EU national units of decision-making. This way to imagine unity at an EU level is different from a European liberal nationalism (option two above) in that it allows for multiple identity allegiances, and most importantly because it builds a principle of equal recognition of national identities into the very conception of European identity, thereby avoiding exclusion. But it also differs from constitutional patriotism, since it continues to talk about culture and nationality in envisioning unity. It does not seek to detach the sources of unity from cultural and national allegiances, and replace them with political and constitutional principles. Instead it builds on the motivational power of the language of nationhood, and seeks to extend it to a European level. But why would we want to conceive a European-wide identity and some form of national-cultural unity for the European political community? Indeed, in response to the above, many nationalists will point to the fact that such supranational identity still does not exist in a robust form, that there is no real European people, and that it is not clear whether national identities will disappear or transform into supranational identities. However, the national pluralism model suggests neither that national identities are eroding, nor that they no longer deserve political recognition. On the contrary, the national pluralism model explicitly sets out politically to accommodate State-wide and sub-State national-cultural identities besides an EU-wide national-cultural identity. It remains a fact, though, that national identities are currently much stronger than supranational identities. Two things may be said in response to this. The first is that a European identity, even if it is weak, does nonetheless exist. And it exists in a way which resembles national identity positions in multinational States: many citizens have dual identities, others have simply local national identities and still others uniquely conceive of themselves as European citizens. Ten per cent of Europeans self-define as ‘European only’ or ‘European first, national second’, 41 per cent self-identify as ‘national only’, and 46 per cent see themselves as national and European.40 European federalism, 40 See ‘Eurobarometer Spring 2004—Public Opinion in the European Union’ (July 2004). I should remark here that there is a crucial difference between the word used in the poll to refer to the identity shared among Europeans, and the one used in the earlier studies in Belgium, Canada and Spain. In

186 Helder De Schutter which includes a European level of decision-making, may be understood as an appropriate structure to accommodate these European identity layers. But this answer will not suffice, especially because European affiliations are much weaker than federal identities in multinational States. But secondly, it accords with the spirit of the national pluralism model to foster such European identities where they do not (or only weakly) exist, for reasons not related to identity. Let us call this the cultural engineering argument. To entertain it, note first that the national pluralism model argues that national identities should be respected and politically recognised (through forms of nation-building, selfgovernment rights, language rights, etc) if these national identities are valuable to the individuals who make up these nations. None of this implies that voluntary assimilation is undesirable. If people take on new identities in addition to or as a substitute for their original national identity, there is no reason to ‘force’ them to re-assume only their original identity. The cultural engineering argument suggests that it is desirable to foster new identity layers compatible with existing national identities by engaging in forms of supranational nation-building, as a result of which a new category of national belonging is offered to citizens. Voluntary assimilation is legitimate, but nothing suggests that the EU polity should want its citizens to copy the identity behaviour of the 10 per cent ‘European only and European first’ citizens. The desirability of cultural engineering may become clear once we realise that nationality can serve functions other than those related to providing identity, functions which are essential to the goals inherent in the liberal democratic project. As liberal nationalists have argued, nations also have an instrumental dimension. Language, for instance, is not only a source of identity but also a means of communication: language is an instrument for conveying messages. And nations are not only freedom-enabling contexts but also means of fostering social justice and solidarity. Take for instance the social justice argument, which says that nationality can be an instrument for social justice. As liberal nationalists have taught us, social justice requires a certain readiness of individuals to make sacrifices for anonymous others, most of whom they will never meet, and a shared nationality can stimulate that readiness. Weaving a nationality around the community can create a common ‘destiny’, so that people will be inclined to consider each other as ‘one of us’. A shared nationality will also foster the belief that sacrifices will be reciprocated. This entails that certain forms of cultural engineering may be positive. I understand cultural engineering as the policy goal to encourage the development of a the latter, citizens are probed for their national identities and whether these are State-wide or subState (eg, a Belgian national identity, a Flemish national identity or both?). The Eurobarometer study contrasts on the one hand national identity, which is equated with the domestic State identity, with on the other hand EU identity, which is thus not described as EU ‘national’ identity. This confusion over the word ‘national’ (taken in the EU-study as State-based only, not as sub-State nor as EU-wide) is based on a general equation of ‘nation’ and ‘State’, which I here argue is ill-founded: descriptively and normatively, sub-State identities and the EU identity resemble State identities.

European Ties that Bind: Political or Cultural? 187 certain national identity layer for instrumental reasons, defined as non-identity reasons: reasons meant to protect not the identity value of national identity, but rather its instrumental value. Let me give a clear example. Imagine it is found desirable to implement some form of European social justice scheme. If so, then sensibly fostering a European national identity will be beneficial to that project. But does fostering this European identity not clash with existing national identities? I do not think so, and my answer will bring us to the heart of the relationship between liberal democracy and nationality. I shall develop this answer through a critical analysis of a counter-argument Kymlicka has developed against stimulating such a European national identity. Kymlicka criticises the attempts of cosmopolitans like David Held to implement forms of supranational or global citizenship through which individuals can participate directly in international organisations, thereby democratising and legitimising them. This might, for instance, mean that additional powers in the EU are transferred from intergovernmental organs (like the Council of Ministers) to the European Parliament, which is directly elected by citizens. Kymlicka disagrees. His view is that we should not make international institutions directly but only indirectly accountable. The result would be a form of representation with the national representative as a medium between the national people and the supranational decision-making body. In his view, citizens should debate at the national level about how they want their national governments and representatives to act in intergovernmental contexts. This should be the case since, for the liberal nationalist Kymlicka,41 true freedom and autonomy and true democracy can be realised only within national boundaries, such as within the Dutch, Danish, Catalan or Flemish units (and not within the Spanish or the Belgian units). I think Kymlicka’s liberal nationalist conclusion about indirect accountability is unwarranted (while not inconsistent): Kymlicka is a proponent of cultural engineering and nation-building as long as this concerns the uniting of earlier identities into what we now call nations like Catalonia, France or Denmark. He wholeheartedly approves of the historical development of national choice contexts that came into existence through active nation-building efforts (though he of course disagrees with their illiberal methods). His theory also supports present-day liberal forms of nation-building. But Kymlicka is not prepared to pursue nation-building at levels above the nation. He supports the development that transformed peasants from the south of France into French citizens, but he is not willing to support European national identity projects. It is not clear, however, why existing nations should have a privileged position here. If recent processes of globalisation and Europeanisation are creating a new civil society in Europe,42 and also new identity structures, there is no reason within the liberal nationalist preoccupation with the importance of identities to inhibit the emergence of a European supranational choice context. What I am arguing here is that, even if this identity is still very tentative or fragile,

41 42

Kymlicka, Politics in the Vernacular, above n 8, 317–26. Ibid 326.

188 Helder De Schutter there may be reasons unrelated to identity to stimulate the development of such a supranational identity layer. Within the European context one might argue, for instance, that a strong European identity would significantly contribute to the development of a generous redistribution from richer to poorer regions. This line of reasoning appeals to the instrumental argument for cultural State-intervention, which says that a shared nationality is conducive to social justice. What forms of nation-building would be appropriate at the European level? Clearly these should be limited to liberal and moderate forms of nation-building. Examples may include some of the following: granting more powers to the EU Parliament; creating incentives for the emergence of European-wide political parties; implementing a European-wide history course in schools that focuses on European history; creating some symbolic attachment to the European anthem and flag; maintaining, honouring and advertising the equal respect language policy; creating funds for writers or historians for projects about European culture, philosophy and art; subsidising the creation of European-wide public media (such as the existing bilingual French-German Arte channel). At this point two crucial remarks need to be made. First, it is important that cultural engineering will not conflict with the national identity argument. The instrumental argument does not have priority over the identity argument, and it is not part of the cultural engineering argument that a European identity should or would undo or supersede all national identities. It is legitimate if in due time people voluntarily shake off their original identities and uniquely identify with a new identity layer. But that does not mean that those who have not done so are not entitled to political recognition of their identities. Therefore, we should limit the range of instrumental reasons for nation-building to stimulating extra identity layers. Stimulating extra layers is possible because, as explained above, identities are not ‘exclusive’: people are capable of assuming multiple identities simultaneously. Secondly, the argument I have developed here rests on the independent desirability of those instrumental motives. I have presupposed for the purpose of the argument that some level of income redistribution on a European scale is desirable. But this argument is not complete without an independent argument in favour of such instrumental motives (nor without an independent argument in favour of having a European Union in the first place). I have not provided this independent argument here. But that a theory of appropriate nationhood does not in and of itself give a full answer to the question of where the boundaries of the nations should be drawn, is not an objection to such a theory (of which liberal nationalism and national pluralism are examples). Constitutional patriotism, the rival theory, is equally unable to do so: also constitutional patriotism is normatively dependent on such an instrumental argument, which is for example appropriately recognised by Müller.43 My main aim here was just to show that there may also be instrumental motives for cultural regulations, and that the logic of national recognition is not incompatible with such motives.

43

Müller, above n 2, 47–48.

European Ties that Bind: Political or Cultural? 189 V. CONCLUSION

When thinking about conceiving unity for political communities who are not united in terms of nationality, three routes may be imagined. The first gives national recognition to none of the identity groups, and shifts the source of identity to shared political principles. This is the rationale behind constitutional patriotism. The second tries to transfer the heart of democracy and justice to smaller nationally homogeneous political communities, or, where doing so is impossible, gives, as it were, all the national recognition to one of the national cultures present in the political community. That is the liberal nationalist route as it has been typically understood. Another option is not to give recognition to none, nor to give it all to one group, but to give to all groups. That is the pluralistic alternative, the route suggested in this chapter. It wholeheartedly embraces nation-building, basing unity on features constitutional patriots typically shy away from, such as historical traditions, mutual appreciation of old national cultures and linguistic cultures, or a European cultural identity. Yet it does so in ways that are inclusive and open, and respectful of sub-EU national identities. The aim of working out the premises of this alternative understanding of unity is to take away one important source of opposition to the language of nationhood as a means of reaching unity, by making that language more inclusive. If the argument has been successful, perhaps it will become clear that, to the extent that constitutional patriotism succeeds in avoiding the exclusion danger (an ambition on which doubt may be cast, as I have tried to do above), realising its central premises may also be done in the ‘national way’, and there is no reason for switching to constitutional patriotism. It appears then that constitutional patriots attempt to reach a conclusion on the basis of an argument that also allows for other conclusions. The premise is the complex nature of contemporary societies and the need not to exclude or disrespect some by imposing an identity on them which they might not share. If my argument here has been successful then perhaps two conclusions can be reached: building unity around shared political principles and building unity around a pluralistic conception of nationality. Here I have contributed to the second task. Notice lastly that this pluralistic understanding is not just able to perceive unity in an attractive way for the current EU project; it also allows for a possible change (or enlargement) of borders in the political community. Liberal nationalism traditionally conceived stipulates that borders should be drawn around nations as they are now understood (and that is seen as an advantage compared with constitutional patriotism which does not have a boundary-drawing principle built into the theory). Constitutional patriotism, in contrast, may be conceived as open with regard to its constituents.44 Fairness can be guaranteed here if the newly-joining nations share the political principles, or if the political principles are changed in the light of the newcomers. In the national pluralist paradigm, dynamism

44

Ibid 125.

190 Helder De Schutter regarding enlargement may be incorporated in a legitimate way by understanding new members as nations with entitlements to recognition and accommodation on the part of the EU, as well as, crucially, new national cultures to take into account in constructing a renewed inclusive national self-understanding of European identity, a process which would have to be re-initiated whenever new members join. Faced with new members, the EU must seek to reinterpret its shared European identity so that it is reasonable for all nations and citizens to embrace the new, enlarged identity.

8 Does EU Decision-Making Take into Account Regional Interests? PIET VAN NUFFEL*

I. INTRODUCTION

D

ECISION-MAKING WITHIN the European Union (EU) is generally presented as an exercise in finding a politically acceptable equilibrium between the various ‘national’ interests of the Member States and the more abstract ‘European’ interest. Since that presentation is based on the notion of Member States as indivisible entities, it ignores the fact that in various Member States ‘regions’ have a political role to play and hence promote ‘regional interests’. The classic portrayal of EU decision-making as a clash between ‘national’ and ‘European’ interests remains a useful way to explain the co-existence of ‘intergovernmental’ and ‘supranational’ institutions within the EU’s decision-making process. However, it does not correspond to the political reality where various public and private stakeholders attempt to defend their interests by lobbying the actors involved in the EU decision-making process. In that context, regional authorities and other promoters of regional interests also try to be heard in the EU decision-making process. This chapter analyses to what extent the EU’s institutional set-up allows for regional interests to be defended in the supranational decision-making process. This decision-making process is understood in a broad sense. It encompasses not only the EU’s legislative process, which is based on the interaction between the European Commission, the European Parliament and the Council, but also the ‘constitutional’ decision-making process, which occurs through Treaty amendments. This contribution sets off with an attempt to define ‘regional interests’, followed by an inquiry as to whether the Treaties protect the autonomy of Member States’ regional authorities. The analysis of the role of regional interests in EU decision-making leads to a conclusion on the position of regional interests under the Treaties.

* Professor of European Law at the Katholieke Universiteit Leuven; member of the Legal Service of the European Commission. This contribution is the author’s personal opinion and cannot be ascribed to the European Commission.

192 Piet Van Nuffel II. HOW POLITICAL INTERESTS BECOME ‘REGIONAL’ INTERESTS

There is no such thing as pre-existing ‘regional’ interests. Like any ‘national’ interest, a regional interest is always the result of political decisions that construe, out of different visions and concerns, the positions that will be advanced as the interests of the region.1 Regional interests exist where political actors identify specific interests as being those that need to be propagated and protected on behalf of a region’s inhabitants or economic operators. Such political actors may be organised within political institutions at regional level, but this is not necessarily the case. Indeed, politicians with regional ‘roots’ may have incentives to promote regional interests within institutions at national or even at supranational level. In principle, any type of political preferences or concerns may be turned into regional interests. In some situations, regional political actors may defend almost ‘natural’ interests of a region, for instance when taking care of geographic, cultural or economic particulars of the region. Thus, regions are prone to defend the interests of dominant economic groups, such as the wine-growing sector in Bordeaux or Champagne, the shipping industry in Cyprus or Malta, or small and medium-size enterprises in Flanders.2 Regions accommodating small linguistic communities will have a tendency to protect and promote their language and local artistic scene. However, in many more areas, there is no such ‘natural’ regional interest, so that politicians, where they need to defend the region’s position, will defend the interests that correspond with their ideological preferences or, in the case of institutional actors, with the ideological position of the political majority. Whether the interest finally advanced as a regional interest corresponds to a region’s ‘natural’ interest or to the ideological preference of an accidental majority, it will normally be supported by the region’s most influential political groups. That does not exclude, of course, that there may be a minority group within the region that does not esteem the ‘chosen’ interest to be ‘in the interest’ of the region. All this means that the interests of one region may be very different from those of other regions. Nonetheless, regions may have some ‘institutional’ interests in common, namely, the interest to protect and promote regional autonomy vis-à-vis other levels of power. The defence of regional interests may be taken up by individual politicians from the region concerned, or by political institutions, created by or within that region, such as regional governments and regional parliaments. As already said, such regional interests may be advanced either within the circumscription of a region or at a broader political level, such as the national or European decisionmaking process.

1 See P Van Nuffel, ‘What’s in a Member State? Central and Decentralised Authorities before the Community Courts’ (2001) 38 CML Rev 871. 2 See also Case C-518/07 Commission v Germany [2010] ECR I-1885, para 35 (in the application of data protection law, a regional government may favour the interests of companies which are economically important for that region).

Does EU Decision-Making Take into Account Regional Interests? 193 III. REGIONAL INTERESTS AND REGIONAL AUTONOMY IN THE TREATIES

Traditionally, the Community Treaties referred only to the Member States, without taking into account the developments in some Member States toward self-government at regional level. This ‘blindness’ towards regional autonomy (Landesblindheit)3 has, however, largely disappeared. In several instances, the Treaties contain obligations for the Union institutions to take into account the specific situation of the Member States’ regions. First, the Lisbon Treaty introduced a general obligation for the EU to respect the fundamental political and constitutional structures of the Member States as regards ‘regional and local self-government’ as part of their ‘national identity’ (Article 4(2) TEU). A closer look at the Treaties shows, however, that they do not recognise regional autonomy as a claim that any region of a Member State may exercise vis-à-vis the central authorities of its State, but only as a right of the Member States to have their internal state structure organised on the basis of regional autonomy.4 What is protected, therefore, is the freedom of a Member State to opt for a federal or decentralised system, and not the right of any particular region to claim some degree of political autonomy within its Member State. That explains why the Treaties conceive ‘regional autonomy’ as part of a Member State’s ‘national identity’ and do not refer to any ‘regional identity’. Secondly, in various areas, the Treaties require the EU to take account of the specific situation of the Member States’ regions and of the differences that may exist between those regions. For example, EU policy-making is to consider ‘regional heritage’ (Article 13 TFEU), the structural and natural disparities between the various agricultural regions (Article 39(2)(a) TFEU), employment in the various regions (Article 46(d) TFEU), the standard of living and level of employment in certain regions (Article 91(2) TFEU), ‘appropriate regional economic policy’ (Article 96(2) TFEU), the ‘regional diversity’ of culture (Article 167(1) TFEU) and the diversity of environmental situations in the various regions (Articles 191(2) and (3) TFEU). Such requirements to respect regional differences may be inspired by different factors. With respect to economic policies, such as the EU’s agricultural or transport policy, one might argue that the need to have a region’s distinct situation taken into account seems to flow from the concern of ensuring the efficiency of the EU’s action. General measures that do not fit the particularities of specific regions may indeed turn out to be ineffective. A different concern is probably lying behind the requirement to consider the regions’ cultural diversity. Here, 3 See, in the German literature, J Wuermeling, ‘Das Ende der “Länderblindheit”: Der Ausschuß der Regionen nach dem neuen EG-Vertrag’ (1993) 28 Europarecht 196; A Epiney, ‘Gemeinschaftsrecht und Föderalismus: “Landes-Blindheit” und Pflicht zur Berücksichtigung innerstaatlicher Verfassungsstrukturen’ (1994) 29 Europarecht 301. 4 In this respect, it goes without saying that, where EU law leaves some discretion to the Member States, the existence of differences between the legislation of a Member State’s regions does not amount to ‘discrimination’. See Case C-428/07 Horvath [2009] ECR I-6355, paras 47–58.

194 Piet Van Nuffel it would appear more important to protect certain political sensitivities than to ensure economic efficiencies. Respect for the ‘regional diversity’ of culture thus appears to reflect the principle of subsidiarity, according to which the EU will not intervene in areas where the Member States—or their sub-entities—are better placed to take action. Likewise, the aim of leaving politically sensitive issues untouched seems to be behind the requirement of Article 13 TFEU to respect the ‘regional heritage’ as regards Member States’ policies in respect of animals.5 In all those areas, the obligation to take regional differences into account does not amount to a requirement to respect regional autonomy. The abovementioned Treaty provisions require the EU institutions to account for regional differences as a matter of fact. Again, there is no obligation for EU institutions to take into account the interests of a particular region as they would be defined by that region’s politicians or institutions. Thirdly, the Treaties lay down a more general obligation for the EU legislator to take into account the interests of Member State regions. That obligation is contained in the Protocol on the application of the principles of subsidiarity and proportionality (Subsidiarity Protocol), as it has been revised by the Lisbon Treaty.6 The Protocol requires that draft legislative acts be assessed for compliance with the principles of subsidiarity and proportionality. According to Article 5 of the Subsidiarity Protocol, whenever an EU directive is proposed, an assessment is to be made of its implications for the legislation of the Member States, including ‘regional legislation’. Draft legislative acts must also take into account the need to minimise any burden falling upon ‘regional authorities’.7 However, ‘regional authorities’ are mentioned at the same level as any other authority of the Member States. Their interests need to be taken into account no more than those of other concerned public authorities or private parties.8 In substance, the Subsidiarity Protocol does nothing more than to spell out the general requirement which the principle of proportionality imposes anyway: the EU legislator must protect all legitimate interests (whether public or private) against unnecessary EU intervention.9

5 Art 13 TFEU refers to the need ‘to pay full regard to the welfare requirements of animals’ (as ‘sentient beings’), respecting the law and customs of Member States relating to religious rites, cultural traditions and regional heritage. This seems to protect customs such as bull fighting in certain Spanish regions or fox hunting in England. 6 Protocol No 2 on the application of the principles of subsidiarity and proportionality [2010] OJ C83/206. 7 Ibid Art 5. 8 Art 5 of the Subsidiarity Protocol requires any draft legislative act to contain a detailed statement with ‘an assessment of the proposal’s financial impact and, in the case of a directive, of its implications for the rules to be put in place by Member States, including, where necessary, the regional legislation’ (ibid). Draft legislative acts are to take account ‘of the need for any burden, whether financial or administrative, falling upon the Union, national governments, regional or local authorities, economic operators and citizens, to be minimised and commensurate with the objective to be achieved’ (ibid). 9 See P Van Nuffel, De rechtsbescherming van nationale overheden in het Europees recht (Deventer, Kluwer, 2000) 298–303 (claims based on the principle of proportionality seek to safeguard both private and public interests; however, public authorities often invoke the principle to defend the interests of specific private economic operators).

Does EU Decision-Making Take into Account Regional Interests? 195 In this respect, it must be recalled that the proportionality principle does not require the EU to modify its legislation such that it takes into account any possible difference that would exist between Member States or between their regions. Under EU law, the proportionality principle is violated only if ignoring certain particularities would render the legislation concerned unsuitable to achieve its objective, or when its objective could be equally achieved with measures that are less intrusive or that better accommodate these differences.10 Instances may be rare where the specific situation of a region would be such that, without accommodation of that region’s specificities, a piece of legislation would fail to achieve its objectives. Likewise, it may be hard for regional authorities to prove that alternative legislative measures, which take better account of their specific situation (eg, by providing for certain exceptions or derogations), would achieve the proposed objectives to the same extent as measures that do not contain such exceptions or derogations. The foregoing does not preclude that there may be cases where specific features of a particular region cannot be ignored by EU legislation. One may conceive of a region whose economy largely depends on one single activity, or whose population is exceptionally old or young. More often, however, it will be possible to express such particularities in objective economic or social characteristics, without linking them to a specific region. In the same way that the EU legislator cannot take into account the particular situation of any private operator likely to be affected, it is not only practically but also legally impossible for the EU to modulate its legislation so as to take into account the specificity of every single region.11 It follows that there is no guarantee in the Treaties for regions to have a certain degree of autonomy respected vis-à-vis the national or European legislator. Regions may, however, invoke the EU’s obligation to respect their Member State’s federal system.

IV. INFLUENCE OF REGIONAL INTERESTS IN EU DECISION-MAKING

As explained above, ‘regional interests’ arise whenever individual politicians and/ or political institutions of a particular region start identifying and propagating specific private or public interests as the interests of that region. The question is to what extent the EU’s institutional architecture allows for such regional interests to be expressed in the decision-making process at the European level, and whether the Lisbon Treaty brought about any changes in this respect.

10 11

Ibid 339. Ibid 345.

196 Piet Van Nuffel A. Direct and Indirect Defence of Regional Interests Within the EU decision-making process, regional interests may be defended in indirect and direct ways. An indirect way for regions to get their interests recognised and protected at the EU level is to exert influence on the politicians or institutions that defend the ‘national’ interest of the Member State within the European institutions. Like any other interest group within a Member State, regional politicians may attempt to influence the setting of political priorities. Within national parliaments, politicians from a given region may find support from their fellow parliamentarians and succeed in having their views defended as an opinion of the national parliament. An even more promising (indirect) route towards the protection of regional interests at the EU level is to convince a national minister to defend regional interests as the ‘national’ interests of the Member State within the Council or the European Council. Yet, from a legal perspective, there is not much to say about the indirect defence of regional interests. Its chance of success depends on the degree to which the national politicians and institutions allow the views of one region to prevail at that national level. More interesting, however, is the possibility for regional politicians and/or institutions to exert influence at the European level in a direct way, that is to say, not through the channel of ‘national influence’ but directly as regional governments, regional parliaments or, for individual politicians, as representatives of the regions.

B. Institutional Avenues for Direct Influence since the Maastricht Treaty It was not until the 1992 Treaty of Maastricht that regional institutions and politicians were given the possibility of having a direct say in the EU legislative process. From then on, ministers of regional governments, on behalf of their Member States, might participate in the activities of the Council. Moreover, representatives of the regions could be appointed to the Committee of the Regions, whose task it is to give advice to the EU legislator. Surprisingly or not, these ‘institutional avenues’ for the direct protection of regional interests were left largely untouched by the Lisbon Treaty. i. Presence of Regional Ministers in the Council As regards the Council, Article 16(2) TEU states that it ‘shall consist of a representative of each Member State at ministerial level, who may commit the government of the Member State in question and cast its vote’.12 Before the EU Treaty entered

12

See K Lenaerts and P Van Nuffel, European Union Law (London, Sweet & Maxwell, 2011) 489.

Does EU Decision-Making Take into Account Regional Interests? 197 into force, each government had to delegate ‘one of its members’ to the Council,13 which meant that only ministers of the national government were allowed to speak in the Council.14 At the instigation of federal Member States such as Germany and Belgium, the Treaty was altered in 1992 so as to allow members of regional governments to represent their Member State in the Council. The Treaties thus permit a federal Member State to have its internal division of powers reflected at EU level. Rather oddly, Article 16(2) TEU mentions that any vote by a Member State’s representative commits ‘the government of the Member State’. Indeed, in reality, it is not the Member State’s central government which is committed by such vote but the Member State itself. Although a Member State’s regional governments do not always defend the same political position, the Treaties allow only one representative per Member State. Consequently, in areas of regional competence, arrangements have to be made at the national level to select the regional minister who will act as the representative of the Member State. Thus, in the same way as Article 4(2) TEU protects regional self-government as an aspect of a Member State’s national identity (rather than as a right of the Member States’ regions themselves), Article 16(2) TEU merely permits Member States to bring their representation in the Council into line with their internal structure. This Treaty provision does not, however, grant regional governments any right to represent their Member State in the Council. Moreover, since each Member State has only one representative in the Council, internal arrangements have to be made to determine: (a) the conditions under which a member of a regional government will be delegated to the Council; and (b) how that representative is to defend as one view in the Council what may be contentious views within the Member State. In Belgium, for example, the representation of the Member State in the Council is regulated by Article 81(6) of the Special Law of 8 August 1980 on institutional reform. This provision was implemented by a ‘co-operation agreement’ between the federal Government and all relevant regional governments.15 The agreement established a system whereby, depending on whether the Council is to debate on an area of federal or regional competence, Belgium is represented by either a federal or a regional minister (with a rotation system where representation is to be done by a regional minister). In areas of parallel competence, the federal or regional minister in question will be accompanied by a minister of the other

13

Art 2(1), Merger Treaty 1965 [1967] OJ 152. See F Eggermont, ‘In the Name of Democracy: The External Representation of the Regions in the Council’ in C Panara and A De Becker (eds), The Role of the Regions in EU Governance (Heidelberg, Springer, 2011) 4–5. 15 Co-operation Agreement of 8 March 1994 between the Federal State, the Communities, and the Regions. See A De Becker, ‘Belgium: The State and the Sub-State Entities are Equal, but is the State Sometimes Still More Equal than the Others?’ in Panara and De Becker (eds), above n 14, 260–65. 14

198 Piet Van Nuffel level of government, who sits as an ‘assessor’. Inevitably, such a system requires prior consultation amongst all authorities concerned to find a common position. In case of disagreement, the co-operation agreement stipulates that a position ad referendum needs to be taken and that Belgium’s final position must be communicated later. In practice, however, Belgium would, in that event, abstain from voting.16 As regards the Member States’ participation in the activities of the Council, the Treaties refer to the sessions of the Council but hardly mention the preparatory work undertaken before a matter reaches the level of the Ministers.17 In practice, however, the possibility for representatives of the regions to act on behalf of the Member State also applies to the representation of the Member State within the numerous working parties and committees that prepare the Council’s work, as well as within the committees that assist the Commission in implementing EU law.18 Likewise, regional delegates may be integrated in the Member States’ permanent representations to the EU.19 However, the fact that a regional minister is to act for the Member State does not necessarily mean that ‘regional interests’ can be defended in the Council. While the prospect of influencing the position expressed in the Council may be real for a regional representative of a Member State with few regions, such as Belgium (where, depending on the subject matter, three Regions and three Communities have legislative power), it may be less so in Member States with a greater number of regions.20 Where a region needs to negotiate at the national level with various other political players—the federal government and/or other regions—there may, at the end of the day, be not much room to advance ‘regional’ interests as identified by that particular region. Indeed, the internal co-ordination which precedes every Council session will most likely have led to a political compromise amongst all relevant interests. ii. Defence of Regional Interests in the European Parliament and the Committee of the Regions As regards the defence of regional interests outside the Council, one naturally thinks of the Committee of the Regions, which was set up by the Maastricht 16 As regards one subject matter area (ie, agriculture), the ‘solution’ of abstention has even been expressly provided for by the co-operation agreement (explanatory memorandum to the agreement, para 4). 17 See Arts 16(7) TEU and 204(1) TFEU (Committee of Permanent Representatives). 18 See Lenaerts and Van Nuffel, above n 12, 628. 19 Curiously, however, the ECJ found that, for the purposes of determining entitlement to an expatriation allowance in accordance with the second indent of Art 4(1)(a) of Annex VII to the Staff Regulations, ‘work done for another State’ only covers work done for the permanent representation of a Member State and not work done for political subdivisions of that State. See Case C-7/06 P Salvador García [2007] ECR I-10265; Case C-8/06 P Herrero Romeu [2007] ECR I-10333; Case C-9/06 P Salazar Brier [2007] ECR I-10357; Case C-10/06 P de Bustamante [2007] ECR I-10381. 20 See also Eggermont, above n 14, 17.

Does EU Decision-Making Take into Account Regional Interests? 199 Treaty to give the Member States’ regions a direct voice in EU decision making. Nonetheless, it should not be forgotten that regional interests may also be directly advanced through the European Parliament. As in the Committee of the Regions, politicians of a particular region may seek support amongst their fellow parliamentarians for the region’s interests. Especially in those Member States where members of the European Parliament (MEPs) are elected in regional circumscriptions and/or where political parties are organised within regional confines, MEPs inevitably act as regional representatives who may choose to promote the interests of their region. As compared to the European Parliament, the Committee of the Regions has a far more restricted role to play in the EU decision-making process. The Committee has a mere consultative role and its opinion is not legally required except in specified areas of EU competence. Moreover, the Committee is not necessarily better placed than the European Parliament to defend a particular region’s interests. First, although the Nice Treaty introduced the requirement for members of the Committee to be elected to, or at least politically accountable to, an elected assembly, the Committee still comprises not only representatives of regions but also representatives of local bodies.21 Secondly, politicians of a particular region cannot be sure to find support in the Committee for whatever ‘regional interest’ they seek to defend. One region’s interests do not necessarily coincide with those of another region, and the representative of a given regional body does not necessarily have interests in common with the majority of other regional (and local) bodies represented in the Committee. The chances of finding support may be higher where ‘institutional’ interests of regions are at stake or, even more, where the institutional interests of the Committee itself are concerned.22 Nonetheless, the Committee of the Regions may be important for regional politicians in other respects. In practice, the Committee appears not to see its role as confined to giving advice in the decision-making process. It tries to achieve broader objectives, such as the fostering of ‘networks’ between regional and local authorities, both amongst themselves and with the European institutions.23 The ‘networking’ that occurs at the margins of the Committee’s working sessions may, of course, be helpful for regional politicians in their defence of regional interests, be it directly or indirectly, in other forums.

21 Thus, the Belgian representation in the Committee (12 members) consists currently of three ministers of regional governments, six (other) members of regional parliaments, one local councillor (schepen/échevin) and two former ministers of a regional government. 22 See, eg, the opinions in which the Committee claimed standing before the ECJ. The Lisbon Treaty has granted this standing for the purpose of protecting the Committee’s prerogatives or to denounce a breach of the subsidiarity principle. See Art 263, para 3 TFEU. 23 S Ricci, ‘The Committee of the Regions and the Challenge of European Governance’ in Panara and De Becker (eds), above n 14, 128.

200 Piet Van Nuffel C. The Lisbon Treaty: New Institutional Avenues Neither the composition of the Council, nor the role and composition of the Committee of the Regions were substantially amended by the Treaty of Lisbon. Thus, the ‘institutional avenues’ which the Maastricht Treaty introduced to enable regions directly to influence the EU legislative process were largely left untouched. Instead, the Lisbon Treaty opted for a strengthening of the control exercised by national parliaments over EU decision-making, and this in two ways. On the one hand, the Lisbon Treaty emphasised the accountability of the members of the Council to their national parliaments,24 and reinforced the guarantees for national parliaments to be informed about the proposed agenda of the Council’s meetings.25 Through these amendments, the Lisbon Treaty facilitated political control by national parliaments on the activities of the Member States’ representatives sitting in the Council. On the other hand, the Lisbon Treaty increased the possibility of national parliaments putting pressure on EU decision making. National parliaments are now entitled to investigate—ex ante—whether the proposed legislation is in line with the principle of subsidiarity.26 Although the Lisbon Treaty’s innovations mainly strengthen the role of national parliaments, they may also increase the chances for regional politicians to advance regional interests. Indeed, by trying to influence the national parliament’s opinion, regional politicians may seek to protect the interests of their region. Thus, the stronger role of national parliaments also facilitates the possibility of regions indirectly influencing EU decision making. The question arises, however, whether the institutional avenues designed to strengthen the role of national parliaments may also be used by regional authorities to gain direct influence over EU decision making. i. Accountability of Ministers to National Parliaments First of all, the Lisbon Treaty introduced amendments that impact on the accountability of the members of the Council and the European Council to their national parliaments. The Treaties now make clear that this accountability is an element of the democratic foundation of the European Union. The new Treaty provisions on democratic principles expressly mention that the EU citizens are ‘directly represented at Union level in the European Parliament’, whereas ‘Member States are represented in the European Council by their Heads of State or Government and in the Council by their governments, themselves democratically accountable either to their national Parliaments, or to their citizens’ (Article 10(2) TEU). The Treaties thus emphasise that representation of citizens takes place, first and foremost, through the European Parliament and, secondly, where the European 24 25 26

Art 10(2) TEU. See Protocol No 1 on the role of national parliaments in the European Union [2010] OJ C83. See Art 7 of the Subsidiarity Protocol.

Does EU Decision-Making Take into Account Regional Interests? 201 Parliament does not take part in the adoption of an act, through the national parliaments to which the government representatives in the European Council and the Council are politically accountable. National parliaments may influence decision making at the EU level by bringing pressure to bear on members of the Council, who are answerable to them. Nonetheless, parliamentary scrutiny of the role played by the national government in EU decision making is fully effective only if three conditions are fulfilled: first, the national parliament must be sufficiently informed of the Council’s activities; secondly, the national parliament must be capable of actually influencing the member of the Council; thirdly, that member can be called to account for how he or she voted in the Council.27 In practice, the influence of national parliaments varies greatly, depending on the Member State considered. Whereas the fulfilment of the second and the third conditions is a matter of national law, recent Treaty amendments have had some bearing on the first condition, as they significantly enhanced the national parliaments’ right to information. Since the Lisbon Treaty, the Protocol on the role of national parliaments provides that all proposals and other draft legislative acts presented to the European Parliament and to the Council must simultaneously be sent to (both chambers of) the national parliaments.28 Before, it was merely provided that Commission proposals for legislation were to be made available in good time, ‘so that the government of each Member State may ensure that its own national parliament receives them as appropriate’.29 This provision did not give the national parliaments an autonomous right to information. Yet in some Member States, the government was bound under domestic law to supply parliament (or the competent parliamentary committee) with all Commission proposals and to inform parliament of the stance it intended to take in the Council.30 The Lisbon Treaty conferred upon the national parliaments an autonomous right to receive legislative proposals. In order to allow enough time for discussion of draft legislative acts, the Protocol on the role of national parliaments requires there to be an eight-week period between the time when a Commission draft is made available (in all languages) to the European Parliament and the Council, and the date when it is placed on a Council agenda for decision. In order to enable the national parliaments to hold the member of the national government accountable, the Treaties now expressly state that the Council meets in public when it deliberates and votes on a draft legislative act (Article 16(8) TEU). The national parliaments are also entitled

27

See Lenaerts and Van Nuffel, above n 12, 615. Art 2 of Protocol No 1 on the role of national parliaments in the European Union. The national parliaments must also be informed in advance if the Commission is to submit a proposal for application of the flexibility clause (Art 352(2) TFEU). 29 Protocol on the role of national parliaments in the European Union, as annexed to the Treaties by the Amsterdam Treaty [1997] OJ C340, para 2. 30 See Lenaerts and Van Nuffel, above n 12, 616–17. 28

202 Piet Van Nuffel to receive the outcome of meetings of the Council, including the minutes of meetings where the Council is deliberating on draft legislative acts.31 The Protocol on national parliaments refers only to (component chambers of) ‘national parliaments’. The question therefore arises whether the information rights granted by the Protocol to national parliaments also apply to parliamentary organs of the Member States’ regions. The Protocol does not define the notion of ‘national parliament’, leaving it to the Member States to designate the authorities they regard as falling within the notion’s ambit. That means that regional parliaments cannot claim any right to be informed on the activities of the Council and on the outcome of its discussions unless they are, under the law of the Member State, considered to be at the same level as ‘national’ parliaments. This is the case in Belgium, for example. At the signature of the Lisbon Treaty, the Belgian State proclaimed in a ‘declaration on national parliaments’ that in accordance with its constitutional law, not only the Chamber of Representatives and Senate of the Federal Parliament but also the parliamentary assemblies of the Communities and the Regions act, in terms of the competences exercised by the Union, as components of the national parliamentary system or chambers of the national Parliament.32

As regards the distribution of information on draft legislative acts, the parliaments of the Belgian regions are put on the same level as the chambers of the federal parliament. In practice, all information sent by the EU institutions to the functional mailbox of the Belgian national parliament (administrated by the Senate) is automatically passed on to the relevant mailboxes of the regional parliaments. In sum, the information rights conferred upon national parliaments by the Lisbon Treaty do not guarantee that those parliaments can put political pressure on the Member State’s representative in the Council, but they certainly make it easier for them to exert influence. For regional parliaments, these information rights may be useful where regional politicians and/or institutions seek to impact on the position taken by national parliaments in matters to be discussed in the Council. The example of Belgium shows that this indirect type of regional influence on EU decision making may also become a direct tool for influence where a Member State takes a broad interpretation of the notion of national parliaments. In this way, regional parliaments will be informed of the Council’s activities, allowing them to steer the activity of the Member State’s (federal or regional) representative in the Council.

31

Art 5 of Protocol No 1 on the role of national parliaments in the European Union. Declaration No 51 by the Kingdom of Belgium on national Parliaments [2010] OJ C83/355. On this issue, see W Pas, ‘The Belgian “National Parliament” from the Perspective of the EU Constitutional Treaty’ in P Kiiver (ed), National and Regional Parliaments in the European Constitutional Order (Groningen, Europa Law Publishing, 2006) 57; L Van Looy, ‘Het Vlaams Parlement als “nationaal parlement” in de Europese Unie (Ceci n’est pas une fiction)’ (2007) 1 Tijdschrift voor Wetgeving 28. 32

Does EU Decision-Making Take into Account Regional Interests? 203 ii. Subsidiarity Scrutiny by National Parliaments Secondly, the Subsidiarity Protocol lays down a procedure allowing national parliaments to intervene in the EU decision-making process if they find legislative proposals to be incompatible with the subsidiarity principle.33 The Subsidiarity Protocol gives any national parliament (or any chamber of a national parliament) eight weeks to issue a reasoned opinion stating why it considers that the draft in question does not comply with the subsidiarity principle. If the draft legislative act concerns a matter for which, under national law, competence rests with regional parliaments with legislative powers, the national parliament is to consult these regional parliaments ‘where appropriate’.34 The Commission, the European Parliament and the Council must take account of the reasoned opinions issued by national parliaments. The Commission is to review its draft where reasoned opinions on non-compliance with the subsidiarity principle represent at least one-third of all the votes of the national parliaments (one-quarter in the case of a Commission proposal or an initiative emanating from a group of Member States related to police co-operation or judicial cooperation in criminal matters). After having reviewed its draft legislative act, the Commission may decide to maintain, amend or withdraw it, while giving its reasons therefor.35 To calculate the threshold of one-quarter or one-third of the allocated votes, every national parliament will have two votes, shared out on the basis of the national parliamentary system. In the case of a bicameral parliamentary system, each of the two chambers will have one vote.36 This procedure received the name ‘yellow card’, after the football term for ‘warnings’ issued by referees. With respect to matters falling under the ordinary legislative procedure (the previous ‘co-decision procedure’), the Subsidiarity Protocol provides for the possibility that a legislative proposal may be rejected. The final decision as to whether a legislative proposal must be discarded is to be made not by the national parliaments but by the EU legislator itself. Under this so-called ‘orange card’ procedure, the Commission must review its draft proposal if the reasoned opinions on non-compliance with the subsidiarity principle represent at least a majority of the votes allocated to the national parliaments. In that event, the Commission can maintain its proposal only if the alleged compliance with the subsidiarity principle is explained in a reasoned opinion. Moreover, in such a case, the EU legislator shall ‘not give further consideration’ to the proposal if the Council (by a majority of 55 per cent of the Member States) or the European Parliament (with a

33

Art 6 of the Subsidiarity Protocol. Ibid. 35 Ibid Art 7(1)–(2). 36 Ibid Art 7(1). This seems to be the case even if these chambers are not equal under domestic laws. See P Kiiver, ‘The Treaty of Lisbon, the National Parliaments and the Principle of Subsidiarity’ (2008) 15 Maastricht Journal of European and Comparative Law 80. 34

204 Piet Van Nuffel majority of the votes cast) considers that it is not compatible with the subsidiarity principle.37 The possibility for the Council and the European Parliament to reject further consideration of a proposal exists only where the Commission ‘maintains’ and not ‘amends’ or ‘withdraws’ it proposal. Of course, were the Commission to decide to amend its proposal, the amended proposal will pass this hurdle only if the amendments actually meet the concerns expressed by the national parliaments. If not, the proposal will have to be considered ‘maintained’. The refusal by the EU legislator to consider a proposal further may appear more radical than it really is. If a majority is found to vote against the proposal, the chance of the proposal being adopted would be rather small anyway, irrespective of whether national parliaments have formally voiced their criticisms.38 The most interesting cases will, therefore, be those where reasoned opinions have been issued by several, though not the majority, of the national parliaments. Through this early warning procedure, the Treaties not only entitle national parliaments formally to issue an opinion on proposed EU legislation, but also require regional parliaments with legislative powers to be consulted by their national (federal) parliament. That does not mean, however, that whenever a regional parliamentary body of a Member State has not been consulted, it can rely on the Subsidiarity Protocol legally to contest this course of action. With respect to regional parliaments, the Subsidiarity Protocol does not lay down a directly enforceable right. Indeed, apart from the fact that the existence of regional parliaments with ‘legislative powers’ has to be determined pursuant to the constitutional law of the Member State concerned, the Protocol provides for the consultation of such regional parliaments ‘where appropriate’. This seems to leave some discretion to the national constitutional order to determine the cases in which regional parliaments need to be consulted.39 As mentioned above, Belgium went further by adding a formal declaration to the Lisbon Treaty which makes clear that the Member State will put the parliaments of its regions at the same level as the chambers of the federal parliament for the application of the Subsidiarity Protocol. Earlier, Belgium had made a similar declaration in relation to the Constitutional Treaty. That declaration was implemented by means of a draft co-operation agreement initialled by both chambers of the federal parliament and the parliaments of the regions.40 Like the Constitutional Treaty, that co-operation agreement never entered into force. After

37

Art 7(3) of the Subsidiarity Protocol. See also Kiiver, above n 36, 81. 39 The Committee of the Regions, however, invites national parliaments ‘effectively and transparently’ to consult the regional parliaments with legislative powers, inviting all regional parliaments expressly to ‘liaise’ with it. See Committee of the Regions, Opinion of 16 November 2005 on guidelines for the application and monitoring of the subsidiarity and proportionality principles (CdR 220/2004 fin), paras 2.22–2.23. 40 Co-operation Agreement of 19 December 2005 between the Federal Parliament, the parliaments of the Communities, and the parliaments of the Regions (Flemish Parliament 2005–2006, No 628/1, docs.vlaamsparlement.be/docs/stukken/2005-2006/g628-1.pdf). 38

Does EU Decision-Making Take into Account Regional Interests? 205 the signature of the Lisbon Treaty, a new (draft) co-operation agreement was initialled by the presidents of the federal and regional parliaments in July 2008.41 However, for internal political reasons, the new co-operation agreement has not yet been formally signed. Interestingly, the envisaged co-operation agreement lays down rules with respect to the division of competences among the various parliamentary bodies and the allocation of the (maximum two) votes to be issued on behalf of the ‘national parliament’. In those instances where either the federal State or the regional entities have exclusive powers, two ‘votes’ will be cast if both chambers of the federal parliament (in matters of exclusive federal power) or at least two regional parliaments from different language groups (in matters of exclusive regional powers) give a reasoned opinion. In areas of concurrent powers, by contrast, two ‘votes’ will be expressed if at least one federal chamber and one regional parliament issue such an opinion. In the Belgian understanding of the Lisbon Subsidiarity Protocol, the rights granted to national parliaments can thus be exercised by the relevant regional parliaments. There is no reason why it should not be for the constitutional order of each federal Member State to determine the internal organs which may represent the ‘national’ point of view at the EU level. This is indeed what happens when a federal State, pursuant to Article 16(2) TEU, sends ministers of regional governments to the Council to represent the Member State. However, since the Subsidiarity Protocol itself distinguishes between ‘national’ and ‘regional’ parliaments, it remains to be seen whether the other Member States and the Union institutions will permit each Member State to interpret the notion of ‘national parliament’ as it sees fit.42 Be that as it may, it would be rather paradoxical to interpret a text designed to give regional entities a say in the EU decision-making process in a way that denies Member States the freedom to achieve this goal. Commentators have rightly stressed the practical difficulties national parliaments may experience when implementing the early warning system. For national parliamentarians, the eight-week period may be too short to identify, amongst the various national and European preparatory acts submitted to them, those European documents that need their attention from the angle of the subsidiarity principle. This is especially true if members of parliament need to consult experts from the national administration, their respective political parties or their colleagues in other parliaments.43 The same concern would apply to members of regional parliaments.

41 Draft Co-operation Agreement of 7 and 9 July 2008 (Flemish Parliament 2007-2008, No 1807/1, docs.vlaamsparlement.be/docs/stukken/2007-2008/g1807-1.pdf). 42 In another context, the Court of Justice has been reluctant to recognise, for instance, any freedom for Member States to have regional entities acting on behalf of the ‘Member State’. See Case C-95/97 Région wallonne v Commission [1997] ECR I-1787, paras 6–8. See also Van Nuffel, above n 1, 871–901. 43 See NW Barber, ‘Subsidiarity in the Draft Constitution’ (2005) 11 European Public Law 204; O Tans, ‘De oranje kaart: een nieuwe rol voor nationale parlementen?’ (2007) 11 Sociaal-Economische Wetgeving 443.

206 Piet Van Nuffel Nonetheless, national and regional parliaments will not act alone. Both the Committee of the Regions and the Conference of committees on European affairs of the Union’s parliaments (COSAC) have shown their eagerness to offer technical assistance through the establishment of networks in which information and expertise may be shared. Without any doubt, the early warning procedure encourages national parliaments to reflect and formulate opinions on proposed EU acts. Even if only a limited number of national parliaments raise doubts as to the compatibility of the proposed act with the subsidiarity principle, the EU institutions are under political pressure to respond to such doubts. The Commission will have to come up with convincing arguments to soothe any unrest that such negative opinions would cause among members of the Council and/or the European Parliament. The Council members and MEPs should also be able to explain the position they take on the concerns expressed by the national parliaments. In practice, the introduction of the early warning procedure may thus lead to higher standards of reason giving as regards the EU’s compliance with the subsidiarity principle.44 Admittedly, the very fact that certain national or regional parliaments consider a proposal to be incompatible with the subsidiarity principle, does not necessarily mean that the proposal, if adopted, would turn out to fail the subsidiarity test laid down in Article 5(3) TEU. Given the difficulties Member States experience in demonstrating their ability to achieve the objectives of the legislation proposed at the EU level,45 the EU legislator would not necessarily take a great risk by adopting a proposal in spite of the negative opinions from one or more national parliaments. Moreover, the doubts expressed by national parliaments may not be limited to the application of the subsidiarity principle but also relate to other concerns. Nevertheless, the early warning procedure’s success should not be measured solely in terms of the degree to which it facilitates judicial subsidiarity review or prevents infringements of the subsidiarity principle. The great advantage of the early warning procedure is that it may stimulate national and regional parliaments to participate in the discussion of proposed EU legislation. Even where a national or regional parliament prefers not to issue a negative opinion, or is unable to take a stance on the proposed EU act, the very existence of the early warning procedure obliges domestic parliaments at least to reflect on EU legislation before its adoption. As one commentator puts it, the very possibility to complain about draft EU legislation makes it embarrassing for national parliamentarians to face criticism of EU legislation when they have not even tried to make use of this possibility.46 For regional politicians and institutions, subsidiarity scrutiny offers another route indirectly to influence the EU legislative process. The early warning procedure may not only put pressure on national (or regional) governments, but also alert other

44 On the statement of reasons with respect to the subsidiarity principle, see P Van Nuffel, ‘The Protection of Member States’ Regions through the Subsidiarity Principle’ in Panara and De Becker (eds), above n 14, 66–68. 45 On the subsidiarity test, see Van Nuffel, above n 44, 58–61. 46 Kiiver, above n 36, 82–83.

Does EU Decision-Making Take into Account Regional Interests? 207 actors in the EU legislative process. Accordingly, the Committee of the Regions has sought the national parliaments’ co-operation in order to develop ‘joint strategies’ for the effective application of the Subsidiarity Protocol.47 The Committee of the Regions announced that it would itself engage in a subsidiarity assessment in its (mandatory or voluntary) opinions, to which the eight-week deadline does not apply.48 Moreover, even where the Commission is unwilling to meet all concerns expressed by national parliaments, such concerns might be taken up by some national governments or by members of the European Parliament.

V. INFLUENCE OF REGIONAL INTERESTS IN CONSTITUTIONAL DECISION-MAKING

The intense debate that has surrounded the failed Constitutional Treaty and its more successful successor, the Lisbon Treaty, demonstrates the political importance of the European Union’s constitutional decision-making process. Indeed, whereas the great majority of EU rules are framed in the context of the EU legislative process based on the interaction between the Commission, the Council and the European Parliament, the most important rules are still those enshrined in the Treaties. Article 48 TEU defines the procedures for amending the TEU and TFEU. It sets out both the ‘ordinary revision procedure’, which is the general amendment procedure, and ‘simplified revision procedures’. The latter category of procedures makes it possible to amend certain Treaty provisions without needing to satisfy all the requirements of the ordinary revision procedure such as the convening of an intergovernmental conference and the ratification of the amendments by all Member States. The ordinary revision procedure49 basically requires Treaty amendments to be determined by a conference of representatives of the Member State governments (the intergovernmental conference or IGC). The Lisbon Treaty has introduced a role for the European Council in the revision procedure. More importantly, since the Lisbon Treaty, the IGC needs to be preceded by a ‘Convention’ consisting of representatives of the national governments, the national parliaments, the European Parliament and the Commission. The task of the Convention is to adopt, by consensus, a recommendation to the IGC. The ultimate decision on Treaty amendments is still left to the IGC.50 Moreover, the amendments do not enter into force until they have been ratified by all the Member States ‘in

47

Committee of the Regions, Opinion of 16 November 2005, above n 39, para 2.22. Ibid paras 3.9–3.12. 49 Art 48(2)–(5) TEU. 50 However, the European Council may decide, after obtaining the consent of the European Parliament, that the extent of the proposed amendments does not justify convening such a Convention (Art 48(3) TEU). The European Council made use of this possibility at the occasion of the first amendment proposed after the Lisbon Treaty’s entry into force. That amendment allowed the number of MEPs to be increased for a transitional period. See Lenaerts and Van Nuffel, European Union Law, above n 12, 77. 48

208 Piet Van Nuffel accordance with their respective constitutional requirements’.51 A Member State is at liberty to frame its constitutional law in such a way that the government has to consult the national parliament before approving the proposed text.52 Likewise, a Member State may make ratification of a Treaty amendment dependent upon the approval of all its competent regions.53 This was the approach taken by Belgium as regards the ratification of all recent Treaty amendments, including the Lisbon Treaty. If a Member State involves its regions in the domestic procedure for the ratification of Treaty amendments, those regions are given a direct say in the EU’s constitutional process. To be sure, even under that hypothesis, regional authorities do not necessarily have much leeway actually to influence the contents of the Treaty amendments at the stage of ratification. Such amendment is the result of negotiations among the Member States’ governments, and authorities involved in the notification process can only approve or reject the amendment but not propose any changes to it. Nevertheless, the Member State’s central government will likely be more willing to take into account certain regional concerns during the negotiation if the consent of the regional authorities is required at the stage of ratification. In those cases where amendments are decided upon by the European Council under the ‘simplified’ revision procedure of Article 48(7) TEU, a national parliament can make known its opposition against a draft amendment initiated by the European Council within six months of the date of the notification of the European Council decision. In several Member States, specific rules have been laid down in order to enable national parliaments formally to ‘oppose’ an amendment proposed by the European Council.54 Although Article 48(7) TEU obliges national parliaments to take the initiative to oppose a proposed amendment, some Member States also require the national government to obtain parliamentary consent before supporting the proposed decision of the European Council.55 In Member States with a federal structure, regional parliaments may be involved in these proceedings. In Belgium, for instance, the aforementioned draft co-operation agreement provides that any competent parliamentary organ, be it at the federal or the regional level, is entitled to pronounce a veto against a decision of the European Council based on Article 48(7) TEU. In that event, the veto would be presented as emanating from Belgium’s ‘national parliament’.56 It is clear that this system gives far-reaching powers to any regional parliament within the Belgian federal context. Indeed, the parliament of any Belgian

51 This ratification requirement is not part of the ‘simplified revision procedures’. Nevertheless, a national parliament can make known its opposition against a draft amendment initiated by the European Council ‘within six months of the date of [the] notification [of the European Council decision]’ (art 48(7) TEU). 52 See Lenaerts and Van Nuffel, above n 12, 610. 53 Ibid 611. 54 See n 51 above. 55 See Lenaerts and Van Nuffel, above n 12, 612. 56 See n 41 above.

Does EU Decision-Making Take into Account Regional Interests? 209 region which has competence in the relevant field (eg, the parliament of the small German-speaking Community) is able to block, at EU level, a decision of the European Council. The logic applied here is the same as that applied to ‘ordinary’ Treaty amendments, whereby any competent region needs to consent to the amendment in order for it to be ratified by Belgium and, hence, for it to enter into force at the EU level.

VI. CONCLUDING REMARKS

The Treaties no longer prevent a Member State from having its internal system of regional autonomy reflected at the EU level. They allow for the expression of regional interests in the EU decision-making process, not only in an indirect way, through the influence exercised by regional politicians on the members of the EU institutions, but also in a more direct manner, in particular through the participation of regional politicians in the activities of the Council. As far as constitutional decision making is concerned, Member States may even give regional parliaments a decisive role, allowing them to participate at the same level as national parliaments. This does not mean, however, that the Treaties themselves contain any legal support for regional autonomy. It remains for every single Member State to decide on the degree of regional autonomy; a decision the EU ought to respect as part of the Member State’s national identity. It would certainly not be incompatible with EU law to recognise a right to self-government for Member State regions. Yet the issue of self-government does not play the same role in every Member State. As a result, imposing a system of regional self-government on the Member States would likely go against the choice of certain Member States to maintain a more centralised political system and, hence, infringe upon their ‘national identity’. European Union law should therefore limit itself to ensuring recognition for the political interests arising in those Member States which have a decentralised or federal system. If a Member State decides to put its regional institutions on equal footing with its ‘central’ political institutions—as exemplified by the Belgian system—EU law should be interpreted so as to give that Member State the freedom to express its political interests, including regional interests, at the EU level.

9 The Role of Sub-State Entities in the EU Decision-Making Processes: A Comparative Constitutional Law Approach NIKOS SKOUTARIS*

I. INTRODUCTION

I

N JUNE 2008, the Finnish Parliament approved the text of the Lisbon Treaty. Three months later, the formal ratification of the Treaty was finalised with the signature of the President of the Republic of Finland, Tarja Halonen. However, it was only on 25 November 2009—a week before the Treaty came into force—that the Parliament of the Åland Islands, an autonomous region of Finland,1 decided with a majority of 24 to 6 that the Treaty would also apply there. In fact, during the previous months, the Åland Government had put forward four requests that had to be resolved before accepting the Treaty. The autonomous region had requested its own seat in the European Parliament; a right to appear before the Court of Justice; participation in the control of the principle of subsidiarity; and participation in the meetings of the Council. All the requests were satisfied except the seat in the Parliament.2 The aforementioned largely unknown episode in the Lisbon Treaty ratification saga sheds light on the efforts of a number of sub-State entities to achieve an enhanced role in the Union decision-making processes. At the same time, certain recent developments concerning both the text of the Treaties and the case law of the Court of Justice might suggest that the European Union (EU) has also become increasingly more willing and able to accommodate similar aspirations of the sub-State entities regarding their role in the political life

* Assistant Professor, Maastricht University. The author participates in the European and National Constitutional Law project which is funded by the ERC (). 1 For a review of the status of the Åland Islands in the Finnish constitutional order, see generally J Huusa, The Constitution of Finland (Oxford, Hart Publishing, 2011) 121–29. 2 For an informative analysis, see P Alilonttinen and S Ruà, Lisbon Treaty Ratification: Will the Åland Islands become Finland’s Greenland? (Brussels, EPIN, Commentary no 1/2008).

The Role of Sub-State Entities in the EU Decision-Making Processes 211 of the EU. With regard to the Treaties, we note that Article 4(2) TEU now states that ‘[t]he Union shall respect the equality of Member States before the Treaties as well as their national identities, inherent in their fundamental structures, political and constitutional, inclusive of regional and local self-government’.3 Moreover, according to Article 5(3) TEU, which provides for the new formulation of the principle of subsidiarity, the Union—outside the areas of its exclusive competence— may act only in so far as the objectives of the proposed action ‘cannot be sufficiently achieved by the Member States, either at central level or at regional and local level’.4 Next to those explicit references to the regional tiers of Member States, one might also note the fact that the Lisbon Treaty includes a number of new provisions for the participation of national parliaments, which are enjoined ‘to contribute actively to the good functioning of the Union’, ‘by seeing to it that the principle of subsidiarity is respected in accordance with the procedures provided for in the Protocol on the application of the principles of subsidiarity and proportionality’.5 This provision together with Protocol No 2 might enhance further the role of the sub-State entities in the EU decision-making process, especially in those States where there is a chamber composed of representatives of the regions, such as in Germany and Austria.6 In addition, a review of recent case law demonstrates that the Court is ready to recognise the constitutional autonomy of certain regions to exercise policy choices that differ from one region of a given State to the next, and that the Court has thereby become more mindful of the regional dimension of the Union structure.7 Such developments in the EU constitutional order cast a serious doubt on that part of the academic literature that has suggested since the early days of the integration process that the Union is ‘blind’ to the internal territorial and constitutional arrangements of its Member States.8 So, it is precisely the aforementioned aspirations of the regional tiers and the political and legal developments in the EU sphere that dictate this examination of the possible channels of regional participation in the Union policy-making processes both at the national and at the EU levels. With regard to the national level, the chapter mainly focuses on the constitutional orders of those Member States 3

Emphasis added. Emphasis added. 5 Art 12(b) TEU. 6 D Edward and J Bengoetxea, ‘The Status and Rights of Sub-state Entities in the Constitutional Order of the European Union’ in A Arnull, C Barnard, M Dougan and E Spaventa (eds), A Constitutional Order of States? Essays in EU Law in Honour of Alan Dashwood (Oxford, Hart Publishing, 2011) 31. See also section III.B. of this chapter. 7 See generally Case C-88/03 Portugal v Commission [2006] ECR I-7115; Joined Cases C-428/06 to C-434/06 Unión General de Trabajadores de La Rioja (UGT-Rioja) and Others v Juntas Generales del Territorio Histórico de Vizcaya and Others [2008] ECR I-6747; Case C-428/07 R (on the application of Horvath) v Secretary of State for the Environment, Food and Rural Affairs [2009] ECR I-6355. For an analysis this case law, see also R Greaves, ‘Autonomous regions, taxation and EC state-aid rules’ (2009) 34 EL Rev 779. 8 See, eg, HP Ipsen, ‘Als Bundesstaat in der Gemeinschaft’ in E von Caemmerer, HJ Schlochauer and E Steindorff (eds), Probleme des europäischen Rechts: Festschrift für Walter Hallstein (Frankfurt am Main, Klostermann, 1966) 248. 4

212

Nikos Skoutaris

where the regional tier enjoys a constitutionally grounded claim for participation in the policy-making processes. Those Member States include the federal Austria, Belgium and Germany, and the regionalised Italy, Spain and the United Kingdom. In particular, we refer to the duty that some of the governments of those Member States have to inform their sub-State entities, and to the mechanisms provided for the participation of the regional tier in the formulation of the EU position of the Member State. Concerning the EU level, we examine the presence of the regional tier in the Union institutions and analyse the new system for subsidiarity compliance. Generally speaking, the examination of both the national and the EU levels underlines the complexity of the institutional framework that allows for the participation of the regional tiers in the Union decision-making processes, and the intertwined9 and composite10 nature of the European constitution. The very presence of such institutional framework renders—partly at least—the ‘regional blindness’ thesis obsolete. However, the fact that only rather a limited number of entities benefit from those arrangements sheds light on the asymmetries of the EU’s constitutional order and points to the fact that we are still far from the enthusiastic vision of a ‘Europe of the Regions’ as described in the 1990s.11

II. PARTICIPATION IN THE INTERNAL DECISION-MAKING PROCESS

The question underpinning this part of the chapter is whether the regional tier enjoys participatory rights in the formulation of the Union position of the respective Member State. In particular, we focus on two distinct but interlinked issues. First, we examine whether the regional authorities have a right—constitutionally enshrined or not—to be informed in the pre-legislative phase, and then we look at what kind of mechanisms are in place for those regional authorities to express their opinion.

A. Informing the Regional Tier in the Preparatory Phase Timely access to relevant information is of critical importance for the effective participation of the regional authorities in the preparatory phase of Union decision-making processes. Although it is difficult to over-state the importance

9 See generally J Ziller, ‘National Constitutional Concepts in the New Constitution for Europe’ (2005) 1 European Constitutional Law Review 247, 452. 10 L Besselink, A Composite European Constitution/Een Samengestelde Europese Constitutie (Groningen, Europa Law Publishing, 2007); M Claes, The National Courts’ Mandate in the European Constitution (Oxford, Hart Publishing, 2005). 11 See generally, S Mazey and J Mitchell, ‘Europe of the Regions? Territorial Interests and European Integration: The Scottish Experience’ in S Mazey and J Richardson (eds), Lobbying in the European Community (Oxford, Oxford University Press, 1993); J Loughlin ‘ “Europe of the Regions” and the Federalization of Europe’ (1996) 26 Publius 141.

The Role of Sub-State Entities in the EU Decision-Making Processes 213 of the availability of information in order for the sub-State entities effectively to influence the EU law-making processes, in most Member States the governments are not legally obliged to inform such entities of Union matters that affect them.12 And even in those cases where such a duty of information exists, the source of regulation used might differ (constitutional, legislative, non-legislative).13 It is hardly surprising that such a constitutional duty to keep the regional authorities informed exists in the three EU Member States that are federations, namely, Austria, Germany and Belgium. In fact, the Austrian Constitution goes as far as to impose a requirement on the federal Government to inform the regional and local authorities both directly14 and indirectly, through the Bundesrat.15 Similar duties to provide information indirectly to the sub-State entities through the legislative chambers representing them are also provided for by the German Basic Law16 and the Belgian Constitution. In the case of Belgium it should be noted that Article 168 of the Constitution provides for such a duty only in the case of the opening of negotiations for the amendment of the EU Treaties.17 However, such right has been extended to the Communities and the Regions under Article 16(2) of the Special Reform Law of 8 August 1980. In addition, Article 92quater of the same law, introduced by the Special Reform Law of 5 May 1993, extended the limited constitutional duty of information to cover Union secondary acts too. The Belgian Senate automatically forwards all EU documents that arrive in its relevant mailbox to the House of Representatives and all legislatures of the sub-State entities. To complete the picture, apart from the constitutions of the three federations, Article 88-4 of the Constitution of the unitary but decentralised France provides for a duty to provide information indirectly to the regions.18 A duty to inform the sub-State entities may also be provided by ordinary national laws. This is the case in Italy and Spain. In Italy, Law No 11/2005, which regulates regional participation in European policy-making, provides in Article 5 that when the relevant EU draft legislation concerns regions and local authorities, it should be transmitted to the competent territorial associations for comment. These associations include the Conference of the Regions and the Autonomous Provinces (Conferenza delle regioni e delle province autonome, hereafter ‘CRPA’) and the Conference of the Presidents of the Assembly of Regional Councils and of Autonomous Provinces. Upon reception, the draft legislation is forwarded by those two associations to the presidents of the regional executive committees and of the regional councils. 12 M Tatham, ‘Devolution and EU policy-shaping: Bridging the gap between multi-level governance and liberal intergovernmentalism’ (2011) 3 European Political Science Review 53, 58. 13 Ibid. 14 Austrian Constitution, Art 23d(1). 15 Austrian Constitution, Arts 23e(1) and 23g(2). 16 German Basic Law, Art 23(2). 17 Belgian Constitution, Art 168. 18 French Constitution, Art 88-4.

214

Nikos Skoutaris

In Spain, the duty to inform the regional authorities used to be contained in a non-legislative act: the Agreement on the Internal Participation of the Autonomous Communities in EC Affairs through Sector Conferences of 30 November 1994.19 That Agreement was concluded three years before the Sectoral Conference for European Affairs (Conferencia para Asuntos Relacionados con las Comunidades Europeas, hereafter ‘CARCE’) was formally institutionalised by virtue of Law 2/97. However, recently, Article 6 of Law 24/2009 of 22 December 2009 has established the national Parliament’s duty to transmit any EU draft legislative act to regional parliaments, without any filtering procedure. Lastly, the duty of the UK Government to inform the sub-State entities is provided for by provisions included in non-legislative acts. The relative framework may be found in the Memorandum of Understanding and the Concordats on Co-ordination of European Union Policy Issues between the UK Government and the devolved administrations.20

B. Mechanisms for Involving the Regional Tier in Internal EU Decision-Making Processes After examining the constitutional, legal and non-legal provisions that provide for the duty of the central State governments to inform the sub-State entities, the next question is what kind of mechanisms are in place for the participation of the regional tier in the internal decision-making processes concerning EU law. In general, the participation of the regional authorities in the formulation of the EU policy of the respective Member State is facilitated by legislative chambers composed of representatives of the regions and inter-governmental bodies, whether interregional or joint national–regional ones. In this section, I focus mainly on two fundamental questions. First, I examine whether the participatory rights of the regions in the internal EU decisionmaking process are either constitutionally or legally guaranteed, or guaranteed by non-legislative means. In the case of the upper chambers, the answer is rather straightforward; in the case of the co-ordination bodies, the picture is more mixed. Secondly, I analyse whether the position adopted by the regions through those mechanisms is binding on the respective Member State.

19 El Acuerdo sobre la Participación Interna de las Comunidades Autónomas en los asuntos comunitarios a través de las Conferencias Sectoriales, de 30 de noviembre de 1994. The Agreement was published in the Resolution of 10 March 1995 of the Ministry of Territorial Administration (Resolución de 10 de marzo de 1995 de la Secretaría de Estado para Administraciones Territoriales) and in the Official Journal (BOE of 22 March 1995 and BOE of 1 April 1995—corrigendum). 20 Memorandum of Understanding, paras 17–20; Concordats on Co-ordination of European Union Policy Issues (Cm 5240, December 2001, Memorandum of Understanding and Supplementary Agreements Between the United Kingdom Government Scottish Ministers, the Cabinet of the National Assembly for Wales and the Northern Ireland Executive Committee), paras B1.2, B1.4, B2.2, B2.4, B3.2, B3.4, B4.2 and B4.3.

The Role of Sub-State Entities in the EU Decision-Making Processes 215 i. Upper Chambers One of the characteristics of the vast majority of federations is the existence of a constitutionally enshrined body that represents the regional tier at the highest level and allows it to play an important role in the national decision-making processes.21 In that sense, the fact that both the Austrian and the German Bundesräte play a pivotal role in the formulation of the respective national policy does not raise any eyebrows. As already mentioned, under Article 23e(1) of the Austrian Constitution, the Government informs the Bundesrat about ‘all projects within the framework of the European Union’ and gives them the chance to express their opinion. Where the proposed Union legislation is required to be implemented in accordance with a procedure which requires the agreement of the Bundesrat,22 then the Government is bound by the opinion of the upper chamber during the negotiations that take place in the EU framework. The Government may deviate from such binding opinion only in the event of ‘imperative foreign and integrative policy reasons’.23 Article 23(4)–(5) of the German Basic Law and an ad hoc Act of Co-operation in 199324 regulate the relationship in EU affairs between the federal Government and the 16 Bundesländer that are united in the Bundesrat. According to Article 23(4), the Government informs the Bundesrat, which may participate ‘in so far as it would have been competent to do so in a comparable domestic matter’.25 Each Land having a weighted vote, the Bundesrat adopts with a majority a common position of the Länder. The opinion of the Bundesrat carries varying degrees of influence depending on what kind of competences the relevant decision concerns. If the relevant decision concerns an exclusive competence of the federal Government, the opinion of the Bundesrat merely needs to be taken into account.26 If the decision affects the legislative powers of the Länder, the structure of Land authorities, or Land administrative procedures … the position of the Bundesrat shall be given the greatest possible respect in determining the Federation’s position consistent with the responsibility of the Federation for the nation as a whole.27

In the event of disagreement, there is a conciliation procedure.28 The federal Government may override the Bundesrat veto in cases where the general political

21 RL Watts, Comparing Federal Systems, 3rd edn (Montreal and Kingston, McGill-Queen’s University Press, 2008) 147–53. 22 Austrian Constitution, Art 44(2). 23 Austrian Constitution, Art 23e(6). 24 European University Institute, Study on the Division of Powers between the European Union, the Member States, and Regional and Local Authorities (Florence, European University Institute, 2008) 148. 25 German Basic Law, Art 23(4). 26 German Basic Law, Art 23(5). 27 Ibid. 28 European University Institute, above n 24, 148.

216

Nikos Skoutaris

responsibility of the federation and its financial interests are at stake.29 To complete the picture concerning the role of the Bundesrat, it should be noted that, by virtue of Article 23(1), its approval by a two-thirds majority is necessary for the ratification of Treaties that amend the Union structure. Such majority is the same as that necessary for the constitutional revision of the Basic Law.30 ii. Interregional and Joint National–Regional Bodies Another mechanism for the involvement of the regional authorities in European policy-making entails institutions that foster co-operation and co-ordination either horizontally, between the regions, or vertically, between the regions and the ‘centre’. Those bodies should not be understood as an alternative to the aforementioned constitutionally enshrined bodies that represent the regional tier at the federal level. In certain cases, they may even co-exist. Prime examples of co-ordination bodies that co-exist with an upper chamber are the Conference of Länder Ministers of European Affairs and the Heads of the State and Senate Chancelleries, and the Conference of Minister-Presidents in Germany.31 Other examples are the Conference of Integration (Integrationskonferenz) of the Austrian Länder (IKL) and the non-institutionalised but very influential Conference of the Presidents of the Länder (Landeshauptmännerkonferenz). The former is composed of the President of each Land and of the Presidents of the regional Assemblies as well as the Presidents of the two houses of the federal Parliament, while the latter comprises only the Presidents of the Länder. As already mentioned, pursuant to Article 23d(1) of the Austrian Constitution, the Government must directly inform the Länder. If the State receives a ‘uniform comment’ from the Länder—through one of the aforementioned bodies—on some Union legislative proposal within Land competence, it is bound to respect that opinion during negotiations and voting at the EU level.32 The Government may deviate from this unitary position ‘only for compelling foreign and integration policy reasons’.33 In that event, the reasons should immediately be communicated to the Länder. Where the EU subject matter lies outside the legislative powers of the Länder but touches on their interests, the federation must take into account the written opinion of the Länder. This obligation does not stem from the Constitution but from a constitutional agreement between the federation and the Länder according to Article 23d(4).34 While the long-standing federal culture of Austria has dictated those constitutionally enshrined obligations to inform and consult the regional tier, the

29 30 31 32 33 34

Ibid. German Basic Law, Art 79. European University Institute, above n 24, 148. Austrian Constitution, Art 23d(2). Ibid. BGB1 775/1992.

The Role of Sub-State Entities in the EU Decision-Making Processes 217 contentious political system of Belgium has led to the establishment of a really inclusive co-ordination procedure. Such procedure is provided for in the 1994 Co-operation Agreement between the federal Government and the sub-State entities.35 Generally speaking, it is the Directorate for European Affairs in the Foreign Ministry which has responsibility for co-ordinating the Belgian positions within the EU. In order to achieve this, it regularly convenes a Co-ordination Committee on European Affairs. Every decision on the Belgian position is reached in the Directorate General by representatives of the federal prime minister and deputy prime ministers, of the minister-presidents of the different sub-State entities and of those ministers who are responsible for the subjects on the agenda. It is important to stress that all the decisions have to be reached by consensus, especially those that touch on regional or community competences. If consensus is not achieved, the matter may be referred to the Inter-Ministerial Conference for Foreign Policy and thence to the Concertation Committee. If agreement is not reached even in that phase, customary practice is that the Belgian representative abstains in the Council. However owing to the Belgian tradition of consensus and to the fact that the Belgian influence in the Council deliberations would otherwise be completely lost, a common Belgian position is regularly reached.36 Joint national–regional bodies like those in Belgium also play an essential role in the participation of the Italian and the Spanish regions in national decision-making processes. Starting with Italy, pursuant to Article 2 of Law No 11/2005, the InterMinisterial Committee for Common European Affairs (Comitato Interministeriale per gli affari comunitari europei, hereafter ‘CIACE’) was set up.37 When the CIACE discusses EU legislation of regional interest, the president of the CRPA, or somebody delegated by him or her, can ask to participate in the meetings.38 Where Union legislation concerns competences of the regions and the autonomous provinces, Article 5 of the same law provides for a sophisticated and inclusive procedure. Upon receipt of the relevant draft legislation, the CRPA and the Conference of the Presidents of the Assemblies of Regional Councils and of Autonomous Provinces forwards it to the presidents of the regional executive committees and of the regional councils. They have 20 days to submit their comments to the Government. If the legislation is of particular importance for the regions and the autonomous provinces, or if one or more of the regions or the autonomous provinces so requests, the Government will convene the Permanent 35 Samenwerkingsakkoord van 8 maart 1994 tussen de Federale Staat, de Gemeenschappen en de Gewesten en het Verenigd College van de Gemeenschappelijke Gemeenschapscommissie, met betrekking tot de vertegenwoordiging van het Koninkrijk België in de ministerraad van de Europese Unie (Belgisch Staatsblad, 17 November 1994). 36 European University Institute, above n 24, 64–66; LR Sciumbata, ‘Belgium’ in Istituto di Studi sui Sistemi Regionali Federali e sulle Autonomie (ed), Procedures for Local and Regional Authority Participation in European Policy Making in the Member States (Luxembourg, Committee of the Regions, 2005) 117–18. 37 P Bilancia, F Palermo and O Porchia, ‘The European Fitness of Italian Regions’ (2010) 2 Perspectives on Federalism E-1, E-122. 38 Law No 11/2005, Art 2(2).

218

Nikos Skoutaris

Conference for the Relations between the State, the Regions and the Autonomous Communities (Conferenza permanente per i raporti tra lo stato, le regioni e le province autonome) with a view to reaching a common position within 20 days. After this period of time lapses, or in cases of urgency, the Government may proceed. If the Permanent Conference so requests, the Government will lodge a ‘reservation of examination’ (riserva di esame) in the Council of the EU.39 In Spain, on the other hand, EU matters within the respective policy fields are handled by the Sectoral Conferences. In 1992, the Sectoral Conference for Union Affairs (CARCE) was set up with top officials from the State and the Autonomous Communities. Five years later, it was formally institutionalised by virtue of Law 2/97. It now acts not only as a forum for the exchange of information and the implementation of Union policies, but also for the participation of the Autonomous Communities in the preparation of the Spanish position in European decisionmaking. More analytically, as regards shared competences, the central Government tries to reach a common position with the Autonomous Communities although it retains the final say. With regard to exclusive regional powers, if the Autonomous Communities reach a common position, the State has to defend that position at the EU level.40 Similar arrangements also exist in two decentralised Member States, namely, France and The Netherlands. In France, this is a rather recent development. During the first National Conference of Executives (Conférence nationale des exécutifs), created in October 2007, it was decided that the General Secretariat for European Affairs (Secrétariat général aux affaires européennes, hereafter ‘SGAE’) would establish a closer relationship with the Association of French Regions (Association des Régions de France) and the Assembly of French Departments (Assemblée des Départements de France) among others. The SGAE meets regularly with the delegates of the territorial associations responsible for European affairs, to discuss the French position within the EU.41 In The Netherlands, the highly decentralised policy-making framework foresees a role for the Association of The Netherlands Municipalities (Vereniging van Nederlandse Gemeenten) and the Association of the Provinces of The Netherlands (Interprovinciaal Overleg). Those two bodies are represented in many areas in interdepartmental working groups of the national Government, preparing the Dutch position for the Council of Ministers. There is also a monthly meeting with the Ministries of the Interior and Foreign Affairs, and on regular base there are meetings on a political level with the Minister for European Affairs.42

39

Law No 11/2005, Art 5(5). J Bengoetxea, ‘The Participation of Infra-State Entities in European Union Affairs in Spain: the Basque Case’ in S Weatherill and U Bernitz (eds), The Role of Regions and Sub-National Actors in Europe (Oxford, Hart Publishing 2005) 55; A Ross and M Salvador Crespo, ‘The Effect of Devolution on Implementation of EC law in Spain and the UK’ (2003) 28 EL Rev 210, 226. 41 European University Institute, above n 24, 132–33. 42 Ibid 229. 40

The Role of Sub-State Entities in the EU Decision-Making Processes 219 Lastly, within the UK framework of devolution, the participatory rights of the regional tier in the domestic EU policy-shaping process are guaranteed by soft non-binding law, as provided for by the aforementioned Memorandum of Understanding and the Concordats. Those agreements between Whitehall and the devolved administrations envisage the full involvement of the devolved regional authorities in the formulation of the UK position.43 In general, the UK negotiating position is discussed at the Joint Ministerial Committee on Europe. Ministers and officials from the three devolved administrations are part of the UK team, with the UK minister determining the final position and retaining overall responsibility.44

III. PARTICIPATION IN THE UNION DECISION-MAKING PROCESSES

It is not only the national constitutional frameworks that provide the sub-State entities with access to the Union policy-shaping process. The EU Treaties themselves also contain provisions that allow for the representation of regional interests at the Union level. The participation of regional ministers in the Council and the role of the regional tier in the application of the subsidiarity principle provide the most tangible examples of this. However, again, the fact that not all regions are able to benefit from these arrangements reminds us of the existence of an asymmetrical ‘Europe with some regions’.

A. Participation of the Regions in the Union Institutions i. The Council One of the first steps the EU made towards responding to the gradual regionalisation process being undertaken by many EU Member States, was the opening-up of the Council of Ministers to representatives from sub-State entities. Indeed, the Maastricht Treaty amended the then Article 146 TEC, dropping the reference to national governments. The new wording allowed Member States to be represented in Council sessions by members of regional authorities. It is difficult to overemphasise the constitutional significance of this amendment that has survived all subsequent Treaty modifications. In essence, it questions—at least partially— the ‘federal blindness’ thesis. However, it is interesting to note that the political science literature is divided on the usefulness of such a provision for regions to represent their Union interests.45 Recently, it has been argued that such a tool may

43 See Memorandum of Understanding, paras 17–20; Concordats on Co-ordination of European Union Policy Issues, above n 20, paras B1.2, B1.4, B2.2, B2.4, B3.2, B3.4, B4.5 to B4.11. 44 Ibid. 45 Tatham, above n 12, 59.

220

Nikos Skoutaris

allow regions to represent distinctive interests at a crucial stage in the EU policy process.46 Be that as it may, Article 16(2) TEU provides: The Council shall consist of a representative of each Member State at ministerial level, who may commit the government of the Member State in question and cast its vote.

It is not prescribed to which internal level of the government that representative must belong. Thus, even ministers of regional governments are allowed to represent their Member States if the internal constitution so provides. In addition, pursuant to Article 5(3) of the Council’s rules of procedure, ‘officials who assist them’ may accompany the members of the Council.47 There is no legal requirement that the official originate from the same government as the representative. Hence, it is possible to have mixed delegations of federal and regional ministers. However, in the composite Union constitutional order, the national constitutional framework must be examined in order to appreciate the importance of this provision. Indeed, it seems that only a small number of regional authorities are able to benefit from this arrangement. The ‘usual suspects’ comprise the sub-State entities of the three federations, the Italian regions and autonomous provinces of Trento and Bolzano, the Spanish Comunidades Autónomas and the UK devolved governments. Starting with Austria, the relevant provisions may be found in Article 23d of the Constitution. According to this provision, if the EU subject matter concerns the legislative powers of the Länder, there are two options. As has been mentioned in the previous section, the first option entails Austria being represented in the Council by a federal minister who is bound by the opinion of the Länder. In fact he or she ‘may deviate therefrom only for compelling foreign and integration policy reasons’.48 Paragraph 3 of the same Article, however, offers the federal Government a second option, to authorise a representative from the Länder to be present in the Council on Austria’s behalf. This representative is bound by the common position of the Länder as expressed in a decision by the 10 Länder prime ministers (Landeshaptleutekonferenz). In the Council meeting, he or she has to consult the competent federal minister who sends an associate to the representative into the Council meeting.49 With regard to Germany, representation in the Council depends on the issue at stake. Article 23(6) of the German Basic Law provides that if the EU subject matter at issue predominantly lies within the legislative powers of the Länder, a minister appointed by the Bundesrat may represent Germany.50 This minister usually 46 M Tatham, ‘Going Solo: Direct Regional Representation in the European Union’ (2008) 18 Regional and Federal Studies 493, 499–502. 47 Annex to Council Decision 2009/937/EU adopting the Council’s Rules of Procedure [2009] OJ L325/35, art 5(3). 48 Austrian Constitution, Art 23d(2). 49 Austrian Constitution, Art 23d(3). 50 German Basic Law, Art 23(6).

The Role of Sub-State Entities in the EU Decision-Making Processes 221 has the mandate for a certain time (one to three years).51 In practice, Germany’s representation by a regional minister designated by the Bundesrat is exceptional. In fact, under the federal reform of 2006, such ministers’ exclusive right to speak for Germany is now restricted to education, culture and broadcasting.52 The Belgian sub-State entities may also represent the federation in the Council. The framework here, however, is more sophisticated and finely tuned than those in Germany and Austria. Following the constitutional reforms of the early 1990s, a Co-operation Agreement was drawn up in 8 March 1994 between the federal Government and the Regions and the Communities.53 The Agreement lays down the procedure for representation and the co-ordination of the Belgian position in the Council, and is based on three principles: consensus, mixed delegation and rotation. It was amended in 2003 following the regionalisation of agriculture and fisheries.54 As far as representation of commonly-agreed positions is concerned, the 1994 Agreement distinguished between four categories: Category I concerns all Council topics which relate to the exclusive federal competences; Category II deals with issues the major part of which concern federal subject matter, while Category III covers those matters of which the major part is of interest to the sub-State entities; and Category IV includes Council topics that touch exclusively on the competences of the sub-State entities. In Category I, Belgium is represented by the federal Government, in Category IV by a representative from the sub-State entities. In the latter case, the sub-State entities decide together who will represent them. In Categories II and III, a system of ‘assistance’ applies. The delegation is headed by a member of the government which has a dominant share, with an assistant being a member of the government which has the non-dominant share. The head of the delegation votes, whereas the ‘assistant’ politically controls his behaviour and has the right to speak. The 2003 Co-operation Agreement added two more categories. Category V concerns Council configurations that touch upon the competences of one regional government. In fact this category refers only to the competence of Flanders with regard to fisheries. Unsurprisingly, in that case, the Flemish government represents Belgium. Lastly, Category VI refers to exclusive regional competences but with the federal Government taking the lead. This applies only to agriculture.55 As was mentioned above, apart from the regional tier within the three named federations, the sub-State entities of the three regionalised States (ie, Italy, Spain

51 Further details may be found in the Law of 12 March 1993 on co-operation of the Federation and the Federated State in EU affairs (BGB1, 1993 I, 313). 52 European University Institute, above n 24, 148. 53 Above n 35. 54 Samenwerkingsakkoord van 13 februari 2003 tussen de Federale Staat, de Gemeenschappen en de Gewesten tot wijziging van het samenwerkingsakkoord van 8 maart 1994 tussen de Federale Staat, de Gemeenschappen en de Gewesten met betrekking tot de vertegenwoordiging van het Koninkrijk België in de ministerraad van de Europese Unie (Belgisch Staatsblad, 25 February 2003). 55 Sciumbata, above n 36, 113–15.

222

Nikos Skoutaris

and the UK) have also benefited from this arrangement. Under Article 5 of Law No 131/2003, Italian regions may participate in the work of the Council of the EU and its working groups, and may work with the Commission and its expert committees in areas of regional legislative competence, following agreement in the Conferenza Stato-Regioni.56 Moreover, in March 2006, the central Government and the sub-State entities signed an agreement which provides, inter alia, that Italy may be represented by a regional official in the Council. However, that will happen only if an agreement is reached within the framework of the Conferenza Stato-Regioni.57 In Spain, the culmination of efforts begun in the 1990s resulted in an agreement on 9 December 2004. The Agreement allows for the participation of the Autonomous Communities in the Council in four configurations: Employment, Social Policy, Health and Consumer Affairs; Agriculture and Fisheries; Environment; and Education, Youth and Culture.58 According to this Agreement concluded by the CARCE, the relevant sectoral conference designates one Autonomous Community to represent all of them in the coming period. This Autonomous Community then seeks the agreement of the others and of the Spanish delegation on their common position and attends the Council.59 Lastly, ministers from the three devolved administrations (Scotland, Wales and Northern Ireland) are allowed to attend the Council after agreement with the UK Government.60 It is the lead UK minister, however, who decides on the composition of the UK team, taking into account that the devolved administrations should have a role to play ‘in meetings of the Council of Ministers at which substantive discussion is expected of matters likely to have a significant impact on their’ competences.61 It is also the head of the delegation who bears the responsibility for the negotiations and agrees to ministers from the devolved administrations speaking for the UK.62 The Concordat clarifies that ‘they would do so with the full weight of the UK behind them’, because the positions to be taken within the Council would have been agreed in advance at the relevant Joint Ministerial Committee.63 ii. The European Parliament Despite its importance for the democratic life of the Union, the academic literature has largely overlooked the role of the European Parliament as a channel for 56

Bilancia et al, above n 37, E-142. Ibid. 58 A D’Atena, ‘Participation of Regional and Local Authorities in the Preparatory Phase of European Policy Making—European Side’ in Istituto di Studi sui Sistemi Regionali Federali e sulle Autonomie (ed), Procedures for Local and Regional Authority Participation in European Policy Making in the Member States (Luxembourg, Committee of the Regions, 2005) 17. 59 European University Institute, above n 24, 286. 60 Concordats on Co-ordination of European Union Policy Issues, above n 20, para B4.14. 61 Ibid para B4.13. 62 Ibid para B4.14. 63 Ibid. 57

The Role of Sub-State Entities in the EU Decision-Making Processes 223 regional representation in the EU political structure.64 The reason for that might be found in the fact that the constituency to elect MEPs in the vast majority of the Member States is a single State constituency. It is only in Belgium, France, Italy, Ireland and the UK that the MEPs are elected on the basis of regional constituencies. In those cases, however, it might be argued that the regional tier indirectly participates in the political life of the Union.65 iii. The Court of Justice It is difficult to overstate the significance of the ‘least dangerous branch’ in the European integration process. In that regard, it would be important to examine not only the conditions under which the regional tier has access to the Court of Justice, but also whether there are constitutional provisions enabling the sub-State entities to request their Member State to appeal to the Luxembourg Court against EU decisions through the annulment procedure set out in Article 263 TFEU. However, given that such tools are available to the regions after an EU decision has been adopted, such an analysis goes beyond the scope of the present chapter. For the purposes of this contribution, it suffices to note that the Court might offer another forum for representation of the interests of the regional tier. iv. The Committee of the Regions Established in 1994, the Committee of the Regions is an EU advisory body. On a proposal from the Commission, the Council unanimously determines the composition of this political assembly, whose members may not number more than 350. However, it is the Member States themselves that assign their representatives to the Committee. The only sufficient and necessary condition that the Treaties provide is that the members of that body should be ‘representatives of regional and local bodies who either hold a regional or local authority electoral mandate or are politically accountable to an elected assembly’.66 This has allowed the States to adopt very different approaches as to the rules concerning their representation. There are, to begin with, formal differences. While in Austria there is a constitutionally-enshrined rule concerning the representation of the federation on the Committee of the Regions,67 in Belgium, Germany and Ireland the rules consist of legislation, and in Italy of regulations. In Spain and Portugal, on the other hand, the members of the delegation are appointed by means of non-legislative parliamentary resolutions.68 Also with regard to the level of administration actually

64 For an account of the role of the European Parliament as a channel of regional interest participation, see generally, Tatham, above n 46, 504–06. 65 Ibid. 66 Art 300(3) TFEU. 67 Austrian Constitution, Art 23c(1). 68 D’Atena, above n 58, 20–21.

224

Nikos Skoutaris

representing the Member States, the diversity cannot be over-emphasised. More specifically, delegations from federal or regionalised Member States (eg, the three federations, Spain, Italy) are predominantly regional, while in non-regionalised Member States such as The Netherlands, Sweden, Denmark, Luxembourg and Ireland, the representatives come exclusively from local authorities.69 Be that as it may, it seems that the Committee of the Regions provides for a forum through which the sub-State entities may exert influence in the EU decisionmaking processes. The obvious question to be posed is in which policy areas this advisory body consults the European Parliament, the Council and the Commission. The answer seems to be: when the Treaties so provide and in all other cases where one of those institutions considers it appropriate.70 Generally speaking, the Treaties provide for the consultation of the Committee in the areas of transport,71 employment policy,72 social policy,73 the European Social Fund,74 education and youth,75 vocational training,76 culture,77 trans-European public health,78 infrastructure networks,79 economic and social cohesion,80 the environment81 and energy.82 However, the Committee of the Regions can influence the shaping of the EU constitutional order by other means as well. According to the Lisbon Treaty, it may also bring annulment procedures before the Court of Justice ‘for the purpose of protecting its prerogatives’.83 This right of direct access to the Court is further elaborated in the Subsidiarity Protocol. Article 8 provides that the Committee may bring ‘actions against legislative acts for the adoption of which the [TFEU] provides that it be consulted’. It remains to be seen when this institution will exercise such right. v. The Regional Representation and Liaison Offices To complete the picture of the representation of regional interests in the EU decision-making processes, we should briefly refer to the regional representation and liaison offices in Brussels. It is important to mention them because they play a crucial role in disseminating and exchanging information on EU policy issues,

69 70 71 72 73 74 75 76 77 78 79 80 81 82 83

Ibid. Art 307(1) TFEU. Arts 90–100 TFEU. Arts 145–150 TFEU. Arts 151–161 TFEU. Arts 162–164 TFEU. Art 165 TFEU. Art 166 TFEU. Art 167 TFEU. Art 168 TFEU. Arts 170–172 TFEU. Arts 174–178 TFEU. Arts 191–193 TFEU. Art 194 TFEU. Art 263(3) TFEU.

The Role of Sub-State Entities in the EU Decision-Making Processes 225 and they are considered to be proof of the Europeanisation of regions and the emergence of a third level in the EU arena.84 As a starting point, it should be noted that these offices have mushroomed since the first ones were set up in the mid-1980s. At present, there are over 250.85 They vary both in terms of the authorities they represent and as regards the legal basis for their establishment.86 As for the first criterion, while some offices belong to a single regional authority, others represent an association of regional governments or even cross-border regions.87 Concerning the legal basis, it suffices to note that some offices are set up by law, others are governed by public law as public bodies and still others are run privately as associations. It seems that the national legal frameworks have progressively become more lenient as regards their existence. A good example of this point is the fact that the Spanish Government had challenged before the Constitutional Court the right of the Basque Country to have a delegation in Brussels, ‘alleging that there could be no relation whatsoever between the Basque public institutions and the European institutions’.88 However, the Court rejected that argument. It held that Union law is internal law and affects the competences of the Autonomous Communities.89

B. The Role of the Regional Authorities in the Application of the Subsidiarity Principle The subsidiarity principle is recognised as a fundamental principle of the Union constitutional order not only by the Treaties,90 but also by the constitutions of Member States, such as the German91 and the Portuguese Constitutions.92 At the

84 See, eg J Magone, ‘The Third Level of European Integration: New and Old Insights’ in J Magone (ed), Regional Institutions and Governance in the European Union (Westport, Conn, Praeger Publishers, 2003) 11. 85 European University Institute, above n 24, 41. 86 D’Atena, above n 58, 40–41. 87 European University Institute, above n 24, 42. 88 Bengoetxea, above n 40, 54. 89 Spanish Constitutional Tribunal, sentencia 165/1994, 26 May 1994. See also M Perez Gonzalez, ‘La onda regional en Bruselas y el ámbito del poder exterior (Comentario a la sentencia del Tribunal Constitucional 165/1994, de 26 de Mayo)’ (1994) 21 Revista de Instituciones Europeas 94. 90 See generally Art 5(3) TEU and Protocol No 2. 91 Art 23(1) of the German Basic Law reads, inter alia: ‘With a view to establishing a united Europe, the Federal Republic of Germany shall participate in the development of the European Union that is committed to democratic, social, and federal principles, to the rule of law, and to the principle of subsidiarity, and that guarantees a level of protection of basic rights essentially comparable to that afforded by this Basic Law.’ 92 Art 7(6) of the Portuguese Constitution reads: ‘Subject to reciprocity and to respect for the fundamental principles of a democratic state based on the rule of law and for the principle of subsidiarity, and with a view to the achievement of the economic, social and territorial cohesion of an area of freedom, security and justice and the definition and implementation of a common external, security and defence policy, Portugal may enter into agreements for the exercise jointly, in co-operation or by the Union’s institutions, of the powers needed to construct and deepen the European Union.’

226

Nikos Skoutaris

same time, it is a guiding principle in the relationship between certain Member States and their respective regional authorities, as is, for example, the case in Italy.93 Arguably, one of the major breakthroughs of the Lisbon Treaty was the reinforcement of this principle by the explicit recognition of its regional dimension and by the formal introduction of the Early Warning System (hereafter ‘EWS’). With regard to the regional dimension, it should be noted that Article 5(3) TEU, which provides for the new formulation of the principle of subsidiarity, states that the Union—outside the areas of its exclusive competence—may only act in so far as the objectives of the proposed action ‘cannot be sufficiently achieved by the Member States, either at central level or at regional and local level’.94 In addition, Article 2 of Protocol No 2 provides for the duty of the Commission to consult widely before proposing legislative acts. Such consultations must, ‘where appropriate, take into account the regional and local dimension of the action envisaged’.95 However, the regional dimension is also linked with the functioning of the EWS. This mechanism gives the right to ‘any national Parliament or any chamber of a national Parliament’ to get involved in the EU legislative process.96 More precisely, the national parliaments are entitled to send a reasoned opinion to the Union institutions containing their objections to the draft legislation if they believe that the draft is in breach of the subsidiarity principle.97 If one-third (one-quarter if the issue touches upon the area of freedom, security and justice) of all the votes allocated in the national parliaments oppose the subsidiarity arguments, the legislative initiator may decide to maintain, amend or withdraw the draft. In any case, reasons must be provided for this decision.98 If, on the other hand, at least a simple majority of all the votes allocated in the national parliaments oppose such draft legislation on grounds of subsidiarity, the proposal must be reviewed. The Commission, being the legislative initiator under the ordinary legislative procedure, may decide to maintain, amend or withdraw the draft. If it decides to maintain its proposal, it has to provide a reasoned opinion why it considers that it is in compliance with the subsidiarity principle. On the basis of the reasoned opinions of the Commission and the national parliaments, the Council (by a majority of at least 55 per cent of its members) or the European Parliament (by a simple majority of the votes cast) shall decide whether the proposal shall be given further consideration.99 Be that as it may, the explicit reference to ‘any chamber of a national Parliament’ is particularly important for the participation of the regional tier in the Union pre-legislative phase in those Member States where chambers composed of

93 Italian Constitution, Art 118. See also Giuseppe Martinico’s contribution in ch 15 of this volume. 94 Emphasis added. 95 Protocol No 2, Art 2. 96 Protocol No 2, Art 6. 97 Ibid. National parliaments must do so within eight weeks from the date of transmission of a draft legislative act. 98 Protocol No 2, Art 7(2). 99 Protocol No 2, Art 7(3).

The Role of Sub-State Entities in the EU Decision-Making Processes 227 regional representatives exist, such as Austria and Germany. Moreover, Article 6 of Protocol No 2 allows national parliaments to consult regional parliaments with legislative powers, to solicit their opinions on whether a certain Union draft legislative proposal complies with the subsidiarity principle. Although the positions of the regional legislatures—with the exception of Belgium—will not bind the national parliaments, the inclusive nature of the procedure recognises the regional dimension of the subsidiarity principle. More specifically, the Austrian Constitution, following the amendment act Lissabon-Begleitnovelle adopted by the Austrian Parliament on 8 July 2010, acknowledges the rights of the Bundesrat to conduct subsidiarity scrutiny100 and the duty of co-operation between the upper chamber and the regions.101 In order to fulfil its role with regard to subsidiarity scrutiny, the Bundesrat has established a specialised EU committee with such a mandate. The committee takes a decision by simple majority. However, if half of the representatives of at least three regions so demand, the committee will delegate the procedure to the plenary assembly.102 In that sense, the Austrian upper chamber plays a significant role in the participation of the regional tier in the EWS. The regional parliaments, by contrast, play a much more modest role. Although every EU proposal is automatically forwarded to all of them,103 the regional parliaments’ opinions do not bind the Bundesrat.104 As far as the German Bundesrat is concerned, it should be noted that its president uses his or her discretion to decide whether a certain Union legislative proposal should undergo subsidiarity scrutiny, following a request from another Bundesrat member or from a Land.105 In that event, the president distributes the proposal to the relevant sectoral committees, including the EU committee. The committee in question presents its report to the plenary assembly, together with a recommendation for a resolution.106 The report may be adopted by tacit assent, or in a formal vote by simple majority.107 Given that the Länder vote en bloc, it is impossible for individual regions to split the vote. Moreover, the opinions of dissenting regions are not considered.108 At the same time, Belgium clarified, in a declaration annexed to the Treaty of Lisbon, that in accordance with its constitutional law, not only the Chamber of Representatives and Senate of the Federal Parliament but also the parliamentary assemblies of the

100

Austrian Constitution, Art 23f(1). Austrian Constitution, Art 23g(3). 102 G Vara Arribas and D Bourdin, The Role of Regional Parliaments in the Process of Subsidiarity Analysis within the Early Warning System of the Lisbon Treaty (Luxembourg, Committee of the Regions, 2011) 10. 103 Austrian Constitution, Art 23g(3). 104 Vara Arribas and Bourdin, above n 102, 19. 105 Ibid 45. 106 Ibid 45–46. 107 Ibid. 108 Ibid 57. 101

228

Nikos Skoutaris

Communities and the Regions act … as components of the national parliamentary system or chambers of the national Parliament.109

This does not mean, however, that consensus must be reached in order for the seven Belgian parliaments to be able to provide for reasoned opinions and to make use of the subsidiarity vote.110 Neither at the federal level nor at the subState level is consensus required. As soon as either the House of Representatives or the Senate considers a certain EU legislative proposal to be in breach of the subsidiarity principle, at least one subsidiarity vote is cast. The same happens when one of the competent sub-State legislatures is of such opinion.111 Thus, in cases where the proposal involves exclusive federal prerogatives, two votes are cast when both chambers issue a reasoned opinion. In cases where the proposal affects both federal and sub-State powers, two votes are cast when at least one federal chamber and one competent sub-State legislature issue a reasoned opinion. For proposals that relate solely to prerogatives of the Regions and/or Communities, two votes are cast when two legislatures belonging to different linguistic segments issue a reasoned opinion. Lastly, in the case of proposals on fisheries, touching upon the legislative competences of Flanders only, its parliament may cast two votes.112 While the constitutional orders of the federal Member States clearly provide for procedures that allow the regional tier to participate in the scrutiny of the subsidiarity of EU legislation, the channels of participation of the sub-State level in the regionalised Member States are less entrenched and more informal. In Italy, for instance, there is no legal duty to involve the regional parliaments in the EWS.113 However, the regions may influence the procedure informally, through the aforementioned Conference of the Presidents of the Assemblies of Regional Councils and of Autonomous Provinces. In Spain, the 17 regional parliaments have four weeks to send a reasoned opinion to the national Parliament. However, only if the national Parliament itself decides to approve of a reasoned opinion will the contribution of the sub-State legislatures be mentioned and accompanied with references to the relevant documents.114 Lastly, in the UK, both chambers at Westminster consider the views of the devolved legislatures to be part of the normal process of scrutiny in the relevant EU Committees.115

109

Declaration by the Kingdom of Belgium on National Parliaments [2008] OJ C115/355. See Ontwerp van samenwerkingsakkoord tussen de Federale Wetgevende Kamers, de parlementen van de Gemeenschappen en de parlementen van de Gewesten ter uitvoering van het Protocol betreffende de toepassing van de beginselen van subsidiariteit en evenredigheid gehecht aan het Verdrag tot vaststelling van een Grondwet voor Europa and Vlaams Parlement, Gedachtenwisseling over de stand van zaken aangaande het intra-Belgische samenwerkingsakkoord noodzakelijk voor de operationalisering van een aantal bepalingen van het verdrag van Lissabon. See also Piet Van Nuffel’s contribution in ch 8 of this volume. 111 Vara Arribas and Bourdin, above n 102, 35. 112 Ibid 37. 113 Ibid 72. 114 Ibid 77–78. 115 Ibid 132–33. 110

The Role of Sub-State Entities in the EU Decision-Making Processes 229 IV. CONCLUSION

In this chapter, I have reviewed EU law and national constitutional law in order to assess the participatory rights of the sub-State entities in the EU decisionmaking processes. Such exercise is necessary given the composite nature of EU constitutional law. And because the EU is an intertwined constitutional order, the diversity of the arrangements that allow the regional tiers to take part in the EU policy-making process cannot be overstated. However, it is quite clear that the autonomous regions with legislative powers have participatory rights—albeit modest—within the Union political structure. Such rights might be constitutionally enshrined, provided for by law or simply a result of a non-legislative political agreement. Clearly, though, the number of sub-State entities that can take advantage of such arrangements is rather limited. We are far from achieving the enthusiastic vision of a ‘Europe of the Regions’, perhaps because most EU Member States do not have regions with legislative powers. Be that as it may, assessing the role of the autonomous authorities in Europe is still important, not least because ‘the regional question, like the poor, is always with us and has been on the political agenda since the formation of the nation-state’.116

116 J Loughlin, ‘The Regional Question, Subsidiarity and the Future of Europe’ in S Weatherill and U Bernitz (eds), The Role of Regions and Sub-National Actors in Europe (Oxford, Hart Publishing, 2005) 157.

10 Autonomous Constitutional Regions in a Federal Europe JOXERRAMON BENGOETXEA*

I. INTRODUCTION

D

OES THE CURRENT EU constitutional set-up correctly represent a special type of sub-State entity, ie regions that:

(a) are constitutionally recognised in the Member States; (b) enjoy a high level of political autonomy; and (c) have autonomous legislatures, executives and even, in some instances, an autonomous administration of justice?

The question is, obviously, loaded with assumptions and evaluative premises. The factual assumption made by the question is that such special regions might already be somehow represented institutionally but perhaps inadequately so. There is also a normative question: Why should sub-State entities be institutionally represented at all? And, also from a normative point of view, we might ask about the extent of an adequate representation. Traditionally, the theory has been that such constitutional regions are purely internal phenomena, of little interest to the European institutions, and that the European Union (EU), if Union it be, is a constitutional order of States.1 Yet reality is often stubborn, and the theory has failed to capture interesting and meaningful changes concerning the status and the role of such regions in the process of European integration. Many of the regional, devolved powers are nowadays shared, mixed or confused with the powers of the EU institutions. In order to exert their powers in law- and policy-making at the EU level as would

*

University of the Basque Country, UPV/EHU. See A Arnull, C Barnard, M Dougan and E Spaventa (eds), A Constitutional Order of States. Essays in EU Law in Honour of Alan Dashwood (Oxford, Hart Publishing, 2011). 1

Autonomous Constitutional Regions in a Federal Europe 231 befit their status and powers, these special, ‘constitutional’ regions have to seek influence: (a) through governance networks, such as the the Conference of European regions with legislative power (REGLEG), and direct communication channels with the EU; or (b) through their Member State representations in the EU, according to their internal mechanisms; or (c) through sharing existing platforms available for sub-State entities in the EU, such as the Committee of the Regions, together with other sub-State entities that have not acquired the same degree of self-government as constitutional regions. Resistance to recognition of such a special status has been detected at all levels and, above all, in the claim that the EU cannot look into the internal organisation of its Member States if it is to respect the principle of institutional autonomy. Historically several attempts have been made to achieve recognition of the special status of the constitutional regions and to make theory fit reality. The Lamassoure proposal for the recognition of a special status for ‘associated regions’, presented during the discussions of the Convention on the Future of Europe, was one such institutional attempt.2 However, it failed and was replaced by the Napolitano report, which makes no distinction as to types of region.3 It seems that the recognition of a special status might be a gradual, matter-of-fact process, rather than a clear institutional recognition brought along by a given constitutional reform. Some European regions are gradually acquiring an autonomous or selfgoverning constitutional status in EU law. This obviously takes the traditional understanding of EU regionalism one step further. This traditional understanding is analysed in section II. of this chapter from an analytical perspective—the project—and in section III. from an historical perspective—the process. Section IV. sets out the paradox of constitutional regions in the EU: on the one hand, they are seen as crucial actors in the EU; but on the other hand, they are not fully recognised as a special type of region. They are represented at the Committee of the Regions alongside city councils and other types of (purely administrative) regions with no constitutional status, and very few powers. One possible solution is suggested in section V., and leads to claiming a special status for ‘autonomous constitutional regions’. There is sufficient support for such a status in the recent

2 Committee on Constitutional Affairs of the European Parliament, ‘Report on the division of competences between the European Union and the Member States’ (‘Lamassoure report’), 2001/2024(INI), 24 April 2002. 3 Committee on Constitutional Affairs of the European Parliament, ‘Report on the role of regional and local authorities in European integration’ (‘Napolitano report’), 2002/141(INI), 4 December 2002.

232

Joxerramon Bengoetxea

case law of the European Court of Justice (ECJ) (Azores,4 Flemish social policy case,5 Basque taxation case,6 Horvath,7 Gibraltar8) and in the EU Treaties after Lisbon. Lastly, section VI. discusses the possible institutional consequences such a status would entail.

II. THE REGIONAL SUB-STATE QUESTION IN THE EU

A. Diversity and Complexity There is enormous diversity of regional formulae and sub-State self-government typologies in the EU. The number of sub-State entities—territorial institutional layers or levels below the Member State level—adds up to 100,000 in the EU. The very concept of ‘region’ does not always refer to the same reality. Leaving aside geographic, geopolitical or ecologic concepts of ‘region’, we can easily observe that economic (administrative or statistically defined) regions are not always co-extensive with constitutional regions. Linguistic and cultural regions (nations or ethnicities) in some cases are larger than political regions. At a wider, global level, strategic or geopolitical regions9 and international organisations (the EU, the OECD) are not necessarily overlapping. Only Member States seem to offer a sufficient normative (and therefore also cognitive) stability when it comes to questions of definition, delimitation and borders. But Member States are very varied and diverse. Some Member States of the EU, large and small, are centralised, unitary States. Others, large and small, are decentralised but unitary States; others, large and small, are decentralised federal States; others are mixed, regional states, partly unitary and partly complex or quasi-federal. And obviously they are all in flux. To this constitutional diversity one might add the different self-conceptions of States. One finds Member States that are officially decentralised or complex, but nonetheless represented by State officials before other States. Moreover, they adopt negotiating positions in supraand international institutions as though they were unitary, taking the national government to represent the whole of the State and to be its only interlocutor, even if no consensus or agreed position has been explored within the State.10 The international community—international law—reinforces this oversimplification, 4

Case C-88/03 Portugal v Commission [2006] ECR I-7115. Case C-212/06 Government of the French Community and Walloon Government v Flemish Government [2008] ECR I-1683. 6 Joined Cases C-428/06 to C-434/06 UGT-Rioja and Others [2008] ECR I-6747. 7 Case C-428/07 Horvath [2009] ECR I-6355. 8 Joined Cases C-106/09 P and C-107/09 P, European Commission (C-106/09 P), Kingdom of Spain (C-107/09 P) v Government of Gibraltar and United Kingdom, Opinion of AG Jääskinen. 9 Europe is seen as a World region. Moreover, until recently, the West was a cultural and ideological project. Perhaps it still is, but it is difficult to see by reference to what alternative project it could be defined, since there are many. 10 Spain is a clear example, in my opinion, although there are some exceptions, obviously. 5

Autonomous Constitutional Regions in a Federal Europe 233 because complexity of interlocutors makes it difficult to systematise international relations. B. Cities and Other Sub-State Entities In some Member States there are no regions at all, only towns or local councils. On the other hand, some of the cities form ‘megalopoleis’ that are often much greater in population and economic weight than many regions. The diversity of phenomena—regions, Länder, autonomous communities, provinces, departments, cities, conurbations, counties, towns, etc—is truly striking and poses a real challenge to any attempt at finding formulae that are simple enough to be institutionally practicable to ensure some representation, but complex enough to capture the multiplicity of competences, structures and idiosyncrasies they represent. The concept of multi-level governance is perhaps a useful approximation, but again, governance goes beyond the purely public institutional dimensions, embracing civil society and non-State actors, which complicates matters further. C. Undeniable Relevance of Regions The regional or sub-State issue is a complex economic, social, cultural, political and institutional reality. Currently it is almost impossible to conceive of the European integration process as propelled exclusively by the Member State governments, although inter-governmentalism was the initial logic of European integration. The monopoly of Member State or national governments in running the Europe show has probably had its day. To begin with, State-national or State parliaments are progressively involved. In the future, their involvement in budgetary policies will only increase, and some parliaments, eg in the Nordic context, have been interested in discussing European integration for a long time. But regional governments and parliaments also claim—and progressively have—a say in European integration, and not only in relation to subsidiarity. New players and stakeholders at the regional level claim a voice in the process of integration too, according to the new governance culture.11 There are many factors (geographic, strategic, economic, social, political, constitutional) that make the regional dimension one of the key features of the EU: — Cohesion policy, and its territorial dimension brought in by the Lisbon Treaty, has been a catalyst for regionalism in the EU. — Sub-State nationalisms in different Member States make political claims for greater recognition—if not full-blown independence—within the EU. Even when secession from their Member States is their goal, many sub-State

11

Commission, ‘European Governance—A White Paper’ COM (2001) 428 final.

234

—

—

—

—

—

Joxerramon Bengoetxea nationalists actually accept the EU as the inevitable reference, thus contributing to its legitimacy and acceptability. It is quite striking that regional independence claims do not question the EU project. Euro-scepticism tends to be State-nationalist rather than regionalist. Federal Member States and their federal chambers (senates) make defensive claims for the protection of their exclusive powers, for participation in EU decision-making relating to these powers, and for the principle of subsidiarity to be respected and extended below the State level. Associations of regional governments (such as REGLEG) and regional parliaments (such as the Conference of European Regional Legislature Assemblies (CALRE)) claim greater involvement and, at the same time, greater protection against increasing EU powers, and the Commission has committed itself to consulting these associations.12 The EU institutions, especially the Commission, are interested in efficient or functional application and implementation of common policies and uniformity of EU law, and see the regions as key actors in this process. The Commission’s proposals for multi-level governance formulae like the European groupings of territorial co-operation13 and the tripartite contracts,14 should be seen in this light. Even as regards reserved Member State powers where the open method of co-ordination is used, the involvement of the regions is crucial, as has been continuously underlined by the Committee of the Regions regarding the Lisbon Strategy and the 2020 Strategy. This is even more so in areas like structural funds or the new 2020 strategy. Likewise, co-operation between different regions15 is one of the aims of trans-frontier co-operation in the EU (being an EU aim in itself), facilitating the creation of stronger ties and links between socio-economic actors in the different Member States. New visions and interpretations of democracy, justice and the common good emphasise the need to remain aware of the governance level at which decisions are or should be made, from the local to the global. In the current (2008–201?) context of severe financial and economic crisis in much of the EU, this issue is even more relevant. Locating solidarity at a higher level means that the lower level is reduced to an executive agency rather than a decisionmaking centre in matters of social welfare. At the same time, the financial and economic crisis has a direct impact on regional budgets, leading to restrictions that seriously curtail regional policy-making, policy implementation or investment in strategic infrastructure (eg, Trans-European networks),

12 Commission, ‘Dialogue with associations of regional and local authorities on the formulation of European Union policy’ (Communication) COM (2003) 811 final. 13 See European Parliament and Council Regulation (EC) 1082/2006 on a European grouping of territorial co-operation (EGTC) [2006] OJ L210/19. 14 Commission, ‘A framework for target-based tripartite contracts and agreements between the Community, the States and regional and local authorities’ (Communication) COM (2002) 709 final. 15 Ie, cross-border neighbouring regions or distant regions with shared agendas.

Autonomous Constitutional Regions in a Federal Europe 235 but might in the long run lead to enhanced co-operation between the layers involved in multi-level governance. As a result, it is now generally recognised that the EU must accommodate the regions’ interest in shaping and debating the process of European integration. This is the result of a difficult and patient evolution, as we shall now briefly examine.

III. THE EVOLUTION OF THE REGIONAL ISSUE: GAINING A VISION

A. Phase One: Regional Blindness At the beginning, the process of European integration neglected regions and sub-State actors; it neglected anything other than State-national governments. Sympathy for the regional phenomenon was to be found only in the other European organisation, the Council of Europe. The initial scenario was one of regional blindness. The Treaties and the European Communities (EC) largely ignored the regional issue. This was explained by reference to two important principles of international law also valid in Community law: the institutional autonomy of the State, and the principle of international legal personality of the State. i. Institutional or Constitutional Autonomy of the State According to the principle of institutional autonomy of the State, the State is free to organise itself as it pleases internally, as long as it honours its international obligations. Later, this was also seen as an expression of the principle of internal self-determination. In principle, for the international community, each State was free to organise as it decided, according to its own constitution. The international community could not interfere in order to impose any particular organisation of the State. Of course, States had to observe their international obligations, but how they chose to do so was the exclusive choice of the individual State. One specific corollary of this principle was that the EC system and the international community would not examine, supervise or control how States organised internally.16 In this sense, the EC system was as region-blind as international law. ii. State Personality The principle of State personality implied that it was the State, and only the State, as a whole, that had legal personality internationally. Any other entity enjoying such personality would be a creation of States through international treaties. 16

Case C-8/88 Germany v Commission [1990] ECR I-2321, para 13.

236

Joxerramon Bengoetxea

In the EC system, this meant that only Member States enjoyed personality, and anything below the Member State was simply a part of the State. The whole, the State, was always involved, even if the act affected only one or some of its parts. These principles clashed with reality in different ways. To begin with, as was mentioned earlier, the Council of Europe was soon aware of the need to represent the regions, and the Conference of Regional and Local Authorities was set up.17 Next, the very legitimacy-seeking discourse of the EU underlined the need for democracy, participation, proximity to the citizens, subsidiarity, multi-level governance, efficiency and functionality. But the EU did not limit itself to discourse only. It also developed policies regarding structural funds, cohesion, and overseas and peripheral regions.18 Furthermore, the Commission set up special directorates for the regions. In a parallel manner, the regions started to organise and create networks as experiments in para-diplomacy, and sometimes lobbies or pressure groups. iii. Internal Organisation Is No Excuse, No Defence The two principles discussed above had an interesting corollary or correlative principle in Community law: responsibility and liability under EC law was that of the State as a whole, even if the act giving rise to responsibility might be attributed or imputed to one or some of its identifiable parts. There was no possible excuse for infringement; the whole State was affected. In EC law, this meant that a Member State could not invoke its own territorial organisation in order to justify an infringement of law. There is consistent jurisprudence on this point.19

B. Phase Two: Regional Myopia The Treaty of Maastricht brought about a new phase in the recognition of the regions. From official blindness, the new situation evolved into a form of myopia: the regions started to be seen but through a rather distorting lens, making it impossible to perceive their differences. They were mingled within a single organ, where they were mixed with townships (the Committee of the Regions). During this phase, the structural principles of institutional autonomy and single legal

17 The culmination of this process can probably be seen in the adoption of the European Charter of Regional Democracy on 27–29 May 2008. The Assembly of European Regions has also been very active within the Council of Europe. See its Declaration on Regionalism in Europe (‘Basel Declaration’) of 4–5 December 1996. The Council of Europe dealt with regionalism in its political dimension, whereas the European Communities saw regions only as economic realities, at best as administrative realities for the effective deployment of structural funds. 18 See, eg, Commission, ‘The outermost regions: an asset for Europe’ (Communication) COM (2008) 642 final. 19 See, eg Case C-87/02 Commission v Italy [2004] ECR I-5975, para 38.

Autonomous Constitutional Regions in a Federal Europe 237 personality of the (Member) States were still paramount, but they developed in interesting and nuanced ways, as we shall attempt to explain. An interesting variant of the State personality-responsibility principle, according to which the whole of the Member State is responsible for breaches of EU law by any of its parts or emanations, was turned over and formulated as a positive obligation. All infra-State bodies of a public nature, or with any legal connection to the State, were expected and obliged to observe Community law. These bodies were even bound to set aside and disapply any internal laws incompatible with EC law.20 This meant that the regions’ administrations would, just like national courts, be directly obliged to give primacy to Community law, setting aside contrary domestic law. This requirement obviously placed them in a difficult predicament. i. Subsidiarity One important novelty of the Maastricht Treaty was the principle of subsidiarity, which, although based on the idea of proximity or closeness to the citizens, stopped short of involving the regions. Another interesting and useful tool that was introduced was the possibility for regional ministers to represent the whole Member State in Council meetings. This made it quite clear that the practice and extent of representation was a matter for each Member State’s internal constitutional arrangements or for the political will of the governments concerned. Many of the changes brought forward by the Maastricht Treaty were long-standing claims of the German Länder. Following a truly federal logic, they assumed that the federal identity of the German Republic had to have some sort of reflection in the EU system, and that the representation of the federal Government in the Council of Ministers was not, therefore, sufficiently representative of the federal State. ii. EU Involvement in Some Regions On the other hand, under the Maastricht and Amsterdam Treaties, the EU progressively got involved in regional issues whenever it was called on to do so by agreement of the Member States concerned, as was the case with the PEACE initiative in Northern Ireland or the creation of the Baltic euro-region, to mention but two examples. But the most interesting development was the renewed strength of the structural funds, which had a real effect on Member States that had no internal regional organisation (statistical regions) for they too started ‘regionalizing’. As a consequence of this perceived need to split Member States into regions that might opt for structural funds, anything went and anything goes when it comes to defining the regional level in the EU: there is no particular condition or requisite for regions to be considered as such by the EU; as long as a Member State declares that some entity is a region, there is in practice no

20

Case 103/88 Fratelli Costanzo SpA v Comune di Milano [1989] ECR 1839.

238

Joxerramon Bengoetxea

EU objection. Institutional autonomy protects the Member State’s artificial or discretionary definition of regions. The 2001 White Paper on governance was another important step in the strategic reflection on the role of all actors and stakeholders (including regions) and their involvement in the Community method, which the Commission wished to preserve and reinstate at all costs through consultation. Consultation is just as important as subsidiarity control because it is preventive rather than reactive; and the Commission undertakes formally to consult the regions and the REGLEG network. Furthermore, under the so-called Barroso initiative, the Commission consults national parliaments as well in an open dialogue. This is a more positive, proactive way to get involved in EU affairs. Yet problems arise when consultation becomes a one-way, unidirectional process, that is, when the Commission does not reply to the feedback. This can generate frustration on the part of the consulted party, who receives no feedback at all other than a polite note thanking it for its contribution. So the major challenge is whether consultation can somehow be structured and subjected to protocol, while still remaining informal and flexible.

C. Phase Three: Regions Gain Visibility, EU Institutions Acquire a Vision The Treaty of Lisbon has brought about some important changes and may be considered as a gaining of a more realistic vision of the regions,21 even if the distortions or blurred images as regards the Committee of the Regions have not been corrected and all sorts of regions are still represented together with the cities. But perhaps that is not such a setback after all. The interests of the cities are worth representing. The poleis need to be heard if the institutional picture is to be complete. The three major Lisbon innovations are, in my opinion, the following: (a) an imposition on the EU of a duty to respect national identity, including regional and local autonomy (Article 4 TEU); (b) inportant new developments concerning the principle of subsidiarity and the participation of Member State—including regional—parliaments as regulated in Protocols 1 and 2, which recognise the existence of decentralised or regional competences in some Member States. As a result, subsidiarity now penetrates below the Member State level, and requires examination of the regional issues. Also, the early warning system, with participation of Member State parliaments foresees the possible participation of regional parliaments either directly or indirectly through the federal chamber; and the possible right of action of the Committee of the Regions before the Court; (c) a clearer division and definition of competences in the TEU.

21 References to the regional reality are numerous throughout the new Treaties. See, eg, Arts 13 TFEU (as regards the protection of animal rights), 167 TFEU (as regards cultural heritage), 191 TFEU (as regards environmental protection) and 196 TFEU (as regards civil protection and risk prevention).

Autonomous Constitutional Regions in a Federal Europe 239 IV. THE PARADOX OF EU REGIONALISM: ‘NEITHER WITH NOR WITHOUT YOU’

As we have seen, it is no longer possible to understand the institutional architecture of the EU or the very process of European integration without taking the regional dimension into account. Yet there is such diversity, heterogeneity and disparity in the regional phenomena that the position of regions with a greater degree of competences is diluted in a system of inadequate representation. If one adds the constitutional EU general principle of the formal equality of the Member States, it seems impossible to change the current situation. Any solution to this dilemma must therefore respect three important, seemingly incompatible, aims: (a) Recognition of a special institutional reality, ie that of certain regions we here call ‘constitutional’, currently relegated to the Committee of the Regions. The recognition of a special EU status and the development of specific institutional fora within or outwith the Committee of the Regions could be a way to enhance such recognition. (b) Recognition of the need to represent all sorts of sub-State or infra-State entities generally (regions and cities of all types), because they are relevant actors in multilevel governance, close to the citizens, and need a voice. The current representation at the Committee of the Regions may be considered a suitable model for these entities. (c) Recognition of the formal and sovereign equality of the Member States and of their institutional autonomy. This recognition seems satisfactory at the formal or constitutional level, but is evolving in interesting ways, as we shall see. Assuming that the recognition of sub-State entities generally at the Committee of the Regions constitutes an interesting system of participation and representation (though there is some room for improvement), and that regions without legislative powers and cities are relatively satisfied with the current constitutional make-up of that organ, the three-pronged tri-lemma becomes a di-lemma: how to reconcile special status for the self-governing constitutional regions—a reality affecting only the complex, quasi-federal Member States- with the principles of equality and institutional autonomy of the Member States, unitary or complex?

V. A POSSIBLE SOLUTION: A SPECIAL STATUS FOR CONSTITUTIONAL REGIONS

A. The Real Change of Vision: the Court of Justice Before the final agreement was reached on the Treaty of Lisbon, the ECJ ruled on 6 September 2006 on the Portugal v Commission case (‘the Azores case’). This was the most important legal development of the regional question up to that point by any of the EU institutions. The problem arose from the strict approach adopted

240

Joxerramon Bengoetxea

by the Commission on regional systems of taxation. The Commission considered benefits like lower tax rates, deductions or exemptions almost automatically as incompatible State aid because of their allegedly selective nature. It held that since regional fiscal advantages are territorially circumscribed, and different from the tax treatment in the rest of the Member State concerned, they must by definition ‘favour certain undertakings or the production of certain goods’, thus contravening ex Article 87 EC. The Court clearly corrected the Commission’s understanding. Systems of regional taxation may be compatible with the Treaty and are not to be considered State aid when three conditions are met: constitutional autonomy;22 decisional autonomy;23 and financial autonomy of the (regional) taxation system.24 In the case of the Azores, the third condition was clearly not met, because the State of Portugal ultimately compensated for the loss of revenue originating in the tax benefits by budgetary transfers to the Azores. These three conditions were confirmed in a decision adopted by the Court two years later, on 11 September 2008, in a preliminary ruling on the Basque tax regime.25 The Court examined in great detail all the relevant rules concerning constitutional autonomy and self-governance in the field of taxation enjoyed by the Basque provinces or ‘Historic Territories’ (in fact federated entities within an historical nation or Autonomous Community). Most notably, the Court considered the decision-making powers enjoyed by the Historic Territories’ parliaments or Juntas Generales, the system of co-ordination without possibility of interference from the Spanish taxation authorities, and the system of cupo or quota, whereby the Basque Autonomous Community calculates the sums it pays to the Spanish Executive for the services it deploys and provides in the Basque Country. In order to assess the extent of the financial autonomy of the regional system, the Court examined the rules and principles governing the Spanish direct taxation system and the regional fiscal policies in extraordinary detail, as Advocate General Kokott had done before in her thorough opinion. Two further relevant cases are Horvath26 and Gibraltar.27 The Opinion given in Gibraltar, a State aid case, confirms the view set out above, though adding interesting developments. Horvath, on the other hand, concerned the EU Single Payment Scheme set up by Regulation 1782/2003, which replaced the previous

22 The decision in question must have been taken by a regional or local authority having, from a constitutional point of view, a political and administrative status separate from that of the central government. See Portugal v Commission, above n 4, para 67. 23 This may be called decisional or procedural autonomy. The decision must have been adopted without the central government being able to intervene directly as regards its content. See ibid, paras 67, 68 and 72. 24 This may be called the economic and financial autonomy. The financial consequences of the decision must not be offset by aid or subsidies from other regions or the central government. See ibid, paras 67, 72 and 75. 25 UGT-Rioja, above n 6, paras 51, 87, 107 and 135. 26 See n 7 above. 27 See n 8 above.

Autonomous Constitutional Regions in a Federal Europe 241 system of agricultural subsidies.28 Farmers receiving direct payments have to maintain their land in ‘good agricultural and environmental condition’.29 Member States are obliged to define, ‘at national or regional level’, minimum requirements for ‘good agricultural and environmental condition’.30 In England, the Secretary of State defined these minimum requirements as including obligations relating to the maintenance of footpaths. In Wales, Scotland and Northern Ireland, by contrast, no such condition was imposed. Mr Horvath claimed—inter alia—that the differential implementation of the Regulation in England and Wales constituted a breach of the fundamental principle of non-discrimination in EU law. The Court, however, ruled that each Member State is free to allocate powers internally and to implement Community acts which are not directly applicable by means of measures adopted by regional or local authorities, provided that that allocation of powers enables the Community legal measures in question to be implemented correctly.31

For the Court, there is no discrimination contrary to EU law where different implementing bodies within a Member State implement Union legislation in different ways, provided that, in each case, the devolved legislation affects equally all persons subject to it and complies with EU law.32

B. A Changing Parameter: the New Interpretation of Institutional Autonomy of the Member States The main development inspired by this line of case law is that the principle of the institutional autonomy of the State is in a process of silent transformation. International law has developed in interesting ways. The scope of ius cogens nowadays is such that the Member States’ obligations under international law are greater than ever, reducing the significance or the scope and intensity of the right to internal self-determination. But much more relevant is the fact that EU law is developing in a way whereby the internal organisation of the Member States has direct implications in EU law as well. A new awareness of regional diversity is emerging. The Member State might still be free to organise as it pleases, but the EU, through its institutions, is now expected and even obliged to take into account how the State organises itself constitutionally; this follows from the obligation to respect the constitutional identity of the State. Even if one were, following Alan

28 Council Regulation (EC) 1782/2003 establishing common rules for direct support schemes under the common agricultural policy and establishing certain support schemes for farmers and amending Regulations (EEC) 2019/93, (EC) 1452/2001, (EC) 1453/2001, (EC) 1454/2001, (EC) 1868/94, (EC) 1251/1999, (EC) 1254/1999, (EC) 1673/2000, (EEC) 2358/71 and (EC) 2529/2001 [2003] OJ L270/1. 29 Ibid Art 3. 30 Ibid Art 5(1). 31 Horvath, above n 7, para 50. 32 Ibid paras 55 and 57.

242

Joxerramon Bengoetxea

Dashwood,33 to admit the qualification of the EU as ‘a constitutional order of States’, respect for such order would necessarily also require respect, rather than disregard, for the constitutional make-up of those States. How the Member State is organised is no longer irrelevant. The EU must now look inside precisely so as not to disrupt the institutional autonomy. As a consequence, if the EU were to treat a complex, decentralised Member State as though it were a unitary State, it would be infringing the principle of institutional autonomy and respect for the national identity of the Member State (Article 4 TEU). This is a novel understanding of the principle of institutional or constitutional autonomy.

C. Recognition and Accommodation Some infra-State entities have special features that should lead to recognition of their special autonomous status: constitutional visibility, defence of their own prerogatives and competences, participation in the ‘community method’, and loyalty to the EU in implementing its law and contributing to furthering its aims. This is why the term ‘self-governing constitutional regions’ may be used to refer to that special status. They are a special type of region within the wider, neutral meaning of ‘region’ in the EU. They are regions that have acquired, developed and exercised an important degree of self-governance, whether or not their powers are seen as original or as devolved, and their self-governing powers are constitutionally recognised. On top of that, these regions are instrumental in achieving the aims of the EU. The term ‘self-governing (or autonomous) constitutional region’ is, in my view, better adjusted to the EU constitutional order than the term ‘associated region’ once advocated by Lamassoure.34 The latter concept implies some type of external relationship rooted outside the Union, as is the case for associated third countries.

D. Special Institutional (Legal) Status The existence, and recognition, of any institution or institutional status depends on a theory of law. The late Neil MacCormick’s institutional theory of law understands law as ‘institutional normative order’ and is one of the few contemporary theories of law that has dissociated the concept of law from the institution of the State.35 It is, therefore, capable of explaining non-State law or legal systems outwith the State. His theory applies to EU law as well as to other transnational laws and sui generis forms of law prevalent in the legal plurality and multiculturalism 33

See n 1 above. See section III. above. 35 N MacCormick, Institutions of Law: An Essay in Legal Theory (Oxford, Oxford University Press, 2007), passim. 34

Autonomous Constitutional Regions in a Federal Europe 243 that characterises contemporary Europe. With ‘law’, a system of primary and secondary legal norms emerges, and especially norms that define the status, powers and positions of officials entitled to create new law, to apply and control the application of law, and to decide disputes by applying such rules. Regional law, institutional normative orders of specific territories, even if understood as sub-systems of larger legal orders of the Member State, are now all integrated in the larger system of EU law, in its constitutional order, itself perhaps a sub-system of international law. The main idea is to understand the connections between legal orders, but at the same time their degree of self-governance and autonomy in specific areas. There is a clear federal logic behind this complex arrangement. Federalism is inspired in the principle of co-operation between public orders. In the EU, federalism is multilevel. Each public order or entity is autonomous within its own field of competence, provided certain general principles of law are respected, and each entity co-operates in the application and achievement of the aims and objectives of the shared system (federal or EU loyalty). As regards shared competences, the principle of subsidiarity offers sufficient guidance to all systems concerned, at least when applied in a federal co-operative rather than competitive context and complemented with the principle of proportionality. The recognition of a special status for the self-governing constitutional regions would precisely imply that the system of EU law includes the Member States’ sub-systems, some of which themselves being, in turn, made up of smaller sub-systems. When considering criminal law, for instance, there is a set of shared norms at the EU level, but the majority of the law is located at the Member State level. In some cases, however, that domestic level (eg, the UK) is simply the accumulation of further sub-systems of criminal law (ie, Scotland, England and Wales, Northern Ireland, arguably, the Isle of Man, a Crown dependency, …). Likewise, as concerns direct taxation, there are very few norms of EU law.36 Most norms belong to the Member State legal systems. Yet in some States, like Spain, the national legal system consists of the accumulation or co-ordination of different sub-systems (Navarre, the three Basque provinces and the rest of Spain). A variable geometry obtains, or ‘asymmetric federalism’ as it is sometimes called. As a result, our very understanding of the concept of ‘State’ is transformed into that of a complex Member State, duly recognised by the EU. Within this view, regions and their legal systems are understood as ‘fragments’37 or sub-systems of Member-Statehood. The point is not that, in order to be considered an ‘autonomous constitutional region’, a region must fulfil the ‘Azores criteria’. This would be to impose a uniform solution on a very diverse reality. These criteria are applicable only as regards differentiated national taxation systems. For other purposes, the conditions may 36 Though there are certain limitations imposed by the fundamental freedoms and the prohibition of discrimination. 37 The term ‘fragments’ is inspired in G Jellinek, Über Staatsfragmente (Heidelberg, G Koester, 1896).

244

Joxerramon Bengoetxea

be different. In the context of social policy, for instance, it is required that the regional body be empowered to adopt decisions that affect and benefit all persons living in its territory.38 The kind of region envisaged in this chapter has legislative powers, and is accountable to its electorate, which gives it a constitutional character. These regions are geographically circumscribed, and have autonomous powers in certain subject matter areas. Ultimately, what is essential in order to be considered an autonomous constitutional region is to have a parliament or legislature with law-making powers in certain fields recognised as regional. That parliament appoints and controls a regional government, which has policy-making and regulatory powers in the same (regional) subject matter areas. Since the distribution of powers between the central State and the regions is different in every Member State, an asymmetric picture of the EU constitutional order arises. Though there is no problem with asymmetry as such, it obviously complicates the EU constitutional make-up.

E. A New Interpretation of the Equality of Member States and their Institutional Autonomy We have already explained the new interpretation of the principle of the institutional autonomy of the Member States, which is evolving in line with Article 4 TEU, that is, with respect for national identity, including regional and local autonomy. We also propose a novel reading of the principle of equality between Member States. This reading is inspired by the Lisbon decision of the German Federal Constitutional Court (Bundesverfassunsgericht) in relation to the principle of equality of votes and the different demographic representation of the Member States in the European Parliament.39 Let us attempt to explain this point. In a purely numerical understanding of democracy, a strictly proportional representation of voters could determine the composition of a parliament. In such a situation, it would be necessary to elect the parliamentary representatives from a single constituency or from constituencies composed of roughly the same number of voters. Yet as far as the European Parliament is concerned, the current electoral system does not neatly follow the principle of equal representation of voters, which is one of the reasons why the EU is said to exhibit a ‘democratic deficit’. However, if one takes into account that the principle of equal representation of voters is to be reconciled with a federal principle of representation of the EU’s plurality of Member States or demoi, the current system makes sense. Indeed, a balance has been sought between equal representation of voters and respect for ‘national’ plurality and diversity by means of a sufficient representation of smaller demoi. 38 Or in two different territories, as was the case for the Flemish Community, which has competence over all persons living in Flanders and over (part of) the Flemish-speakers living in the Brussels Region. 39 BVerfG, 2 BvE 2/08, 30 June 2009, paras 278–97.

Autonomous Constitutional Regions in a Federal Europe 245 In federations, the said balance is often achieved through a bicameral system. In the EU system, however, the balance is secured by a variety of measures, such as the need for ratification by all Member States; the guarantee of a minimum number of MEPs, leading to the over-representation of the smaller Member States or under-representation of the larger ones;40 and the double majority system in the Council, which somehow consolidates the comparatively greater weight of the smaller Member States.41 As a result of all these factors, smaller Member States could be said to be over-represented in the EU. Obviously, this means that, from a formal point of view, the Member States are not equal. Strict uniformity is not a prerequisite for equality; on the contrary, equality might require making some, acceptable, distinctions. But what distinctions are adequate is a matter for rational practical argumentation. In the light of the foregoing, the moment has probably come also to rethink the way plurinational, complex, federal or decentralised Member States (the UK, Germany, Italy, Spain, Belgium) and Member States with overseas, peripheral territories and regions (France, Portugal) are represented at the EU level. As the aforementioned representation of smaller and larger Member States exemplifies, the (material) equality of States can be preserved without uniformity needing to be imposed. The contention in this contribution is that, in a similar vein, the EU should recognise the special position of those regions and territories that, although they are part of a larger State, enjoy internal constitutional status, have their own legislature and government,42 and have been granted autonomous decision-making powers, including the power to apply and implement EU law.

VI. RECOGNITION, ACCOMMODATION AND INSTITUTIONAL CONSEQUENCES

The recognition of the special status of autonomous constitutional regions implies—or should imply—some institutional consequences. The following are some sketchy suggestions for adjustments that might be made to the current EU architecture, some of them requiring amendments to the Treaties, others not. A. Committee of the Regions The most obvious organ to start with is that representing the regions, ie the Committee of the Regions. Apart from important constitutional regions,

40 The weight of a Maltese citizen’s vote is about 12 times the weight of a German citizen’s vote (ibid paras 284–85). 41 On this issue of over-representation or underrepresentation, see D Edward and J Bengoetxea, ‘The Status and Rights of Sub-state Entities in the Constitutional Order of the European Union’ in the Dashwood Festschrift, above n 1. 42 And sometimes even their own judiciary, as is the case in Scotland.

246

Joxerramon Bengoetxea

sub-State entities seem to be satisfactorily represented in the Committee.43 The possibility of creating a special chamber composed of the constitutional regions within the Committee is worth exploring, though. Another option would be to accord greater weight to constitutional regions in decisions on reports dealing with issues as regards to which only these regions have legislative powers. Lastly, it may be appropriate for constitutional regions to have a special voice when the Committee is considering starting judicial proceedings on the basis of the subsidiarity principle. This last change may result from a ‘natural evolution’ within the Committee, without there being any need to force the formula into a legislative text.

B. European Parliament Electoral constituencies for the European Parliament should correspond to the territory and population of the constitutional regions. Moreover, the CALRE should be recognised as a special interlocutor of the European Parliament, in the same way that the representative of the Member States’ Parliamentary Assemblies (COSAC) has been so far. The CALRE might also be designated as a special interlocutor of COSAC and the Committee of the Regions. The parliamentary commission on regional development should have closer ties with the Committee of Regions, especially in relation to the principles of subsidiarity and proportionality.

C. European Commission The Commission should develop a clearer view of the regional phenomena and a special sensitivity to regional particularities. The abovementioned solution of the tripartite contracts between the Commission, the Member States and the regional governments for the implementation of EU policies could be revived.44 Moreover, the European groupings of territorial co-operation45 might be reinforced and spread into many new areas. They might be further endowed with a special EU status and operating legislation, so that they need no longer depend on the intricacies of the laws of the Member States involved in the territorial co-operation, each imposing its own requirements on extraterritorial legal relationships (eg, as regards social security). In the context of network governance, the Commission might consult those regions (and associations of regions) affected by its proposals for specific measures

43 Nevertheless, the Committee has some phenotype (rather than genotype) shortcomings, such as the grouping of its members on the basis of the Member State to which they belong. 44 See n 14 above. 45 See n 13 above.

Autonomous Constitutional Regions in a Federal Europe 247 and policies, as it already does in certain subject matter areas. Furthermore, greater weight should be given to transfrontier interregional co-operation. Regions (or associations of regions), for their part, should improve their communication with and representation in Commission agencies (eg, Human Rights, Health and Safety, Environment, Food). Certain constitutional regions host such European agencies and decentralised bodies.46

D. Council of Ministers Where there are, in the Council of Ministers, rotating presidencies of federal Member States, regional ministers might take the lead in some of the Council meetings and co-ordinate with COREPER. This interesting (but rare) practice might be generalised for some Council formations and occasionally be extended to some European Council meetings. Regional ministers should be actively participating in Council meetings, and COREPER delegations should better reflect regional realities. As regards powers usually reserved to the regions but subject to the open method of co-ordination, the Member State should be represented in the Council by a delegation of the regions concerned, or by one of the regional ministers. Moreover, it should be explored whether a split vote system might be introduced in the Council. Lastly, it would be proper for REGLEG to adopt common positions in order to have an impact on important Council and European Council meetings.

E. Economic and Social Council of the EU The Economic and Social Council of the EU should enter into dialogue with the regional Economic and Social Committees, and with the Member States’ ECOSOC Councils (in those Member States where they exist). A network of the different economic and social committees of the various levels—European, Statenational or regional-national—could be a very effective consultative body, where much of civil society is represented.

F. Court of Justice Some of the classical claims of autonomous constitutional regions insist on the active judicial legitimation of the Committee of the Regions to bring actions before the Court of Justice. Right now, this possibility is limited to subsidiarity 46 See, eg, the Agency for Health and Safety at Work in the Basque Country, the Food Authority in Emilia Romagna, the Aviation Safety Agency in Nordrhein-Westfalen, the GSA Agency in Brussels and the European Training Foundation in Piemonte.

248

Joxerramon Bengoetxea

control, but privileged standing might be extended to other issues. Furthermore, locus standi could be accorded to the regions as privileged plaintiffs, such that they would have a status similar to that of the Member States. This would be a great improvement compared to the current system in which regions have access to the Court only as ‘individual’ plaintiffs. After all, autonomous constitutional regions are involved in the implementation of EU law and legislate in areas affected by EU law. Another suggestion would be for the Court to make sure that the various legal systems internal to the Member States are properly represented within its research and documentation service. In addition, it would be appropriate if the Court’s translation services were also to operate in the languages which have (co-)official status within the constitutional regions, even if they have not been recognised as official languages of the EU.

G. Appointments and Nominations Appointments to the Court of Justice (judges as well as Advocates General) should be arranged by Member State governments so as to ensure the representation of the Member States’ various legal systems. This would be simply a matter of ‘regional sensitivity’ or awareness of the State’s internal legal and/or linguistic diversity.47 This concern also extends to appointments to the highest posts in other EU institutions, agencies and bodies. Just as Member State governments take gender diversity into account, they should be equally concerned about regional diversity. Indeed, it should be ensured that the highest posts at the EU level are not always reserved for professionals from the Member State capitals or financial centres.

VII. CONCLUSION

This chapter contains a variety of proposals for institutional change. However, change often comes from new understandings of existing concepts and principles. The major change might therefore result from a novel conception of the status of constitutional regions in EU law, as explained in section V. Obviously, such evolutions take time. Yet they are far-reaching. The principle of institutional autonomy of the Member States, for instance, has silently transformed into a much more subtle and nuanced principle of respect for internal variety and difference, of sensitivity to plurality and diversity within the Member States. This novel conception echoes the ideas following from the politics of recognition, reasonable accommodation, and cultural and national diversity management. 47 The UK used to provide a good example in this sense. Also Belgium seems to exhibit sensitivity to its internal diversity.

11 The European Court of Justice and the Devolution of Taxation Powers SUZANNE KINGSTON*

I. INTRODUCTION AND BACKGROUND

F

OR A NUMBER of Member States, the constitutional balance between certain regions and central government has been struck by means of devolution. While there are probably as many varieties of devolution as there are devolved States, it can, at its broadest, be understood as the transfer of competences from a superior to an inferior regional political authority.1 Such transfer may occur in relation to legislative, executive and judicial powers, but may differ considerably per State, and indeed within States, as to its nature and degree, as well as the legal means by which it is achieved. Thus, while in Member States with written constitutions, devolution tends to be constitutionally enshrined, with further detail added by implementing laws,2 in the UK, without any written constitution, devolution was achieved by ‘mere’ Acts of Parliament.3 Similarly, while in some Member States devolution of powers has been achieved on a symmetrical basis, with all regions granted devolved powers pursuant to a federal model,4 in others devolution takes place on an asymmetrical basis, with only certain regions or communities granted a (greater) level of autonomy from the central State.5 The fundamental challenges to national identity that may be posed by devolution have been addressed at length in the political science literature.6 While, from one perspective, the lessening of control by central State powers may represent a politically palatable compromise in the face of separatist demands, from another it may represent the first step in an irreversible process leading, potentially, to the *

Barrister; Lecturer, University College Dublin. See V Bogdanor, Devolution in the United Kingdom (Oxford, Oxford University Press, 2001) 2. 2 See, eg the Spanish and Portuguese examples in relation to the Basque Country and the Azores, set out below. 3 See the Scotland Act 1998, the Government of Wales Act 1998 and the Northern Ireland Act 1998. See also the legal framework in relation to Gibraltar, described below. 4 An obvious example being Germany. 5 As is the case, for instance, in the UK, Spain and Portugal (see further below). 6 See, eg, J Wilson and K Stapleton, Devolution and Identity (Aldershot, Ashgate, 2006). 1

250 Suzanne Kingston dissolution of the State.7 Thus, in his 2010 autobiography, Tony Blair, the onetime hero (or villain, depending on your perspective) of devolution in the UK due to his crucial role in bringing in Scottish and Welsh devolution, admitted deep concerns about the process, describing it as a ‘dangerous game to play. You can never be sure where nationalist sentiment ends and separatist sentiment begins.’8 For many Member States, therefore, the decision to engage in the devolution process is a highly sensitive one, entered into only after a long period of political reflection and consultation, with stakes of the highest order. In this context, the challenge to EU law posed by devolution is evident. As with any international treaty, the signatories of the EU Treaties are of course limited to sovereign States. The constitutional law of the EU has, therefore, long grappled with the difficulties of reconciling its nature as a Union of Member States with the particularities of those Member States’ constitutional settlements that challenge traditional notions of sovereignty. From one perspective, the solution found by the Treaties is clear: Union constitutional law recognises and seeks to respect the Member States’ competence to decide on their own particular constitutional settlement. This stance is epitomised in the current version of the Treaty on European Union (TEU) by Article 4(2), which states that [t]he Union shall respect the equality of Member States before the Treaties as well as their national identities, inherent in their fundamental structures, political and constitutional, inclusive of regional and local self-government. …

This provision represents a considerable strengthening of emphasis in comparison to its pre-Lisbon form, which simply stated that the Union shall respect Member States’ national identities, with no specific mention of regional and local selfgovernment.9 The Union has sought to recognise regional identities in a variety of ways, not least by giving regions an input into the EU policy- and law-making process via the Committee of Regions, as well as by supporting financially weaker regions via its Structural Funds. Yet perhaps the greatest challenge posed by regionalisation in the EU context arises where devolution of powers to regions appears to come into conflict with fundamental principles of the EU legal order and, in particular, those of the internal market and EU economic law. In such circumstances, how does one genuinely respect the Article 4(2) TEU principle in the light of the

7 A classic illustration of this fear is the refusal by the Westminster Government to accede to demands for Home Rule (ie, devolution) for Ireland made in the late 19th century, led by John Redmond of the Home Rule Party. The Irish Home Rule movement was ultimately superseded by the nationalist movement and 1916 Easter Rising, which led to the creation of the Irish State. 8 T Blair, A Journey (New York, Random House, 2010). See also R Davies, Devolution: a Process not an Event (Cardiff, Gregynog Papers, Institute of Welsh Affairs, 1999); J Mitchell, Devolution in the UK (Manchester, Manchester University Press, 2009). 9 Art 6(3) TEU (Nice version).

The ECJ and the Devolution of Taxation Powers 251 ECJ’s consistent jurisprudence that the specificities of national law cannot lead to differences in the application of rules that lie at the core of the EU Treaties? This chapter will consider how the ECJ has answered this question in its jurisprudence to date in the particular context of devolution of taxation powers, an area where the potential conflict between EU economic law and regional devolution is perhaps most acute, for a number of reasons. First, direct taxation remains a key, and highly sensitive, area of competence of the Member States.10 Secondly, the power to tax is traditionally viewed, along with foreign policy competence, as one of the key hallmarks of State sovereignty; devolution of taxation powers clearly forces us, and the ECJ, to rethink traditional concepts of sovereignty and the State. In some contexts, this has not posed undue difficulty. Thus, in relation to the four freedoms of EU internal market law, it is long-established that taxation rules resulting in restrictions on the free movement of goods, persons or services are, unless justified, prohibited by the Treaty even if imposed by regions rather than by the State.11 As this principle does not in itself question the devolution of taxation powers to regions but rather the manner in which such powers are exercised, it is relatively uncontroversial. It is in the field of State aid law, however, where the most obvious potential for conflict has arisen; in particular, in the context of regional tax systems that provide for a lower tax burden, or greater tax advantages, than in the rest of the Member State at issue. Article 107(1) TFEU, which defines the scope of the concept of aid, provides: Save as otherwise provided in the Treaties, any aid granted by a Member State or through State resources in any form whatsoever which distorts or threatens to distort competition by favouring certain undertakings or the production of certain goods shall, in so far as it affects trade between Member States, be incompatible with the internal market.

As is well known, this Article qualifies as aid State measures that satisfy a four-step test: (a) The aid must ‘confer on recipients an advantage which relieves them of charges that are normally borne from their budgets’, a requirement clearly satisfied where an undertaking is subject to a lower tax rate than would otherwise apply.12 (b) The aid must be granted ‘by a Member State or through State resources’, a requirement satisfied where the more favourable tax rules are laid down by regional authorities.13

10 See S Kingston, ‘The Boundaries of Sovereignty: the ECJ’s Controversial Role Applying Internal Market Law to Direct Tax Measures’ (2006–2007) 9 Cambridge Yearbook of European Legal Studies 287; S Kingston, ‘A Light in the Darkness: Recent Developments in the ECJ’s Direct Tax Jurisprudence’ (2007) 44 CML Rev 1321. 11 See, eg Case C-169/08 Presidente del Consiglio dei Ministri v Regione Sardegna [2009] ECR I-10821, paras 29–30, and the jurisprudence cited therein (legality of regional legislation establishing tax on stopovers for tourist purposes by certain aircraft). 12 Commission, Notice on the application of the State aid rules to measures relating to direct business taxation [1998] OJ C384/3, para 9. 13 See Case C-248/84 Germany v Commission [1987] ECR 4013.

252 Suzanne Kingston (c) The aid must favour ‘certain undertakings or the production of certain goods’. (d) The aid must affect trade between Member States. It is at the third step, commonly known as the selectivity requirement, where most issues arise in the context of regional devolution of taxation powers.14 The nub of the challenge lies in defining the boundaries of geographical selectivity: that is, in which circumstances should measures favouring undertakings in a certain region or regions be considered selective? This issue has arisen in a number of high-profile, and highly politicised, recent cases before the ECJ and the Court of First Instance (‘CFI’, now the ‘General Court’), to which we now turn.

II. REGIONAL TAXATION AND THE SCOPE OF ARTICLE 107(1) TFEU

A. Early Judgments Prior to the ECJ’s judgment in the Azores case,15 it had never explicitly considered the question of the circumstances in which regionally-specific tax measures should be considered to be selective under the Article 107(1) TFEU definition of State aid. Rather, its judgments applying the definition of aid in the context of regional powers had been confined primarily to the issues of the scope of the concept of ‘State resources’ and the imputability of regional measures to the State, and had considered the selectivity issue only obliquely. This was understandable, given the relative rarity of asymmetrically-devolved taxation powers, and the fact that what is now DG Competition initially focused its energies primarily on antitrust, rather than State aid, enforcement. Thus, as far back as 1987, the Court had in Germany v Commission confirmed that aid granted by a regional body (in that case, aid granted by the German Land of North Rhine-Westphalia under regional legislation) could be viewed as ‘aid granted by a Member State or through State resources’ under Article 107(1) TFEU ‘if the conditions laid down in that provision are satisfied’ (this last caveat, of course, being crucial).16 Similarly, the Court had on a number of occasions considered measures of central government granting tax relief to particular regions, meaning that the selectivity criterion was clearly satisfied.17 It has also long been

14 As consistently held by the ECJ, Art 107(1) TFEU requires assessment of whether, under a particular statutory scheme, a State measure is such as to ‘favour certain undertakings or the production of certain goods’ in comparison with other undertakings which are in a legal and factual situation that is comparable in the light of the objective pursued by the measure in question (see, eg Case C-143/99 Adria-WienPipeline and Wietersdorfer & Peggauer Zementwerke [2001] ECR I-8365, para 41; Case C-308/01 GIL Insurance and Others [2004] ECR I-4777, para 68). 15 Case C-88/03 Portugal v Commission [2006] ECR I-7115 (‘the Azores case’). 16 Germany v Commission, above n 13, para 17. See similarly, in relation to aid granted by Wallonia, Joined Cases 62/87 and 72/87 Exécutif régional wallon and SA Glaverbel v Commission [1988] ECR 1573. 17 Case C-156/98 Germany v Commission [2000] ECR I-6857 (measure of central German Government granting tax relief to a Land).

The ECJ and the Devolution of Taxation Powers 253 clear that it is the Member State, not the region, which is formally responsible for breaches of the State aid rules carried out within its territory, even if such breach is due to the actions of the region rather than central government.18 These cases were interpreted by certain commentators as meaning that regional tax measures were always geographically selective.19 Upon careful reading, however, the better view is that, while the judgments concerned cases of regional tax measures, they never considered the specific question of the extent to which tax measures taken by autonomous devolved regions should be considered to be selective. Prior to Azores, the nearest the Court came to considering this issue was the first Basque case in the late 1990s, Juntas Generales de Guipúzcoa, which concerned the compatibility with Article 107 TFEU of the provincial laws (‘normas forales’) adopted by three Territories belonging to the Autonomous Community of the Basque Country.20 These laws provided for a more favourable tax regime than that generally applicable within the Spanish territory, with the aim of stimulating economic activity and promoting investment. While the case was withdrawn before the Court could hand down its judgment, it was not before Advocate General Saggio had the opportunity to deliver an Opinion, in which he took a rather strict approach to the definition of State aid in the context of the exercise of devolved taxation powers. In the Advocate General’s view, the Basque laws were selective in nature as they were ‘intended exclusively for companies situated in a particular region of the Member State in question’, giving such companies an advantage not available to similar companies operating in other parts of the State.21 Further, in explicitly rejecting the Spanish Government’s argument that finding against the laws at issue would be ‘tantamount to issuing a value judgment on the constitutional structure of the Spanish State’, the Advocate General argued that, if a sub-national reference framework were accepted, this would enable the State to ‘easily avoid’ the application of the State aid rules ‘simply by making changes to the internal allocation of competence on certain matters’—changes which the Advocate General considered to be merely formal, rather than substantive, in nature.22 In one sense, therefore, the Advocate General’s view represents a rather cynical take on devolution, rejecting the idea that the fiscal autonomy of the Basque Territories represented ‘any specificity’ of such Territories.23 However, he stopped short of setting out generally applicable principles for the assessment

18 Case 130/83 Commission v Italy [1984] ECR 2849. It was no defence in that case that Italy had tried to persuade the Sicilian authorities to repeal the measures at issue. 19 See further, the literature cited at n 40 below. 20 Joined Cases C-400/97, C-401/97 and C-402/97 Juntas Generales de Guipúzcoa and Diputación Foral de Guipúzcoa [2000] ECR I-1073. 21 Juntas Generales de Guipúzcoa and Diputación Foral de Guipúzcoa, Opinion of AG Saggio, para 35. 22 Ibid paras 37–38. 23 Ibid para 38.

254 Suzanne Kingston of regional tax variations, preferring to base his conclusion on the fact that the measures at issue constituted mere ‘short-term measures’ aiming to improve the competitiveness of the companies in that area, rather than measures of general economic policy.24 While the Court did not get to rule on the issue in that case, a similar question arose before the CFI some two years later in Territorio Histórico de Álava, which concerned, inter alia, the tax credit of 45 per cent granted to companies making investments of over €2.5 billion in new fixed assets within this Basque Territory.25 Amongst many arguments made in defence of the Territory’s laws was an argument that the Territory had historically enjoyed constitutionallyprotected autonomy in taxation matters. Referring to Germany v Commission, the CFI rejected this argument, noting that measures adopted by intra-state entities (decentralised, federated, regional or other) of the Member States, whatever their legal status and description, fall … within the ambit of Article 92(1) of the Treaty [now Article 107(1) TFEU] ‘in the same way as measures taken by the federal or central authority,’ but once again adding the crucial caveat, ‘if the conditions laid down in that provision are satisfied’.26 The issue of regional selectivity was, therefore, kicked down the road, although the compatibility of Basque taxation laws would shortly be back before the Court once again, as detailed below.

B. The Azores Judgment—Taking Devolution Seriously? Six years after Advocate General Saggio’s Opinion in the first Basque case, the Court was faced head-on with the question of how to apply the regional selectivity concept in the context of taxation powers that had been devolved to an autonomous region, and rose to the challenge by setting out, for the first time, a test to determine the circumstances in which the exercise of devolved taxation powers constitutes State aid under what is now Article 107(1) TFEU. The case concerned the special tax regime that applied in the form of tax reductions for residents of the Autonomous Region of the Azores, a part of the Portuguese State. In particular, the regional legislative assembly of the Azores had passed a decree providing for reduced rates of income and corporation tax, as compared to the national levels, for all Azores residents, whether legal or natural

24

Ibid. Joined Cases T-127/99, T-129/99 and T-148/99 Territorio Histórico de Álava—Diputación Foral de Álava and Others v Commission [2002] ECR II-1275, appealed on other grounds in Joined Cases C-183/02 P and C-187/02 P Demesa and Territorio Histórico de Álava v Commission [2004] ECR I-10609. See also Joined Cases T-346/99, T-347/99 and T-348/99 Territorio Histórico de Álava— Diputación Foral de Álava and Others v Commission [2002] ECR II-4259. 26 Territorio Histórico de Álava—Diputación Foral de Álava and Others v Commission, above n 25, para 142. 25

The ECJ and the Devolution of Taxation Powers 255 persons.27 The decree was passed pursuant to a Portuguese State law establishing the principles of financial regional autonomy which provided, inter alia, for certain regional tax competences for regional bodies, including the power to reduce national corporation and income tax and VAT rates by up to 30 per cent.28 In turn, this law reflected provisions of the Portuguese Constitution, which provides that the autonomous regions shall receive their own tax revenue as well as part of State tax revenue, as established by a principle of national solidarity, and have the right to exercise their own fiscal competence.29 The case concerned an action to annul a Commission decision of 2002 finding the Azores tax reductions to be unlawful State aid, on the ground that they constituted an advantage granted through State resources that favoured certain undertakings (namely, those resident in the Azores), which affected trade and competition between Member States.30 Rather unsurprisingly, the United Kingdom and Spain, as the principal other jurisdictions with significantly devolved taxation powers, intervened in support of the Portuguese challenge. As will be recalled, the Azores, along with the French overseas departments, Madeira and the Canary Islands, enjoy a special status under EU law, constituting ‘outermost regions’ within the meaning of Article 349 TFEU (ex Article 299(2) TEC). This legal basis provision specifically recognises the particular structural social and economic situation of these regions, and ‘their remoteness, insularity, small size, difficult topography and climate, economic dependence on a few products, the permanence and combination of which severely restrain their development’. In his Opinion, Advocate General Geelhoed identified the issue of geographical selectivity as key in assessing the extent to which devolved taxation rules constituted State aid under what is now Article 107(1) TFEU and, more specifically, the issue of which reference framework should be selected in comparing the regional rules with the ‘normal’ rules in any given case.31 In rejecting the Commission’s view that the reference framework should always be the entire territory of the relevant Member State, the Advocate General began by distinguishing the situation at hand, where certain, but not all, regional bodies enjoy autonomous taxation powers, from situations where the central government passes tax reductions for a specific region (which is evidently selective) and situations where all local authorities enjoy autonomy of taxation powers (which is evidently not selective). For such asymmetrical cases, the Advocate General proposed that the exercise of taxation powers by local authorities that are ‘truly’ autonomous from the central

27 Regional legislative Decree No 2/99/A of 20 January 1999. The rates were set at 15% to 20% for income tax and 30% for corporation tax. 28 Law 13/98 of 24 February 1998. 29 Portuguese Constitution, Arts 227(1)(i) and (j). 30 Commission Decision 2003/442/EC [2003] OJ L150/52. 31 Portugal v Commission, above n 15, Opinion of AG Geelhoed. The present author was the référendaire dealing with the case in the Advocate General’s cabinet.

256 Suzanne Kingston government should not be selective, and adopted a three-prong test of selectivity in this respect, namely:32 (a) Is the local authority institutionally autonomous, with its own constitutional, political and administrative status separate from the central Government? (b) Was the taxation measure taken by the local authority in a procedurally autonomous manner, ie, so the central Government did not have any power to intervene directly in the procedure of setting the tax rate, and the local authority did not have any obligation to take the interest of the central State into account in setting the regional tax rate? (c) Was the taxation measure taken by the local authority in circumstances of economic autonomy, ie, such that the lower tax rate was not cross-subsidised or financed by central Government, so that the economic consequences of these reductions are borne by the region itself?

Applying this test to the case at hand, the Advocate General found that while the measures at issue satisfied the requirement of institutional autonomy (as taken by a government body with its own constitutional and politico-administrative status), they did not satisfy the requirements of procedural and economic autonomy. In particular, the regional government was obliged, pursuant to the principle of national solidarity, to co-operate with the central Government in furthering the cause of redistribution across the Portuguese State. This principle also required budgetary transfers between the region and the State, to reduce economic inequalities. As a result, the regional tax measures were selective and, as they were not justified by the nature and general scheme of the taxation system, constituted State aid. In its judgment, the Grand Chamber essentially followed the Advocate General. The Court began by identifying the issue of selectivity as critical, and within this issue, the question of identifying the correct reference framework against which the tax measures at issue might validly be compared.33 In this regard, it explicitly rejected the Commission’s argument that the only legitimate reference framework was the totality of the Member State at issue; rather, in cases where an infra-State body enjoys a legal and factual status which makes it sufficiently autonomous in relation to the central government of a Member State, with the result that, by the measures it adopts, it is that body and not the central government which plays a fundamental role in the definition of the political and economic environment in which undertakings operate,34

the correct reference framework would be the area in which that infra-State body responsible for the measure exercises its powers. Approving the Advocate General’s tripartite classification of types of regionally-differentiated taxation (centrally set, symmetrically devolved and asymmetrically devolved), the Court

32 33 34

Ibid para 54. Ibid para 56. Ibid para 58.

The ECJ and the Devolution of Taxation Powers 257 accepted that only in the final case (asymmetrical devolution of taxation powers) was there a need for a test of autonomy, though the Court preferred to use the term ‘sufficient’ autonomy rather than the Advocate General’s ‘true’ autonomy.35 In relation to this test, it approved the Advocate General’s three-pronged formula of constitutional/politico-administrative, procedural and financial autonomy: In order that a decision taken in such circumstances can be regarded as having been adopted in the exercise of sufficiently autonomous powers, that decision must, first of all, have been taken by a regional or local authority which has, from a constitutional point of view, a political and administrative status separate from that of the central government. Next, it must have been adopted without the central government being able to directly intervene as regards its content. Finally, the financial consequences of a reduction of the national tax rate for undertakings in the region must not be offset by aid or subsidies from other regions or central government.36

It followed, therefore, that ‘sufficient’ autonomy would require not only that the regional body had the competence to adopt measures reducing the tax rate, but also that it assumed ‘the political and financial consequences’ of such measures.37 On the facts of that case, while the Azores’ constitutional/politico-administrative autonomy was not disputed, the Portuguese Government had not put forward any arguments to show that the region was not compensated financially for the lower tax rates applied, particularly given the possibility of budgetary transfers to less well-off regions expressly provided for pursuant to the principle of national solidarity under Portuguese law.38 As a result, the measure was selective. Neither was it justified on the basis of the nature or overall structure of the Portuguese tax system, as the tax reductions applied across the board within the Azores and were not dependent on ability to pay.39 As a result, Portugal’s action was dismissed. The Azores case generated a considerable amount of commentary and, in some instances, controversy. While many welcomed the introduction of a test setting out the parameters against which to measure the selectivity of devolved taxation powers, others were critical of certain aspects of the judgment, arguing that the ambiguity of the test would itself give rise to further litigation, but also that the judgment could mean circumvention of the State aid rules by enabling numerous small local authorities to distort competition by adopting slightly different taxation rates and benefits.40 While some criticism of any significant judgment is 35

Ibid paras 64–66. Ibid para 67. 37 Ibid para 68. 38 Ibid paras 73–76. 39 Ibid para 82. 40 See, eg, J Winter, Case note on Portugal v Commission (2008) 45 CML Rev 183, 195 (‘If tax reductions adopted by regional authorities possessing the requisite amounts of institutional, procedural and economic autonomy are considered to be general measures, the current trend towards greater decentralization or federalization in the Member States (Belgium?) is likely to lead to a further erosion of the effectiveness of the mechanism of Articles 87 and 88.’); B Kurcz, ‘How Selective is Selectivity? A Few Thoughts on Regional Selectivity’ (2007) 66 CLJ 313, 320 (‘Although such a test, in its intention, is supposed to reduce the legal uncertainty of the application of Article 87 EC, 36

258 Suzanne Kingston perhaps inevitable, it is difficult to avoid the conclusion that any attempt to set out a test in previously uncharted legal territory is bound to suffer, in the first instance, from a certain amount of ambiguity. No single judgment, based in a specific factual context, could or should reasonably hope to answer all possible issues on the application of a fundamental concept such as regional selectivity; rather, uncertainties in defining the concept are destined to be clarified, and evolve, in a casuistic manner applied to different factual contexts. This, indeed, is precisely what has happened in two subsequent judgments.

C. UGT-Rioja—Defining the Limits of Azores In UGT-Rioja,41 the ECJ got an opportunity to apply, and clarify, the test set down in the Azores judgment. The case concerned, once again, taxation laws passed by certain of the Historical Territories of the Basque Country within Spain which, inter alia, applied a lower corporation tax rate and a more favourable regime of tax deductions within such Territories than those applicable elsewhere in Spain.42 Despite the relatively high level of autonomy accorded to the Basque Country, and its three constituent Territories, under the Spanish constitutional settlement, the national court in its Order for Reference voiced uncertainty about whether the Azores test was satisfied. In relation to procedural autonomy, while no direct intervention in passing the regional laws took place by the central Government, a non-coercive conciliation mechanism existed to ensure that the law ultimately passed complied with the Economic Agreement between the Autonomous Community of the Basque Country and the Spanish State. The Economic Agreement laid down a variety of requirements to which the regional tax laws were subject (including, for instance, respect for the principle of solidarity laid down in the Spanish Constitution). In relation to economic autonomy, while the national court stated that the Basque Country was the responsible tax body for that region, nonetheless its economic autonomy was limited by the exclusive competences enjoyed by the State in certain areas such as the monetary system, social security and spending on public works.

one may wonder if this is achieved.’) For further commentary on the regional selectivity issue, see, eg P Nicolaides, ‘Fiscal State Aid in the EU: The Limits of Tax Autonomy’ (2004) 27 World Competition 365; P Nicolaides, ‘Developments in Fiscal Aid: New Interpretations and New Problems with the Concept of Selectivity’ (2007) 6 European State Aid Law Quarterly 43; S Moreno Gonzales, ‘Regional Fiscal Autonomy from a State Aid Perspective: The ECJ’s Judgment in Portugal v. Commission’ (2007) 47 European Taxation 328. 41

Joined Cases C-428/06 to C-434/06 UGT-Rioja and Others [2008] ECR I-6747. Namely, 32.5% compared to 35% corporation tax rate. The Territories at issue were Álava, Vizcaya and Guipúzcoa. 42

The ECJ and the Devolution of Taxation Powers 259 In its judgment, the Third Chamber repeated and confirmed the three-limbed test of autonomy set out in Azores,43 rejecting the Commission’s argument that any additional test of autonomy applied.44 Further, the fact that a region’s laws could be subject to procedures for judicial review to test their validity was not, in itself, such as to mean that region lacked authority.45 Applying the test to the case at hand, as with Azores the criterion of institutional autonomy was relatively easily met, as the Historical Territories and the Basque Country had a distinct political and administrative status from that of central government.46 Similarly, in relation to procedural autonomy, the fact that a conciliation process existed between regional and State governments did not prevent this condition from being satisfied, as long as the final decision was taken solely by the regional authority (even if that authority was obliged by law to take certain factors into account in so doing, such as the principle of solidarity).47 Once again, however, the final condition of the Azores test—economic and financial autonomy—raised more difficulties, with virtually all interveners arguing that the condition was satisfied, but the Commission contending the contrary. The Court used the opportunity to add clarification to this element of the test, underlining that the core of the matter was whether the financial consequences of a reduction of the national tax rate were offset by aid or subsidies from other regions or central government (ie, a causation-based approach).48 In order to ascertain whether this was so, the Court was forced to look in detail at the provisions of the Economic Agreement between the Autonomous Community of the Basque Country and the Spanish Government which provided, inter alia, for a complex quota system laying down the Basque Country’s financial contribution to the State to be specified by law on a periodic basis. On this point, the Court accepted the Basque and Spanish arguments that essential elements of such quota were determined by ‘political negotiations’ between the Basque Country and the Spanish State, and were not ‘necessarily’ influenced by a reduction in the Basque tax rate.49 Unsurprisingly, however, the Court held back from giving a definitive view on the interpretation of this law, preferring to leave this for the national court to decide50 as the court that had sole jurisdiction to ‘determine, interpret and apply the relevant national law and to examine the facts’.51 However, the Court gave some additional guidance in this regard, noting that financial compensation may not only be ‘declared and specific’, but may also be ‘hidden and result only from the actual examination of the financial flows

43 44 45 46 47 48 49 50 51

UGT-Rioja and Others, above n 41, paras 51–52. Ibid paras 53–60. Ibid paras 76–83. Ibid para 87. Ibid paras 95–110. Ibid para 123. Ibid para 127. Ibid paras 130–31. Ibid para 139.

260 Suzanne Kingston existing between the infra-State body concerned, the Member State which it comes under and the other regions of that Member State’.52 Nonetheless, agreeing with Advocate General Kokott on this point, the Court noted that the mere fact that financial transfers took place between the region and the State was not itself inconsistent with financial autonomy, as such transfers may have taken place for reasons other than the tax reduction.53 As noted by the Advocate General, in contrast to the Azores case, it was not obvious that the Basque Country had been given the power to reduce its tax rate in order to compensate for structural disadvantages for the region’s undertakings (which arose in the Azores case as a result of insularity).54 Prima facie, therefore, UGT-Rioja offered some relief to regions (and States) worried about the breadth of the Azores test in relation to economic and financial autonomy: only financial transfers from State to region caused by, and offered in compensation for, regional taxation reductions satisfy this test. As with the Azores judgment, therefore, this represents a concern to respect each Member State’s constitutional order, a point expressly highlighted by Advocate General Kokott in noting the significance of what is now Article 4(2) TEU which, as noted above, in the Nice version in force at the time required the EU to respect national identities.55 Upon closer examination, however, the Court’s careful qualification that compensation may be ‘hidden’ (for instance, in the form of transfers in sectors such as social security or minimum public service guarantees) means that such investigation will be highly complex and, ultimately, will necessitate making a call, at judicial level, on the reasoning behind delicate financial settlements achieved between region and State. In that case, it is again perhaps unsurprising that, upon referral back to the referring court, the Basque Supreme Court found its region to have sufficient autonomy within the meaning of the Azores test.56 This was so despite the fact that the Spanish Supreme Court had itself, some years previously, declared that the Basque tax rules constituted State aid, with the Basque Supreme Court holding that the intervening two judgments of the ECJ (Azores and UGTRioja) had changed the applicable legal framework.57 In this sense, the case highlights once again the potential significance of the procedural route by which a case ends up before the ECJ. In preliminary reference cases, such as UGT-Rioja, the ECJ will naturally tend towards leaving difficult questions of interpretation of national constitutional law to the referring court, which will often be the court of the region at issue (and perhaps, one might

52

Ibid para 133. Ibid para 135. 54 UGT-Rioja and Others, above n 41, Opinion of AG Kokott, para 61. 55 Ibid para 54. 56 Tribunal Superior de Justicia del Pais Vasco, judgments of 22 December 2008. 57 Tribunal Superior de Justicia del Pais Vasco, judgments of 17 November and 9 December 2004. The matter was unsuccessfully appealed to the Spanish Supreme Court. 53

The ECJ and the Devolution of Taxation Powers 261 think, be likely to consider its region to be autonomous).58 In contrast, in actions for annulment of a Commission decision, as was the case in Azores, the ECJ has no such luxury, and has to make the call on the national constitutional settlement itself. In the latter case, therefore, the written and oral submissions of the lawyers of the relevant State and region will be particularly crucial. Indeed, one sees this clearly in the fact that the ECJ in Azores expressly supported its conclusion that the Portuguese rules were selective on the fact that Portugal had not made submissions rebutting the argument that lower tax rates were compensated for by financial transfers.

D. Subsequent Cases: Gibraltar and Regione Sardegna Three months after the ECJ handed down its UGT-Rioja judgment, the CFI gave judgment in Gibraltar and United Kingdom v Commission,59 an action brought by Gibraltar to annul a 2004 Commission decision finding the proposed reform of Gibraltar’s corporation tax system, which had been notified by the UK to the Commission, to constitute State aid that was incompatible with the common market.60 This decision followed an unsuccessful attempt by the Commission to initiate State aid proceedings in relation to Gibraltar’s previous corporate tax rules, which attempt was annulled by the CFI.61 At issue in the present case, however, were the reforms applying a top-up penalty tax solely to companies in the financial services sector and to utilities. This tax was to be in addition to a payroll tax of £3,000 per employee per year, a business property occupation tax and a registration fee. In terms of constitutional status, Gibraltar is a Crown colony (or British Overseas Territory). Although the UK is responsible for its external relations (meaning that the EU Treaties, including the competition and State aid rules, apply to it),62 it is not part of the UK and has its own Constitution.63 However, its Governor is appointed by, and a representative of, the Queen. It has its own legislature (constituted by its House of Assembly and Governor) and courts, although appeal from the highest court goes to the Judicial Committee of the Privy Council of the UK. In the present case, the legislation implementing the reforms was to be 58 It should be noted that the wording of the question referred in fact asked the ECJ to decided itself whether or not the relevant measures were selective; the ECJ, as it often does, avoided giving a specific answer by reformulating the question on the basis that it impinged on the division of competences in Article 267 TFEU proceedings between the ECJ and national courts. 59 Joined Cases T-211/04 and T-215/04 Gibraltar and United Kingdom v Commission [2008] ECR II-3745. 60 Commission Decision 2005/261/EC [2005] OJ L85/1. 61 Joined Cases T-195/01 and T-207/01 Gibraltar v Commission [2002] ECR II-2309. 62 See ex Art 299(4) EC. 63 See further, Case C-145/04 Spain v United Kingdom [2006] ECR I-7917 (legality of right to vote in European Parliament elections for Commonwealth citizens residing in Gibraltar but not EU citizens).

262 Suzanne Kingston passed by Gibraltar’s House of Assembly and implemented by the Government of Gibraltar. While a variety of arguments were raised by Gibraltar in support of its action for annulment, the most relevant for our purposes was the first, arguing that the Commission had wrongly concluded that Gibraltar’s tax regime was regionally selective, on the basis that the UK was the appropriate reference framework. Interestingly, while Spain had quite foreseeably (given Gibraltar’s proximity, the economic threat it poses, and the political tensions it arouses between Spain and the UK) intervened in support of the Commission’s defence of its decision, it was forced to tread a fine line in distancing itself from the decision’s stance on regional selectivity, given the UGT-Rioja case ongoing at the time. In its judgment, the CFI found in favour of Gibraltar on this point, after carefully considering the implications of the Azores judgment for the case. As the ECJ had done in UGT-Rioja in response to a similar argument, the CFI rejected the Commission’s attempts to argue that Azores effectively imposed a fourth condition (in addition to those of institutional, procedural and economic autonomy), namely, that the infra-State body must occupy a fundamental role in the definition of the political and economic environment in its territory.64 Further, the three Azores conditions were, in the present case, met: (a) institutional autonomy, which was not disputed by the principal parties; (b) procedural autonomy, as executive competence in relation to corporate tax lay with Gibraltar’s Council of Ministers (although the UK retained a residual power of last resort to legislate for Gibraltar, this power had been exercised only exceptionally and never in tax matters);65 (c) economic autonomy as, just as the ECJ had held in UGT-Rioja, Azores required there to be a causal link between the regional tax measure and the financial support from central government, which was not the case here.66 Although the CFI made no reference to the ECJ’s UGT-Rioja judgment, therefore, it is clear that there are strong parallels between the reasoning in both cases, and one might imagine that while the ECJ’s judgment was not published in time to allow reference to be made to it, informal consultation may have taken place between the two courts. The CFI’s judgment is currently, however, on appeal before the ECJ, with Advocate General Jääskinen handing down his Opinion on 7 April 2011 and endorsing the CFI’s approach to regional selectivity in a virtually wholesale manner.67 Of particular interest is the ground of appeal put forward by Spain, raising

64 Gibraltar and United Kingdom v Commission, above n 59, paras 87–88; UGT-Rioja, above n 41, paras 53–60. 65 Gibraltar and United Kingdom v Commission, above n 59, paras 93–95. 66 Ibid paras 106–07. 67 Joined Cases C-106/09 P and C-107/09 P Commission and Spain v Gibraltar and United Kingdom, pending.

The ECJ and the Devolution of Taxation Powers 263 the issue of what scope of review the ECJ should have on appeal of the CFI’s appraisal of the applicable national law. On this point, it is difficult to dispute the Advocate General’s conclusion that such appraisal should be considered a question of fact, meaning that the ECJ’s review is confined to manifest error.68 If followed by the ECJ, this will, given the key importance of interpretation and characterisation of national constitutional provisions in regional selectivity cases, mean the scope for successful appeals from CFI (now General Court) judgments in this field will be substantially reduced. While Gibraltar may illustrate the significance of Azores,69 a further recent case, Presidente del Consiglio dei Ministri v Regione Sardegna, demonstrates its limitations and that it does not constitute carte blanche for regions to argue that their regional tax system falls outside the scope of the EU State aid rules.70 The case concerned the legality of Sardinian regional legislation establishing a tax on tourist stopovers by aircraft and pleasure boats used for private transport or recreation, but which was imposed only on operators tax resident outside Sardinia. In arguing that the tax fell outside Article 107(1) TFEU, the Region of Sardinia attempted to rely on Azores to conclude that, as an autonomous region, its tax rules were not selective. The ECJ was quick to reject this argument, noting that, as the tax did not apply to all operators making a stopover in Sardinia, but only to those not tax resident in Sardinia, it was clearly materially selective and could not be saved by the Azores principles.71

III. CONCLUSION

In being asked to define the boundaries of the concept of regional selectivity in the definition of State aid, the ECJ has been forced to consider carefully the practical implications of the Union’s commitment to respect national identities and, as part of this, national attitudes to regional autonomy and subsidiarity. As is evident from the early Opinion of Advocate General Saggio in Juntas Generales de Guipúzcoa, and the arguments of the Commission in cases like Azores, it has not always been clear how the Court might react to the suggestion that a State could validly be ‘split’ into geographic regions when determining the reference framework in assessing selectivity. The Court’s positive answer to such suggestion in Azores reflects a mature approach to the application of the State aid rules, and one unwilling to ignore the realities of national constitutional settlements out of fears that this could 68 Commission and Spain v Gibraltar and United Kingdom, above n 67, Opinion of AG Jääskinen, para 48. 69 See also Case T-75/03 Banco Comercial dos Açores v Commission [2009] ECR II-143 (applying Azores in the same constitutional context as the original judgment). 70 See n 11 above. See also, eg Case C-537/08 P Kahla Thüringen Porzellan GmbH (ECJ, 16 December 2010) (aid granted by, amongst others, a German Land contrary to Art 107 TFEU). 71 Presidente del Consiglio dei Ministri v Regione Sardegna, above n 11, paras 60–66.

264 Suzanne Kingston lead to the circumvention of the State aid rules by Member States. This maturity of approach to the issue of regionalism finds obvious parallels in non-State aid judgments such as Horvath.72 In this sense, the Court’s Azores judgment displays a confidence in its ability to devise a test that distinguishes genuine situations of sufficiently autonomous regions from others, and, importantly, a confidence in its own ability to assess and make a call on the interpretation of national constitutional law (on the basis, of course, of the parties’ submissions).73 In reality, as demonstrated by subsequent judgments, the Azores principles will only ‘save’ tax measures imposed by genuinely autonomous regions, which will exclude the vast majority of regional taxes from the ambit of the judgment.74 Ultimately, of course, the State aid rules are by no means the only tool at the Commission’s disposal in eliminating State-caused competitive distortions, and difficulties arising from distortive autonomous regional tax systems may well, depending on their nature, fall to be tackled by another means (in particular, the free movement provisions or Articles 115–117 TFEU). Nonetheless, the ECJ’s approach in the State aid context represents a welcome affirmation of its ability to adopt a dynamic interpretation of fundamental concepts of EU law in recognising, respecting and accommodating the complex and varied nature of national, and regional, identities.

72 Case C-428/07 Horvath v Secretary of State for Environment, Food and Rural Affairs [2009] ECR I-6355 (the mere adoption by devolved administrations of differing Good Agricultural and Environmental Condition (GAEC) standards pursuant to Regulation No 1782/2003 did not constitute unlawful discrimination ‘where the constitutional system of a Member State provides that devolved administrations are to have legislative competence’ (para 58)). 73 Commission and Spain v Gibraltar and United Kingdom, above n 67, Opinion of AG Jääskinen, para 47. 74 See, eg Joined Cases T-254/00, T-270/00 and T-277/00 Hotel Cipriani [2008] ECR II-3269, para 175 (regional selectivity of social security alleviation for Venice and Chioggia undisputed).

12 The Impact of EU Law on the Devolution of Social Powers in the Member States HERWIG VERSCHUEREN*

I. INTRODUCTION

T

HIS CHAPTER EXAMINES the impact of European Union (EU) law on the devolution of social powers in the EU Member States and more specifically on the relationships between the sub-national entities of a Member State with devolved competences in the field of social protection. Several Member States with a regionalised structure have introduced a residence-based distribution of competence between sub-national entities in this field. In Member States like Belgium, Germany, Spain, Austria and the United Kingdom, regions have some power to introduce elements of social protection and have created social benefits, in most cases limited to persons living on the territory of the relevant region. Furthermore, in both regionalised and non-regionalised Member States, local authorities have also introduced some social benefits, albeit usually limited to residents of their territory. Hence, internal rules for the distribution of powers commonly delimit regional and local circles of solidarity on the basis of a place-of-residence criterion. At first sight, EU law does not seem to impact on Member States’ internal rules regarding the distribution of social powers between their regions and the definition of internal circles of solidarity. Yet, in this chapter, we shall see that EU law on the free movement of persons may indeed influence these matters. The starting point of the analysis is that the EU allows its Member States great freedom in developing their own social protection systems. This is, first and foremost, a competence of the Member States themselves. However, we enter the realm of EU law as soon as we consider cross-border situations involving two or several Member States. The European legislative framework provides a number of solutions for such situations. If an individual has ties with more than one Member

* Professor of International and European Social Law at the University of Antwerp and Visiting Professor at the University of Brussels (VUB).

266

Herwig Verschueren

State, confusion may arise regarding the question of which Member State’s social protection schemes apply to that individual. Such cross-border situations may present themselves not only in nationally organised systems of social protection, but also in social protection schemes organised by sub-national entities of a Member State. In section II. of this chapter, we shall consider the solutions that the European legal framework provides for situations that are cross-border between Member States. These solutions are expressly laid down in secondary EU law. They are based on the principles of the free movement of persons under the European Treaties and have been the subject of extensive case law of the Court of Justice of the European Union (CJEU). European Union law has thus delineated the boundaries of circles of solidarity between the Member States in order to clarify, on behalf of EU citizens in cross-border situations, which Member State’s social protection schemes apply. In other words, the territorial and personal boundaries of the circles of solidarity of the Member States are, in principle, drawn by EU law, which aims to guarantee the free movement of persons between the Member States. In section III., we shall examine how these European rules affect systems of social devolution developed (or under development) in Member States, particularly in relation to the execution of regional (and local) competences in the field of social protection. The key questions in this respect concern: (a) the extent to which Member States, in their internal legal distribution and execution of regional and local competences in social protection, must take account of the European rules; and (b) the extent to which such internal arrangements may deviate from the legal provisions agreed upon at EU level. In other words: What are the contours of social devolution in the Member States under EU law? And to what extent are these contours a constraint on the ability of regionalised Member States autonomously to distribute internal powers in the field of social protection between different levels of government?

II. THE EUROPEAN STRAND OF SOCIAL FEDERALISM

A. No Separate European Social Protection The starting point of EU law and policy is that the Member States are entirely free to organise their internal social protection systems as they see fit. In so far as social protection is concerned, the European Treaties assign hardly any harmonising powers to the European institutions. Article 153(1)(c) TFEU (ex Article 137 TEC) does entitle the Union to support and complement the activities of the Member States in the field of ‘social security and social protection of workers’, including by means of directives laying down minimum requirements. However, the Union

Impact of EU Law on Devolution of Social Powers in Member States

267

has never used this competence, neither is it likely to do so in the short to medium term. The only harmonising initiatives that have been taken are a number of directives concerning the equal treatment of men and women in matters of social security.1

B. The Impact of the European Principle of Free Movement of Persons on Social Protection Schemes in the Member States i. European Social Security Co-ordination These starting points do not prevent EU law from indirectly affecting national legislation in the field of social protection. This impact is due primarily to the principle of the free movement of persons. Additionally, the social policies of the Member States are affected indirectly by European rules relating to the free movement of goods and services, as well as by competition rules, including the prohibition of State aid. However, in the present contribution we shall leave aside the impact of these latter aspects of EU law, as the focus of our study is on questions relating to the personal and territorial boundaries of solidarity schemes in Member States and their sub-national entities.2 Most tangible of all in this respect is European social security co-ordination. Such co-ordination is necessary with a view to ensuring the right to free movement of persons. The right of European citizens to move and reside freely within the territory of the Member States, to seek employment, to work, to pursue selfemployed activities or to provide services in another Member State is guaranteed under the European Treaties themselves.3 The purpose is to co-ordinate the social security systems of the Member States in such a way as to eliminate any negative consequences for the migrating individual that may arise from differences between the various systems. To this end, the European legislature has worked out an extensive co-ordination system. Until 1 May 2010, the applicable regulations were Regulation (EEC) 1408/714 1 See, eg Council Directive (EEC) 79/7 on the progressive implementation of the principle of equal treatment for men and women in matters of social security [1979] OJ L6/24. Additionally, we refer to the abundant case law of the CJEU prohibiting discrimination on the basis of gender in relation to rights stemming from occupational social security systems, such as supplementary pensions rights. See, eg C Barnard, EC Employment Law (Oxford, Oxford University Press, 2006) 488–535. 2 On the impact of European internal market and competition law, see among others the following recent publications: M Krajewski, U Neergaard and J van de Gronden (eds), The Changing Legal Framework for Services of General Interest in Europe (The Hague, Asser Press, 2009); E Mossialos, G Permanand, R Baeten and T Hervey (eds), Health Systems Governance in Europe. The Role of EU Law and Policy (Cambridge, Cambridge University Press, 2010). 3 Arts 21, 45, 49, 56 TFEU (ex Arts 18, 39, 43 and 49 TEC). 4 Council Regulation (EEC) 1408/71 concerning the application of the social security schemes to employees and self-employed persons, as well as to their family members travelling within the Community (‘Regulation 1408/71’). This Regulation was amended on several occasions. A consolidated version may be found on Eur-lex at .

268

Herwig Verschueren

and Regulation (EEC) 574/72.5 These Regulations were replaced on 1 May 2010 by Regulation 883/20046 and Regulation 987/2009.7 It is not the intention of this co-ordination system to in any way harmonise or approximate the systems of the Member States. It does not touch upon the material and formal differences between the social security regimes of the various Member States and hence it does not affect the divergent rights of the individuals working or residing in those States.8 This implies among other things that (labour) migration between the Member States may give rise to more extensive or less extensive social protection depending on the system that is in place in the Member State where the individual concerned is working or residing.9 ii. Determination of the Legislation Applicable in Cross-border Situations: State of Employment and State of Residence One of the most important tasks of the co-ordination system is to determine the legislation applicable in cross-border situations. The relevant rules are contained in Title II of Regulation 1408/71 and Regulation 883/2004. These provisions are intended not only to prevent the simultaneous application of several national legislative systems and the complications that might ensue, but also to ensure that a person in a cross-border situation between Member States is not left without social security coverage because there is no legislation applicable to him.10 The person in question shall be subject to the legislation of a single Member State only,

5 Council Regulation (EEC) 574/72 laying down the procedure for implementing Regulation 1408/71. This Regulation was also amended on several occasions. A consolidated version may be found on Eur-lex, above n 4. 6 European Parliament and Council Regulation (EC) 883/2004 on the co-ordination of the social security systems [2004] OJ L200/1 (‘Regulation 883/2004’), as amended by European Parliament and Council Regulation (EC) 988/2009 [2009] OJ L284/43. For an introduction to this new Regulation, see the contributions in (2009) 1–2 European Journal of Social Security. Special Issue on 50 Years of European Social Security Coordination 1–241. See also F Pennings, European Social Security Law (Antwerp, Intersentia, 2010). 7 European Parliament and Council Regulation (EC) 987/2009 laying down the procedure for implementing Regulation (EC) 883/2004 on the co-ordination of social security systems [2009] OJ L284/1. 8 See, eg, Case 41/84 Pietro Pinna v Caisse d’allocations familiales de la Savoie [1986] ECR 1, para 20; Case C-340/94 de Jaeck v Staatssecretaris van Financiën [1997] ECR I-461, para 18; Case C-493/04 Piatkowski v Inspecteur van de Belastingdienst grote ondernemingen Eindhoven [2006] ECR I-2369, paras 19–20; Case C-208/07 von Chamier-Glisczinski v Deutsche Angestellten-Krankenkasse [2009] ECR I-6095, para 84. 9 See, eg, Cases C-393/99 and C-394/99, Institut national d’assurances sociales pour travailleurs indépendants (Inasti) v Claude Hervein and Hervillier SA and Guy Lorthiois and Comtexbel SA [2002] ECR I-2829, paras 50–51; Piatkowski, above n 8, para 34; von Chamier-Glisczinski, above n 8, para 85. 10 See, eg Case C-196/90 Fonds voor Arbeidsongevallen v De Paep [1991] ECR I-4815, para 18; Case C-275/96 Kuusijärvi v Riksförsäkringsverket [1998] ECR I-3419, para 28; Case C-202/97 Fitzwilliam Executive Search Ltd v Bestuur van het Landelijk instituut sociale verzekeringen [2000] ECR I-883, para 20; Case C-347/98 Commission v Belgium [2002] ECR I-3327, para 27.

Impact of EU Law on Devolution of Social Powers in Member States

269

and that legislation is to be determined in accordance with the provisions of the 2004 Regulation.11 In so far as the determination of the applicable legislation is concerned, the governing principle is that of the State of employment (lex loci laboris). It means that a person employed in the territory of one Member State shall be subject to the social security legislation of that State, even if he resides in the territory of another Member State or if his employer is registered in another Member State.12 The choice of this principle is inspired by, among other things, the legal context in which European social security co-ordination is applied: the principle’s aim is to ensure the free movement of workers. Such freedom of movement entails a prohibition of discrimination based on nationality by the Member State where the migrating worker is employed (Article 45(2) TFEU). This prohibition holds not only for all conditions of employment and remuneration that apply in the State of employment, but also for social security provisions.13 Hence, the State-ofemployment principle is an expression of the premise that a migrating worker is entitled to the same rights in the State of employment as workers of that Member State.14 In the Flemish care insurance judgment, the CJEU appears to confirm that the lex loci laboris principle, which constitutes the basis for European social security co-ordination, is already contained in the Treaty provisions relating to the freedom of movement of employed and self-employed persons.15 Nonetheless, the application of the State-of-employment principle does not imply that a migrant worker is invariably entitled to social security benefits provided in the State of employment. If such a person resides in another Member State (eg, frontier workers) then, for certain benefits (eg, coverage of medical costs), he or she shall be entitled first and foremost in accordance with the legislation of the State of residence (albeit at the expense of the State of employment).16 Moreover, the CJEU has confirmed that the application of the State-of-employment principle does not preclude a migrant worker from entitlements pursuant to the

11

Regulation 883/2004, Art 11(1). Regulation 883/2004, Art 11. 13 See, eg Pinna, above n 8. 14 The relevance of the principle of equal treatment to the rules governing the determination of the applicable legislation was underlined by the CJEU in, among other judgments: Case C-68/99 Commission v Germany [2001] ECR I-1865, paras 22–23; Inasti, above n 9, para 31; Piatkowski, above n 8, para 19; Case C-103/06 Derouin v Union pour le recouvrement des cotisations de sécurité sociale et d’allocations familiales de Paris—Région parisienne (Urssaf de Paris—Région parisienne) [2008] ECR I-1853, para 20. See also recital 17 of the preamble to Regulation 883/2004, stating that: ‘With a view to guaranteeing the equality of treatment of all persons occupied in the territory of a Member State as effectively as possible, it is appropriate to determine as the legislation applicable, as a general rule, that of the Member State in which the person concerned pursues his/her activity as an employed or self-employed person.’ 15 Case C-212/06 Government of the French Community and Walloon Government v Flemish Government [2008] ECR I-1683, para 48. 16 See Art 17 of Regulation 883/2004. 12

270

Herwig Verschueren

national legislation of the State of residence. The State of residence is not required to grant such social security rights, but EU law does not exclude it.17 In addition, the fact that the State-of-employment principle is the starting point for economically active persons does not prevent the application of the social security system of the State of residence in a number of situations, more particularly if those persons are simultaneously active in more than one Member State. The provisions of Regulation 1408/71 and Regulation 883/2004 then refer to the State of residence, provided that occupational activities are pursued in that country. In a number of other cases, it is the registered office or the place of business of the employing undertaking that determines to which country’s social security legislation the person in question is subject.18 Moreover, as a rule, the State-of-residence principle applies to those who are not (or no longer) economically active.19 The category of economically inactive persons is understood to include pensioners.20 Unemployed frontier workers are likewise subject to the legislation of the State of residence.21 That is not to say, though, that inactive persons are exclusively entitled to social security benefits under the legislation of their State of residence. Retired persons who do not receive their pension from their State of residence, for example, are also subject to the legislation of the State that does pay their pension in so far as sickness benefits in cash, such as long-term care benefits, are concerned. Consequently these Member States must export those benefits.22 Regulation 833/2004 also contains a number of rules relating to the payment of benefits to persons who do not reside in the territory of the Member State under whose legislation they are entitled to benefits (the so-called export of benefits).23 This holds first and foremost for benefits like old-age pensions or invalidity and long-term care benefits. There are, however, a number of exceptions to the export principle. One such exception is the special co-ordination regime that applies for so-called special non-contributory benefits that are situated in between social security and social assistance. For these kinds of benefits, a State-of-residence-based co-ordination system is in place. Another exception is unemployment benefits, which only need to be exported for a maximum of three months if the benefit-entitled person moves to another Member State in order to seek employment.24 In other words, there is in effect a residence requirement for unemployment benefit entitlement.

17 Case C-352/06 Bosmann v Bundesagentur für Arbeit—Familienkasse Aachen [2008] ECR I-3827, paras 27–29 and 33; von Chamier-Glisczinski, above n 8, paras 55–56. 18 Regulation 883/2004, Art 13. 19 Ibid Art 11(3)(e). 20 Ibid Art 11(2). 21 Ibid Art 11(3)(c). 22 Ibid Art 29. 23 Ibid Art 7. 24 Ibid Art 64.

Impact of EU Law on Devolution of Social Powers in Member States

271

Moreover, in situations that are beyond the scope of these specific European co-ordination rules (or other rules of secondary EU law), the CJEU has always sought the most appropriate basis for delimiting the circles of solidarity of the Member States. In such instances, the Court commonly refers directly to the Treaty provisions on free movement of persons. In some cases, particularly those relating to economically active individuals, the CJEU has granted rights pursuant to the legislation of the State of employment (or former State of employment).25 In others, particularly those concerning economically inactive persons, it has granted rights in accordance with the legislation of the State of residence.26 And in yet some other cases, the Luxembourg Court has recognised the right of the individuals concerned to export to a Member State other than the State of residence a social benefit that is not within the scope of the European co-ordination system.27 It appears from the above analysis that European legislation and case law have tried to resolve the matter of determining to which circle of solidarity a person migrating within the European Union belongs. The answer to this question depends on the circumstances. In the case of economically active persons, the starting point is the State-of-employment principle, whereas in the case of those who are not (or no longer) economically active, it is the State-of-residence principle. However, these starting points are qualified in both legislation and case law: in some situations involving economically active persons, the legislation of the State of residence is applicable, while in specific cases involving economically inactive persons, the legislation applicable is that of the former State of employment. Invariably, the European legislature and judges appear to have tried to establish with which Member State the person concerned is linked most closely from a socio-economic perspective. Even though social protection is an almost exclusive competence of the Member States, one discerns a form of European ‘social federalism’ in this body of regulations and legislation, in the sense that it is for the European lawmaker to determine to which circle of solidarity persons migrating within the EU belong. In setting the boundaries of these circles, however, Europe struggles with the question of which criteria to apply, and particularly with the choice between the State-of-employment and the State-of-residence principles. At present, it relies

25

Case C-35/97 Commission v France [1998] ECR I-5325. Case C-184/99 Grzelczyk v Centre public d’aide sociale d’Ottignies-Louvain-la-Neuve [2001] ECR I-6193; Case C-456/02 Trojani v Centre public d’aide sociale de Bruxelles (CPAS) [2004] ECR I-7573. However, in the case of benefits with characteristics of social assistance, the Court has, in a number of judgments, recognised that the host country where the migrant citizen is residing may ascertain whether there is a real link between the claimant and this country or its employment market. See Case C-138/02 Collins v Secretary of State for Work and Pensions [2004] ECR I-2703; Case C-158/07 Förster v Hoofddirectie van de Informatie Beheer Groep [2008] ECR I-8507. 27 Case C-499/06 Nerkowska v Zakład Ubezpiecze Społecznych Oddział w Koszalinie [2008] ECR I-3993; Case C-221/07 Zablocka-Weyhermüller v Land Baden-Württemberg [2008] ECR I-9029. 26

272

Herwig Verschueren

on a complex and intricate combination of these two principles, each of which is, moreover, applied in a qualified way and with room for exceptions.

III. THE IMPACT OF EUROPEAN ‘SOCIAL FEDERALISM’ ON SOCIAL DEVOLUTION IN THE MEMBER STATES

In the Member States, rights and obligations in respect of social protection are usually laid down at the federal level. Still, that is not to say that sub-national entities (even at the local level) cannot have powers in the field of social policy. In the domestic context, social devolution concerns, first, the question of what policy-making levels are empowered to take what initiatives in the field of social protection. Once the federal, regional and local powers in a Member State have been fixed, the question arises as to how to define the boundaries of the circles of solidarity of the comprising entities. What criteria should be applied in determining to which sub-national entity’s circle of solidarity an individual belongs? We leave aside how the various regionalised Member States of the European Union have answered this question.28 Instead, we focus in this chapter exclusively on whether EU law affects how Member States delimit their circles of solidarity internally. Are regionalised Member States left entirely free to lay down in internal legislation the personal and territorial scope of application of social protection schemes developed by that State’s sub-national entities? Or is the European system of defining circles of solidarity a model that can or even must be adopted by regionalised Member States in the organisation of relationships between their sub-national entities?

A. Can EU Law Intervene in a Regionalised Member State’s Internal Distribution of Competences in the Field of Social Protection? The delineation of the circles of solidarity of the various sub-national entities comprising a Member State would initially appear to be a matter which is purely internal to the Member State in question and to its constitutional order. However, since the CJEU’s judgment in the Flemish care insurance case, this is no longer a foregone conclusion.29 The Flemish care insurance scheme was intended to cover

28 On the Belgian case, see B Cantillon, P Popelier and N Mussche (eds), Social Federalism: The Creation of a Layered Welfare State. The Belgian Case (Antwerp, Intersentia, 2011). 29 Government of the French Community and Walloon Government, above n 15. For a detailed discussion of this judgment, see C Dautricourt and S Thomas, ‘Reverse Discrimination and Free Movement of Persons under EU law: All for Ulysses, Nothing for Penelope?’ (2009) 34 EL Rev 433; P Van Elsuwege and S Adam, ‘Situations purement internes, discriminations à rebours et collectivités autonomes après l’arrêt sur l’assurance soins flamande’ (2008) Cahiers de droit européen 655; H Verschueren, ‘La régionalisation de la sécurité sociale en Belgique à la lumière de l’arrêt de la Cour de Justice européenne portant sur l’assurance soins flamande’ (2008) 2 Revue belge de sécurité sociale 173.

Impact of EU Law on Devolution of Social Powers in Member States

273

the cost of non-medical assistance and services for persons who were unable to perform daily tasks necessary for their basic needs or other related activities.30 Only persons residing in the territory of Flanders were (compulsory) affiliated to this scheme. They were obliged to pay an annual contribution and were entitled to a monthly benefit, provided they fulfilled the conditions of being in need of assistance. Affiliation to the Flemish care insurance scheme was optional for persons residing in Brussels. It was a typical example of a residence-based regional social security benefit, limited to persons residing on the territory of a specific region within a regionalised Member State. However, this residence condition was in conflict with the abovementioned requirements of the EU social security co-ordination, more particularly with the State-of-employment principle. At the request of the European Commission, the Flemish legislature amended the decree in order to bring it in line with this EU principle.31 The applicability of the State-of-employment principle to economically active persons entailed that, in a European cross-border context, the Flemish care insurance scheme must also comply with the principle. More specifically, employees or self-employed persons who lived in Flanders (or Brussels) but worked in another Member State could not be compelled to join the Flemish care insurance scheme. Moreover, employees and self-employed persons who lived in a Member State other than Belgium but worked in Flanders (or Brussels) could not be excluded from the Flemish care insurance scheme on the basis of the fact that they did not reside in Flanders (or Brussels). In addition, the principle of the export of benefits implied that persons working in Flanders (or Brussels) but residing in another Member State were entitled to this benefit, so that the residence requirement laid down in the Flemish legislation was discarded. As a result, persons working in Flanders (or Brussels) but residing in another Member State fell within the ambit of the Flemish care insurance scheme. However, persons working in Flanders (or Brussels) but living in Wallonia remained excluded. The Flemish legislature assumed the latter situation to be a purely internal one, to which EU law did not apply. For that reason, the amendments to the Care Insurance Decree did not address cross-border situations between Flanders/Brussels and another Community in Belgium (ie, in practice, Wallonia). Yet the amendments were challenged before the Belgian Constitutional Court as being discriminatory against persons residing in Wallonia. In these proceedings, the Belgian Constitutional Court referred a number of questions to the CJEU for a preliminary ruling. The Constitutional Court wished to learn, in particular, whether the exclusion from the Flemish care insurance scheme of persons working in Flanders or Brussels but living in Wallonia was contrary to the provisions of Regulation 1408/71 and to the Treaty provisions regarding the free movement of persons. The Court made reference to the free 30 Decree of the Flemish Community of 30 March 1999 concerning the organisation of the care insurance scheme. 31 This amendment was made by the Decree of 30 April 2004.

274

Herwig Verschueren

movement of workers (Article 39 EC; now Article 45 TFEU), self-employed persons (Article 43 EC; now Article 49 TFEU), and EU citizens in general, irrespective of whether they are pursuing an economic activity (Article 18 EC; now Article 21 TFEU).32 The case was essentially about the impact of EU law on the delimitation of circles of solidarity in the context of the relationship between subnational entities within a given Member State. In its answer, the CJEU first reiterated that EU law cannot be applied to purely internal situations.33 At the same time, however, it defined the notion of a ‘purely internal situation’ strictly. According to the Court, this category consisted of ‘Belgian nationals working in the territory of the Dutch-speaking region or in that of the bilingual region of Brussels-Capital but who live in the German- or French-speaking region and have never exercised their freedom to move within the European Community’.34 Hence, the CJEU did not rule out that certain EU citizens living in Wallonia but working in Flanders (or Brussels) nonetheless fell within the ambit of EU law. The Court referred, first and foremost, to nationals of Member States other than Belgium working in the Dutch-speaking region or in the bilingual region of Brussels-Capital but living in another part of the national territory, but also to Belgian nationals in the same situation who had made use of their right to free movement.35 It was in this context that the CJEU considered the residence requirement in the Care Insurance Decree. It held that the requirement presented a potential obstacle to the free movement of employed and self-employed persons. In the Court’s view, migrant workers pursuing (or contemplating the pursuit of) employment or self-employment in Flanders or Brussels might be dissuaded from making use of their freedom of movement. Moving from their Member State of origin to certain parts of Belgium would cause them to lose the opportunity of eligibility for the benefits which they might otherwise have claimed. In other words, the Court argued that the fact that workers find themselves in a situation in which they suffer either the loss of eligibility for care insurance or a limitation of the place to which they transfer their residence is, at the very least, capable of impeding the exercise of the right to free movement conferred by EU law.36 Moreover, the Court found this impediment to be unjustified. It rejected the Flemish government’s argument that the non-applicability of the Flemish care insurance scheme to residents of Wallonia was due to the requirements inherent in the distribution of powers within the Belgian federal structure and, particularly, to the fact that the Flemish Community had no power in relation to care insurance vis-à-vis persons residing in the territory of other linguistic

32 33 34 35 36

Belgian Constitutional Court, No 51/2006, 19 April 2006. Government of the French Community and Walloon Government, above n 15, para 38. Ibid paras 37–38. Ibid para 41. Ibid para 48.

Impact of EU Law on Devolution of Social Powers in Member States

275

communities of Belgium.37 The position of the CJEU implies that Member States must ensure that their internal legal distribution of powers in relation to social security does not impede the exercise of the right to free movement between Member States. Thus, in the Court’s opinion, the principle of the free movement of persons within the European internal market takes precedence over the internal constitutional organisation of a Member State. The general nature of the Court’s position entails that all regional and local authorities which are empowered to confer social benefits are precluded from imposing a requirement of residence in their region/municipality upon migrant EU workers or self-employed persons (and their families) who wish to work there, unless there is an adequate justification for that prerequisite.38 The above also implies that, in so far as the relationship between sub-national entities of a Member State is concerned, a distinction is to be made between two categories of European citizens: those with recourse to EU law on grounds of their having exercised their right to free movement as workers or self-employed persons, and those without recourse to EU law. This gives rise to what we might refer to as ‘reverse discrimination’.39 As a consequence, the majority of those living in Wallonia but working in Flanders (or Brussels) are unable to invoke EU law to claim eligibility for Flemish care insurance. The Court expressly rejected the contention that these persons could resort to the principle of citizenship of the Union set out in Article 17 EC (now Article 20 TFEU), which includes, in particular, the right of every citizen of the Union to move and reside freely within the territory of the Member States (Article 18 EC, now Article 21 TFEU). More particularly, the Court held that Union citizenship is not intended to extend the material scope of the Treaty to internal situations that have no link with EU law.40 The CJEU nevertheless remarked that its interpretation of the provisions of EU law might be of use to the Belgian Constitutional Court, even as far as purely internal situations are concerned.41 The CJEU appeared to be implicitly referring to the general principle of non-discrimination laid down in Articles 10 and 11 of the Belgian Constitution. Yet the Belgian Constitutional Court did not accept this

37

Ibid paras 57–58. An example of such a justification is found in the judgment of the EFTA Court of 3 May 2006 in Case E-3/05, EFTA v Norway (see ). This case concerned the application of a residence requirement for eligibility for a regional supplement to Norwegian family allowance in the northern regions of Norway (Finnmark and part of Troms). The purpose of the supplement is to reverse the depopulation trend in these areas and to provide additional support to residents with children, among other things to compensate for the harsh living conditions. Eligibility is conditional upon the family actually living in the region. People working in the region but living elsewhere with their family are not entitled to the supplement. The EFTA Court ruled that the regional residence condition for this social security benefit is legitimate given the overall aim, and that it is proportional to this aim and therefore not in breach of Regulation 1408/71. 39 On the issue of reverse discrimination, see further below. 40 Government of the French Community and Walloon Government, above n 15, para 39. 41 Ibid para 40. 38

276

Herwig Verschueren

suggestion in its ensuing ruling of 21 January 2009.42 The Constitutional Court did acknowledge that the ambit of the Flemish Care Insurance Decree should, pursuant to EU law,43 be extended to persons living in Wallonia but working in Flanders (or Brussels) and who were either nationals of another Member State or Belgian nationals who had exercised their right to move freely within the EU.44 In this context, the Constitutional Court also took explicit account of the fact that only a relatively small group of individuals would benefit from this extension.45 However, as regards persons who were unable to invoke EU law, the Constitutional Court reaffirmed the exclusively territorial distribution of powers between Belgium’s different Communities. It concluded that the Flemish Care Insurance Decree was not applicable to Belgians living in Wallonia and working in Flanders (or Brussels) who had never exercised their right to free movement within the EU. According to the Constitutional Court, the fact that these residents were consequently not eligible for care insurance, even if they worked in Flanders (or Brussels), was entirely due to the fact that no such insurance was provided by the other Belgian Communities, nor by the Belgian federal Government.46

B. Critical Reflections The CJEU’s assertion that EU law is, in principle, applicable to nationals of a Member State who work (or reside) in another Member State and to Member State nationals who have exercised their right to free movement is, in itself, not surprising. Such individuals must indeed have the possibility to invoke the prohibition of discrimination on grounds of nationality as laid down in various European legal instruments.47 Moreover, unjustified restrictions on the free movement of employed, self-employed and economically inactive persons cannot be tolerated under the rules of EU law.48 However, it is questionable whether the CJEU’s ruling is satisfactory as far as those principles are applied in the context of

42 Belgian Constitutional Court, No 11/2009, 21 January 2009. For an analysis of this ruling, see T Vandamme, annotation of Government of the French Community and Walloon Government v Flemish Government (2009) 46 CML Rev 287; P Van Elsuwege and S Adam, ‘Belgium. The Limits of Constitutional Dialogue for the Prevention of Reverse Discrimination. Constitutional Court, Judgment 11/2009 of 21 January 2009’ (2009) 5 European Constitutional Law Review 327; J Velaers, ‘Social Federalism and the Distribution of Competences in Belgium’ in Cantillon, Popelier and Mussche (eds), above n 28, 152–57. 43 Belgian Constitutional Court, No 11/2009, 21 January 2009, para B.10.2. 44 Ibid para B.10.1. 45 Ibid para B.10.2. 46 Ibid para B.16. 47 See, among others, Arts 18 and 45 TFEU. 48 On non-discriminatory obstacles to the freedom of movement of workers, see, eg, Case C-415/93 Union royale belge des sociétés de football association ASBL v Bosman, Royal club liégeois SA v Bosman and others and Union des associations européennes de football (UEFA) v Bosman [1995] ECR I-4921; Case C-109/04 Kranemann v Land Nordrhein-Westfalen [2005] ECR I-2421; Case C-208/05 ITC Innovative Technology Center GmbH v Bundesagentur für Arbeit [2007] ECR I-181.

Impact of EU Law on Devolution of Social Powers in Member States

277

the relationships between sub-national entities of a single federal Member State. In what follows, we shall make some critical observations regarding the Court’s judgment. i. Failure to See the Internal Distribution of Powers in the Field of Social Protection as an Aspect of Domestic Social Security Legislation By applying the State-of-employment principle to persons who have exercised their right to free movement within the EU but find themselves in a cross-border situation between sub-national entities of a single Member State, EU law interferes with the internal organisation of social devolution within that Member State. A residence-based distribution of competence between sub-national entities exists in several regionalised Member States. Likewise, the jurisdiction of local authorities is usually limited to residents of their territory, in regionalised as well as nonregionalised Member States. Hence, internal rules concerning the distribution of powers commonly delimit regional and local circles of solidarity on the basis of a residence criterion. In consequence of the CJEU case law, this internal, residencebased delimitation system must be repealed in relation to a limited category of persons. As far as those persons are concerned, the system is substituted by the framework designed for the delineation of circles of solidarity between different EU Member States, which is predominantly workplace-centred. Although the Court—understandably—wanted to ensure the application of the principles of the free movement of persons as laid down by EU law, it did not need to go that far. As explained above, Regulation 883/2004 contains rules determining which Member State’s social security legislation applies in crossborder situations. The legislation which is applicable according to that Regulation encompasses all provisions of the Member States’ internal law in the field of social security. In previous judgments, the CJEU repeatedly held that all national laws directly connected with and sufficiently relevant to the legislation on the branches of social security to which Regulation 1408/71 applies, were envisioned.49 In our view, the constitutional distribution of powers in the fields of social security and social protection between the sub-national entities of a regionalised Member State is also sufficiently and directly linked with the social security legislation of that State. The applicability of the federal and regional social security schemes is—to a considerable degree—dependent on whether these schemes were established in accordance with the internal constitutional system of power division. If the European co-ordination rules determine that a given Member State’s legislation applies, then that Member State’s constitutional rules

49 Case C-327/92 Rheinhold & Mahla NV v Bestuur van de Bedrijfsvereniging voor de Metaalnijverheid [1995] I-1223, para 15; Case C-34/98 Commission v France [2000] ECR I-995, para 35; Case C-169/98 Commission v France [2000] ECR I-1049, para 33. In the Commission v France cases, the Court applied the rules concerning the determination of the applicable legislation to legislation that was classified as tax legislation under internal law.

278

Herwig Verschueren

regarding the distribution of powers in the field of social security and social protection should apply as well. As a result, a migrant worker who is employed in a regionalised Member State should be subject to that Member State’s constitutional provisions regarding the distribution of powers in respect of social security and social protection schemes. Entirely in line with the rationale behind the European system of social security co-ordination, the CJEU could have held that the State of employment’s internal distribution of powers regarding social security legislation is applicable to all migrant employees and self-employed persons who work in the State concerned. Any ensuing disadvantages50 could then have been interpreted as a consequence of the characteristics inherent in the social security system of the Member State in question, more particularly of its regionalised structure. In this regard, the Court could have relied on its earlier case law, according to which migration between Member States may result in better or in worse social protection, depending on the system prevailing in the Member State where the individual is employed.51 ii. Legal Uncertainty over Who Precisely Has Recourse to EU Law The application of the European place-of-employment principle to relationships between sub-national entities of a Member State has created considerable legal uncertainty. In so far as concerned the legal position of migrant EU employees and self-employed persons working in Flanders (or Brussels) and living in Wallonia, the judgment of the CJEU did not rely on the provisions of Regulation 1408/71 (now Regulation 883/2004), but merely on Articles 39 and 43 EC (now Articles 45 and 49 TFEU). This is understandable to a certain extent, as the Regulation is not intended to co-ordinate different social security systems that may co-exist within a single Member State. There is nothing in the Regulation to suggest otherwise, as it merely refers to the Member States as such. However, the non-applicability of the detailed co-ordination rules of Regulations 1408/71 and 883/2004 does give rise to certain specific difficulties, which the more general Articles 45 and 49 TFEU cannot resolve. To begin with, the CJEU failed adequately to specify which migrant employees and self-employed persons should be considered to be impeded in the exercise of their right to free movement by the Flemish care insurance scheme. Did the Court merely intend to refer to those persons who move from another Member State to Belgium, in order to live in Wallonia and work in Flanders (or Brussels)? In its judgment, the CJEU appeared to regard the residence condition as a potential obstacle to

50 In the Flemish care insurance case, non-entitlement to the Flemish scheme due to residence in one of Belgium’s other Communities. 51 See, eg Inasti, above n 9, paras 50–51; Piatkowski, above n 8, para 34; von Chamier-Glisczinski, above n 8, para 85.

Impact of EU Law on Devolution of Social Powers in Member States

279

the free movement of persons, rather than, as the Advocate General had suggested,52 as an instance of discrimination against migrant EU citizens on grounds of their nationality. This may imply that the Court was primarily concerned with those rules that cause ‘workers to lose, as a consequence of the exercise of their right to freedom of movement, social security advantages guaranteed them by the legislation of a Member State’.53 The CJEU seemed to point to a disadvantage suffered as a consequence of the exercise of the right of free movement between Member States. This would entail that persons who had previously moved from a Member State other than Belgium to Wallonia as employees or self-employed persons, and who only subsequently took up work in Flanders (or Brussels) while living in Wallonia, would have no recourse to the Flemish care insurance judgment. After all, the latter category of persons’ non-eligibility for the Flemish care insurance scheme is attributable to their economic migration between two sub-national entities of a single Member State rather than to the exercise of their right to free movement from one Member State to another. In addition, there is a lack of clarity regarding the categories of individuals who have exercised their right to free movement not as employed or self-employed workers but as persons, for example as students. Does it suffice for this kind of EU citizen to have resided in another Member State in the past in order to be able to rely on the CJEU’s judgment on the Flemish care insurance scheme?54 iii. The (Non-)Applicability of the Place-of-Employment Principle to Purely Internal Situations and the Issue of Reverse Discrimination The application of the EU rules concerning the demarcation of circles of solidarity to relationships between a Member State’s sub-national entities (at least in so far as ‘EU migrants’ are concerned) gives rise to ‘reverse discrimination’. Pursuant to the Flemish care insurance scheme judgment, the European State-of-employment principle is held to apply to a certain group of persons who find themselves in a cross-border situation between sub-national entities of a single Member State. Yet persons who find themselves in a very similar situation but have no recourse to EU law are governed by national criteria, which may be different (eg, the place-ofresidence principle). This state of affairs may create instances of so-called ‘reverse discrimination’, whereby EU law grants more rights to persons who can rely on EU law than the rights enjoyed by persons who find themselves in a purely internal situation and are, therefore, merely subject to the relevant domestic legislation.

52 Government of the French Community and Walloon Government, above n 15, Opinion of AG Sharpston, paras 71–83. 53 Government of the French Community and Walloon Government, above n 15, para 46 (emphasis added). 54 For other examples of legal uncertainties created by the CJEU’s judgment, see Verschueren, above n 29, 213–15; Dautricourt and Thomas, above n 29, 445–46; Van Elsuwege and Adam, above n 29, 667.

280

Herwig Verschueren

In the Flemish care insurance case, the government of the French Community and the Advocate General suggested that reverse discrimination is irreconcilable with Union citizenship, a status enjoyed by all Member State nationals. They regarded this form of ‘discrimination’ as a consequence of the very operation of EU law itself and, hence, as the responsibility of that legal order.55 However, the CJEU pointed out that it is settled case law that the Treaty rules governing freedom of movement for persons—and the measures adopted to implement them—cannot be applied to activities which have no factor linking them with any of the situations governed by EU law and which are confined in all relevant respects within a single Member State.56 The Court reiterated that citizenship of the Union is not intended to extend the material scope of the Treaty to internal situations which have no link with EU law.57 Hence, a difference in treatment is maintained between those who are able to invoke EU law in situations concerning the relationship between sub-national entities of a given regionalised Member State, and those who are not. The issue of reverse discrimination has given rise to considerable legal debate in recent years.58 Amongst other things, it has been pointed out that the requirement of a cross-border element to trigger the applicability of EU law is somewhat artificial. It is apparent from the case law of the CJEU in relation to freedom of movement of persons that such a cross-border element is quite easily found. It suffices, for example, for a national of a Member State to have been born in another Member State, or for the individual concerned occasionally to travel to other Member States for business purposes.59 Some authors have suggested that EU citizenship and the rights associated with it belong to all nationals of the Member States, including those who have not migrated within the EU. Hence, sedentary EU citizens should be able to invoke the prohibition of discrimination on grounds of nationality as laid down in Article 18 TFEU in order to claim, from the Member State of which they are nationals and where they reside, the same

55 Government of the French Community and Walloon Government, above n 15, Opinion of AG Sharpston, para 154. 56 Government of the French Community and Walloon Government, above n 15, para 33. 57 Ibid para 39. 58 See, eg, Dautricourt and Thomas, above n 29; E Papadopoulou, ‘Situations purement internes et droit communautaire: un instrument jurisprudentiel à double fonction ou une arme à double tranchant?’ (2002) 38 Cahiers de droit européen 95; C Ritter, ‘Purely Internal Situations, Reverse Discrimination, Guimont, Dzodzi and Article 234’ (2006) 31 EL Rev 690; E Spaventa, ‘Seeing the Wood despite the Trees? On the Scope of Union Citizenship and its Constitutional Effects’ (2008) 45 CML Rev 13; A Trifonidou, ‘Reverse Discrimination in Purely Internal Situations. An Incongruity in a Citizens’ Europe’ (2008) 35 Legal Issues of Economic Integration 43; Van Elsuwege and Adam, above n 29. 59 See, eg Case C-60/00 Carpenter v Secretary of State for the Home Department [2002] ECR I-6279; Case C-200/02 Zhu and Chen v Secretary of State for the Home Department [2002] ECR I-9925.

Impact of EU Law on Devolution of Social Powers in Member States

281

rights as those granted to EU migrants on the basis of EU law.60 However, the CJEU has more recently confirmed its traditional position on this issue.61 iv. Treating Sub-National Entities as Member States in the Application of the Freedom of Movement for Persons Disregards the Member States’ Singular Nature Some authors believe that a solution to the problem of reverse discrimination lies in the application of the same rules, in particular the rules on freedom of movement of persons, to relationships between the Member States and relationships between sub-national entities of a regionalised Member States. According to this view, the rules governing free movement of persons should apply not only to movements between Member States, but also to movements between the sub-national entities of those Member States.62 The authors in question refer to, amongst other things, the case law of the CJEU in relation to the free movement of goods, where it is held that any restriction on the free movement of goods between sub-national entities or regions of a single Member State may be seen as an unjustified restriction on the free movement of goods among the Member States.63 Hence, it is argued, the EU rules on freedom of movement of persons should equally apply between sub-national entities of a Member State, particularly in matters where those sub-national entities possess regulatory powers. Thus, those sub-national entities should, in their mutual relations in matters which are governed by EU common market law, effectively be treated as EU Member States. Hence, they should be bound by the specific EU laws applicable to cross-border situations between such Member States. The fact that a Member State’s sub-national entities may be relevant to the application of EU law is, in itself, not surprising. The CJEU has, for example, also recognised these entities’ relevance to the application of EU law on State aid. If a sub-national entity of a Member State reserves the benefit of a measure for undertakings established in that sub-national entity, that measure may be selective and in contravention of EU law. However, in situations where a Member State’s 60 See among others Spaventa, above n 58, 36–39 and 44. See also the pleas to resolve this reverse discrimination in the literature referred to in n 43. 61 Case C-127/08 Metock and others v Minister for Justice, Equality and Law Reform [2008] ECR I-6241, paras 76–77; Case C-245/09 Omalet v Rijksdienst voor Sociale Zekerheid (CJEU, 22 December 2010), para 12; Case C-434/09 McCarthy v Secretary of State for the Home Department (CJEU, 5 May 2011), para 45. 62 See, eg Dautricourt and Thomas, above n 29, 453–54; Van Elsuwege and Adam, above n 29, 705–08. See also Government of the French Community and Walloon Government, Opinion of AG Sharpston, above n 15, paras 122–32. 63 See especially Case C-163/90 Administration des Douanes et Droits Indirects v Léopold Legros and Others [1992] ECR I-4625; Joined Cases C-363/93, C-407/93, C-408/93, C-409/93, C-410/93 and C-411/93 Lancry SA v Direction Générale des Douanes and Société Dindar Confort, Christian Ah-Son, Paul Chevassus-Marche, Société Conforéunion and Société Dindar Autos v Conseil Régional de la Réunion and Direction Régionale des Douanes de la Réunions [1994] ECR I-3957; Joined Cases C-485/93 and C-486/93 Simitzi v Kos [1995] ECR I-2655.

282

Herwig Verschueren

sub-national entities possess either symmetric autonomous powers or sufficient institutional, procedural and economic/financial autonomy in a particular matter, the CJEU regards the sub-national entity, not the territory of the entire Member State to which the sub-national entity belongs, as the relevant context wherein the selectivity of the measure should be assessed.64 The fact that undertakings established in another sub-national entity of the Member State cannot benefit from the measure does not make that measure territorially selective and potentially in breach of EU law on State aid, since the measure was taken by a sub-national entity possessing autonomous powers.65 Some authors have referred to this case law in support of their assertion that a Member State’s sub-national entities possessing sufficient autonomy to regulate in a particular area must be treated as Member States in the context of the application of this regulation. This would also imply that relations between the various sub-national entities of the Member State in question are governed by the same rules applicable among Member States.66 We think, however, that the aforementioned case law of the CJEU recognises, first and foremost, the internal constitutional rules of Member States with regard to the autonomy of their subnational entities. In the context of the application of EU law, such a sub-national entity may thus take measures the scope of which is limited to its own territory. These measures may therefore differ from measures taken by other sub-national entities of the same Member State, without these measures having to be regarded as regionally selective and therefore unlawful under EU law on State aid.67 In our opinion, this case law does not warrant the conclusion that a sub-national entity, in its relations with other sub-national entities of the same Member State, must always apply EU internal market law. In its judgment of 17 November 2009, for example, the CJEU ruled that a regional Sardinian tax on stopovers made by aircraft or boats operated by persons having their tax domicile outside the territory of the region constituted an obstacle to the freedom of movement of services. The Court held, more precisely, that persons ‘having their tax domicile outside the territory of the region and established in other Member States’ suffered

64 Case C-88/03 Portugal v Commission [2006] ECR I-7115, paras 52–68; Joined Cases C-428/06 to C-434/06 Unión General de Trabajadores de La Rioja (UGT-Rioja) and Others v Juntas Generales del Territorio Histórico de Vizcaya and Others [2008] ECR I-6747, paras 53–60. See more recently also Case C-169/08 Presidente del Consiglio dei Ministri v Regione Sardegna [2009] ECR I-10821, para 60. 65 See, on this discussion in the context of regional tax measures, K Lenaerts and N Cambien, ‘Regions and the European Courts: Giving Shape to the Regional Dimension of Member States’ (2010) 35 EL Rev 609, 629–33; B Peeters, ‘European Legal Limitations to the Repartition of Fiscal Competences in a Federal State Structure’ in Cantillon, Popelier and Mussche (eds), above n 28, 227–53. 66 Dautricourt and Thomas, above n 29, 453; Van Elsuwege and Adam, above n 29, 707–08. 67 Vandamme, above n 42, 298–99. See also JA Winter, annotation of Portugal v Commission (2008) 45 CML Rev 183, 195. The CJEU applied the same reasoning to divergences between measures adopted by the different sub-national entities of a Member State in order to comply with a regulation in an area in which these entities have autonomous competences. Such divergences do not constitute discrimination in breach of EU law. See Case C-428/07 Horvath v Secretary of State for Environment, Food and Rural Affairs [2009] ECR I-6355.

Impact of EU Law on Devolution of Social Powers in Member States

283

an additional cost.68 Thus, in order for EU internal market law to apply, the CJEU continues to require a cross-border element involving another Member State. A cross-border situation involving another sub-national entity of the same Member State clearly does not satisfy this requirement. This conclusion also holds in relation to EU law on the free movement of persons. Moreover, if a Member State’s sub-national entity were also subject to EU free movement law in internal cross-border situations, the rules governing these internal situations could not be different from the EU rules governing the relations between the Member States. This implies, for instance, that the notion of ‘economic and monetary union’ could not be interpreted differently in the domestic and EU legal orders. Thus, it has been argued that the notion of ‘economic and monetary union’ laid down in Belgian constitutional law69 should be interpreted in the same way as its counterpart in EU internal market law, particularly in so far as the applicability of the place-of-employment criterion for access to social protection is concerned.70 Yet it appears that ‘economic and monetary union’ does not necessarily have the same meaning in Belgian constitutional law as the corresponding notion in EU law.71 In its judgment of 21 January 2009, the Constitutional Court confirmed this viewpoint.72 Indeed, to ‘purely internal’ Belgian situations, involving a person living in Wallonia and working in Flanders (or Brussels), the Constitutional Court did not apply the EU place-of-employment principle but the internal Belgian constitutional rule regarding the delimitation of the personal and territorial scope of a social protection scheme introduced by the Flemish Community, that is, the place-of-residence principle. It appears that the Belgian Constitutional Court gave precedence to the territorial distribution of powers laid down in the Belgian Constitution over the principles of EU internal market law. Applying the EU place-of-employment principle to a purely internal cross-border situation would have been incompatible with the Belgian constitutional territorial distribution of powers, which is based on residence. Besides, the mandatory application of EU internal market law to relationships between a Member State’s sub-national entities would render meaningless

68

Presidente del Consiglio dei Ministri v Regione Sardegna, above n 64, para 32 (emphasis added). See Art 6(1) VI(3) of the Belgian Special Law on Institutional Reform. Pursuant to this provision, the regions must, when exercising their economic powers, take due account of the freedom of movement of persons, goods, services and capital, of the freedom of commerce and industry, and of the economic and monetary union. 70 H Storme, ‘Vlaamse zorgverzekering. Een lange weg naar een geschikte aanknopingsfactor’ (2008) Nieuw Juridisch Weekblad 614, 622–23. 71 E Van Brutsem, ‘L’affaire de la “zorgverzekering”: comment concilier la libre circulation des travailleurs et l’organisation fédérale des états?’ (2009) Chroniques de droit social 193, 198–99; Vandamme, above n 42, 294–95; Van Elsuwege and Adam, above n 42, 336; J Velaers and J Vanpraet, ‘De materiële en territoriale bevoegdheidsverdeling inzake sociale zekerheid and sociale bijstand (II)’ (2009) 64 Tijdschrift voor Bestuurswetenschappen en Publiekrecht 195, 210–11. On this issue, see also Stef Feyen’s contribution in ch 17 of this volume. 72 Belgian Constitutional Court, No 11/2009, 21 January 2009, para B.16. 69

284

Herwig Verschueren

the federal structure of that Member State. In a regionalised Member State, the demarcation of the material, personal and territorial scope of application of the legislation of that State’s sub-national entities proceeds on the basis of national political agreements and constitutionally-established criteria. These criteria may vary from those incorporated in EU law to govern the relationships between the Member States. By denying that national and European criteria may diverge, one threatens to make any regionalisation superfluous and insubstantial. The treatment of these sub-national entities as if they were Member States would, for example, preclude any internal legal delimitation of the personal and territorial scope of application of social protection schemes established by sub-national entities. A simple reference, mutatis mutandis, to EU rules governing the relationships between Member States would then suffice. Furthermore, the treatment of such sub-national entities as if they were Member States might also threaten the federal cohesion within regionalised Member States. Indeed, it would then be only a small step towards such sub-national entities acquiring the status of fully-fledged autonomous Member States. Lastly, the assumption that sub-national entities of a single Member State should, in their mutual relationships, be regarded as fully-fledged Member States appears to be contrary to the EU Treaty. The new Article 4(2) TEU proclaims that the Union shall respect the equality of Member States before the Treaty as well as their national identities, inherent in their fundamental structures, political and constitutional, inclusive of regional and local self-government. …73

IV. CONCLUSION

This chapter started with the following question: To what extent does EU law impact on the relationship between the sub-national entities of a Member State if these sub-national entities have regulatory powers in the field of social protection? More specifically, we set out to explore whether the criteria relied upon in EU law for determining the scope of the circles of solidarity in relationships among the Member States should also be applied in the context of relationships between the sub-national entities of regionalised Member States. The judgment of the CJEU in the Flemish care insurance case answers this question in the affirmative, but only for the limited group of people who, in view of their migratory history, are able to invoke EU law. For the large majority of persons who find themselves in a cross-border situation between a Member

73 See also Vandamme, above n 42, 298. For a recent reference to this provision and the relevance of the national identity of Member States for the implementation of EU law, see Case C-208/09 Sayn-Wittgenstein v Landeshauptmann von Wien (CJEU, 22 December 2010), para 92. The CJEU referred to the status of the Austrian State as a Republic.

Impact of EU Law on Devolution of Social Powers in Member States

285

State’s sub-national entities, by contrast, only internal legislation applies. This does mean, however, that any territorial and personal delimitation of the circles of solidarity by a sub-national entity of a Member State is in part governed by rules of EU law and is, therefore, no longer a matter to be determined by domestic constitutional law alone. This evolution has given rise to so-called ‘reverse discrimination’ between persons who, in a cross-border situation between sub-national entities, are able to invoke EU law and those who are not. The CJEU has held that this issue is beyond the ambit of EU law. The Belgian Constitutional Court, for its part, has confirmed the applicability of the Belgian rules governing internal distribution of powers to the group of sedentary persons, as otherwise the domestic, exclusively territorial, system of power division would be threatened. This case law has created considerable legal uncertainty, and the ensuing reverse discrimination has been met with incomprehension. It remains an open question whether the European rules for delimiting circles of solidarity between Member States in situations involving migrant EU citizens, and which constitute a form of European social federalism, should be applied in their entirety to relationships between the sub-national entities of a single Member State. Although this would resolve both the legal uncertainty and the reverse discrimination problem, we remain unconvinced. It would be legally more consistent to resolve this unsatisfactory situation by applying, in the context of relations between sub-national entities of a regionalised Member State, the domestic constitutional rules governing the distribution of powers to all categories of persons, including EU migrants.74 We feel this solution follows from the rationale behind the EU Regulations concerning the co-ordination of social security within the Union. Economically active EU migrants ought to be subject to the legislation of the country where they are employed, including to any constitutional rules regarding the delimitation of circles of solidarity within that Member State. Migrant EU citizens who find themselves in a cross-border situation between two sub-national entities must, in other words, be treated in the same way as the group of non-migrants who find themselves in the same interregional situation. The alternative proposal that the European rules for determining circles of solidarity should equally apply to sedentary persons who find themselves in a cross-border situation between a regionalised Member State’s sub-national entities, by contrast, threatens to prejudice the often delicate and precarious political agreements underlying the structure of such States. According to the latter ‘solution’, the sub-national entities of a regionalised Member State should be treated as if they themselves were Member States, which ultimately compromises the federal nature of the State to which those sub-national entities belong.

74 See also AP van der Mei, ‘Editorial. Combating Reverse Discrimination: Who Should Do the Job?’ (2009) 16 Maastricht Journal of European and Comparative Law 379, 381–82.

286

Herwig Verschueren

This is clearly not to say that the European model of social federalism, which has developed detailed and intricate criteria for determining circles of solidarity in cross-border situations, cannot serve as a source of inspiration for the delimitation of circles of solidarity in cross-border situations between sub-national entities of a single Member State. We do feel, however, that there is no legal reason to impose the solutions that have been worked out at the EU level to these internal issues. What is more, such an imposition may be regarded as an improper interference with the constitutional relationships established within regionalised Member States.

13 EU Law and Language Regulation in (Quasi-)Federal Member States ELKE CLOOTS* AND STEFAN SOTTIAUX**

I. INTRODUCTION

I

N THE TWENTIETH century, several Western European countries embarked on a path to devolution or even federalisation. By reforming their governance structure, traditionally unitary States such as Finland, Italy, Spain, Belgium, and the United Kingdom sought to accommodate the aspirations for selfgovernment of sub-State national societies living in their territory.1 Significantly, these processes of devolution almost invariably coincided2 with the granting of (co-)official status to the language of sub-State nations, at least within those nations’ own region.3 In the (quasi-)federal States under consideration,4 regional institutions such as the legislature, the administration, courts and

* PhD Fellow of the Research Foundation—Flanders (FWO), Institute for Constitutional Law, Katholieke Universiteit Leuven. ** Professor of constitutional law and discrimination law, Katholieke Universiteit Leuven. We should like to thank the participants in the workshop for helpful suggestions and discussions. 1 See, eg W Kymlicka, Politics in the Vernacular (New York, Oxford University Press, 2001) 94–95; J McGarry, B O’Leary and R Simeon, ‘Integration or Accommodation? The Enduring Debate in Conflict Regulation’ in S Choudhry (ed), Constitutional Design for Divided Societies (New York, Oxford University Press, 2008) 63–69. 2 To be sure, there may be a time gap between the granting of official language status and the conferral of self-government rights. The federalisation of Belgium, for instance, started only in the 1970s, while the use of Dutch in certain public institutions and official documents had already been imposed at the end of the 19th century. See, eg, the so-called Equality Act 1898, repealed by the Act on Language Use in Legislation in 1961. 3 See, eg, the Act on the Autonomy of Åland 1991, ss 1, 18 and 36; Belgian Constitution, Arts 1–4 and 127–130; Spanish Constitution, Arts 2–3, 143 and 148; Scotland Act 1998; Gaelic Language (Scotland) Act 2005, s 1; Welsh Language Act 1993; Government of Wales Act 2006; Welsh Language (Wales) Measure 2011, s 1; Italian Constitution, Arts 116–117; Special Statute of the Trentino–Alto Adige Region and the Provinces of Trento and Bolzano, Art 99. 4 For the purposes of this chapter, the term ‘(quasi-)federal States’ refers to multinational Member States which are either nominally federal (eg, Belgium) or have gone through a process of devolution without adopting the label ‘federal State’ (eg, the UK and Spain).

288 Elke Cloots and Stefan Sottiaux public schools now operate in the local majority language, either exclusively or on a par with the State’s majority language. In addition, a requirement to use (at least) the local language is often imposed on a wide range of private actors (eg, the media) or communications (eg, in employer–employee relations, or on product labels). It is hardly surprising that regional self-government and the protection of the regional majority language co-occur. Since they are commonly distinguished from the dominant national society by language, sub-State nations tend to strive not only for political autonomy, but also—and often primarily— for the predominance of their mother tongue in the region where their members live and work.5 The aim of this chapter is to explore how the language policies featuring in the abovementioned (quasi-)federal States—which are all Member States of the EU— relate to the requirements of EU law. The Treaty articles on freedom of movement and non-discrimination are especially relevant in this regard. Although the shortage of case law of the European Court of Justice (ECJ) on the matter may suggest otherwise,6 the said Treaty provisions are frequently in tension with the language policies under consideration.7 This chapter seeks to identify the major areas of conflict between language policies in (quasi-)federal Member States and EU law, and to develop a theoretical framework for addressing these conflicts. The framework put forward distinguishes between three types of regulation: (a) public language regulation in contexts where State neutrality is impossible; (b) public language regulation in contexts where State neutrality is possible; and (c) private language regulation. A separate section is devoted to each of these categories. The normative framework proposed in this chapter will, it is hoped, provide the ECJ with a helpful tool for monitoring these sorts of policies. The first test cases are already underway. The ECJ has recently received references for preliminary rulings on two sensitive regional language laws: a South Tyrolean measure stipulating that only people who declare themselves to belong to one of the three regional linguistic groups

5

For a broader account of these constitutional reforms, see Kymlicka, above n 1, 212–13. Only a limited number of judgments have so far dealt with language regulation in (quasi-)federal Member States. See especially Case 137/84 Mutsch [1985] ECR 2681; Case C-369/89 Piageme and others v BVBA Peeters [1991] ECR I-2971; Case C-85/94 Groupement des Producteurs, Importateurs et Agents Généraux d’Eaux Minérales Etrangères, VZW (Piageme) and others v Peeters NV [1995] ECR I-2955; Case C-274/96 Bickel and Franz [1998] ECR I-7637; Case C-33/97 Colim [1999] ECR I-3175; Case C-281/98 Angonese [2000] ECR I-4139; Case C-250/06 United Pan-Europe Communications Belgium SA and Others [2007] ECR I-11135; Case C-222/07 Unión de Televisiones Comerciales Asociadas (UTECA) [2009] ECR I-140. 7 See generally, B De Witte, ‘Common Market Freedoms versus Linguistic Requirements in the EU States’ in A Milian i Massana, B De Witte, J Woehrling, I Urrutia Libarona and MA Orriols i Sallés (eds), Mundialització, lliure circulació i immigració, i l’exigència d’una llengua com a requisite (Barcelona, Institut d’Estudis Autonómics, 2008) 109–31. 6

EU Law and Language Regulation 289 are eligible for a certain allowance, and a Flemish law imposing the use of Dutch in employment relations.8

II. PUBLIC LANGUAGE REGULATION WHERE NEUTRALITY IS IMPOSSIBLE

It is widely acknowledged among political theorists that every polity inevitably has to determine the language or languages in which its institutions shall operate. As a consequence, certain languages—and their native speakers—unavoidably receive more favourable public treatment than others. The strategy of ‘benign neglect’, often pursued by States for coping with religious pluralism, is simply not available to governments confronted with multilingualism.9 Political theorists are divided on whether the promotion of certain languages resulting from their use in public institutions is to be celebrated or deplored.10 However, whether it is to be regretted or not, the plain fact is that complete State neutrality to these linguistic issues is unattainable. The question then becomes how a polity, in particular a multinational and multilingual one, is to determine the languages which shall have official status in its territory. This section starts with a brief overview of the various ways in which the question has been answered in political theory. Next, we look at how this dilemma has been resolved by the (quasi-)federal Member States under consideration. The third part of this section explores whether, and to what extent, EU law should have an impact on the Member States’ choices in this regard. Although our focus is mainly on Member States where minority languages have acquired official status, we also briefly investigate whether our findings can be extended to Member States that have so far largely refused to accommodate their national minorities.

A. Political Theory and Official Language Policy in Multinational Member States The quest for the ‘best’ official language policy for multinational States has been occupying political philosophers for a while now. The suggestions offered so far vary widely, depending on, amongst other things, the values to which one primarily adheres. It is widely acknowledged, for example, that deliberative democracy 8 Case C-571/10 Kamberaj [2011] OJ C46/7 (reference from the Tribunale di Bolzano); Case C-202/11 Anton Las v PSA Antwerp NV [2011] OJ C219/3 (reference from the Antwerp Labour Court). See also Case C-571/10 Kamberaj, Opinion of AG Bot. 9 See eg, B Barry, Culture and Equality (Cambridge, Polity Press, 2001) 107; JH Carens, Culture, Citizenship, and Community (New York, Oxford University Press, 2000) 53–54 and 77–78; W Kymlicka, Multicultural Citizenship (Oxford, Oxford University Press, 1994) ch 6; Y Tamir, Liberal Nationalism (Princeton, NJ, Princeton University Press, 1993) 140–41 and 145–50. But see A Patten, ‘Liberal Neutrality and Language Policy’ (2003) 31 Philosophy and Public Affairs 356. 10 Yael Tamir, for example, explicitly encourages public institutions to reflect the culture and language of their national society (Tamir, above n 9, 74). But see Kymlicka, above n 1, 250–52.

290 Elke Cloots and Stefan Sottiaux and economic solidarity are easier to realise in a linguistically homogeneous polity. When the citizens of a State are tied by linguistic bonds, they experience fewer difficulties in communicating and identifying with one another. Effective communication and a common identity, in turn, facilitate democratic debate and the pursuit of distributive justice.11 Given these beneficial consequences of linguistic homogeneity, it might be argued that a multilingual and multinational State would do well to encourage the spread of a common language in its entire territory. From that perspective, it makes perfect sense for such a State to designate the national majority language as the only official State language. Other principles, however, prescribe an alternative linguistic regime for multinational States. So, several authors embrace conceptions of justice entailing that national minority languages too must be endowed with official status. Their arguments are typically structured as follows. First, one or more essential interests people have in connection with language policy are identified. In a next step, it is argued that justice requires that people be equally respected in relation to these essential linguistic interests. Consequently, if government cannot avoid supporting the linguistic interests of one national group, justice requires it to support the interest of other national groups as well. For Will Kymlicka, for instance, people have an essential interest in membership in their own ‘societal culture’, which he understands as a territorially concentrated and institutionally complete culture, centred on a shared language.12 Access to one’s own societal culture is valuable, in Kymlicka’s view, because it provides individuals with a meaningful context of choice, and a secure sense of self-identity and belonging.13 Kymlicka then posits that fairness requires that the good of cultural membership be protected equally for the members of all national groups.14 This means, amongst other things, that ‘if [the government] supports the majority culture, by using the majority’s language in schools and public agencies, it cannot refuse official recognition to minority languages’.15 A similar but different justice-based argument endorsing official status for national minority languages may be found in Philippe Van Parijs’s work. Van Parijs contends that people have a fundamental interest in respect for their identity, including their linguistic identity, and that justice requires people to be granted

11 Carens, above n 9, 84; Kymlicka, above n 1, 25–27, 212–16 and 224–27; A Patten, ‘What Kind of Bilingualism?’ in W Kymlicka and A Patten (eds), Language Rights and Political Theory (New York, Oxford University Press, 2003) 310–13; P Van Parijs, ‘Linguistic Diversity as Curse and as By-Product’ in X Arzoz (ed), Respecting Linguistic Diversity in the European Union (Amsterdam, John Benjamins Publishing Company, 2008) 33–36. 12 Kymlicka, above n 9, 76; Kymlicka, above n 1, 76. For a critique of Kymlicka’s basic premise, see J Waldron, ‘Minority Cultures and the Cosmopolitan Alternative’ (1992) 25 University of Michigan Journal of Law Reform 751. 13 Kymlicka, above n 9, 75–105. 14 Ibid 113. 15 Ibid 111.

EU Law and Language Regulation 291 equal dignity.16 In Van Parijs’s view, the only way to secure equal dignity of the various languages and the associated identities is to make each language the official language of its native speakers’ political community.17 For instance, in the multilingual and federal State of Belgium, Dutch should be the official language of the Flemish political community, and French of the Walloon political community.18 Although many theorists would now agree that the languages of sub-State nations deserve official status, the controversy is far from settled. One major source of dispute is whether and to what extent the dominant language of a multinational State should retain official status in the territory of the sub-State nation. Political philosophers are sharply divided as to whether a ‘one region/one language’ regime should be adopted or, rather, a regime of bilingual institutionalism in which persons have the choice between two languages. Once again, arguments based on effective communication, deliberative democracy, economic solidarity and respect for identity are being invoked to support either a system of monolingual or bilingual public institutions.19 Some philosophers strongly favour a language regime in which the sub-State nation’s language is the only official language in its territory (the so-called linguistic territoriality principle/regime). At least in Will Kymlicka and Philippe Van Parijs’s view, this is the only way the essential linguistic interest of the members of the national minority group can be paid the same respect as the national majority’s interest.20 At the other end of the spectrum, authors like Helder De Schutter advocate a system of bilingual institutionalism and personal language choice for regions where native speakers of both the national minority language and the national majority language live. According to De Schutter, such a language regime succeeds where the linguistic territoriality principle fails: it ensures that the identities of the members of the two language groups—and of people who simultaneously feel attached to both language groups—are equally recognised

16 P Van Parijs, ‘Linguistic Justice and the Territorial Imperative’ (2010) 13 Critical Review of International Social and Political Philosophy 181, 182. 17 Ibid 185. 18 P Van Parijs, ‘Must Europe Be Belgian? On Democratic Citizenship in Multilingual Polities’ in C McKinnon and I Hampsher-Monk (eds), The Demands of Citizenship (London, Continuum, 2000) 247. For analogous fairness-based theories, see Carens, above n 9, 77–87; Patten, above n 9, 378–84. Leslie Green, by contrast, seems to give a more central role to legitimate expectations and the fact of agreement; see L Green, ‘Are Language Rights Fundamental?’ (1987) 25 Osgoode Hall Law Journal 639, 663–65. 19 See, eg, Patten, above n 11. Significantly, certain arguments have been invoked on both sides of the debate. So, the importance of identity considerations, and of people’s linguistic identities being treated with equal dignity, is stressed by proponents as well as critics of the linguistic territoriality principle. Cp, eg, Van Parijs, above n 16, and H De Schutter, ‘The Linguistic Territoriality Principle—A Critique’ (2008) 25 Journal of Applied Philosophy 105. 20 Kymlicka, above n 1, 79; Van Parijs, above n 16, 183. As already said, in Kymlicka’s view, this is the interest in access to one’s own societal culture, while what matters for Van Parijs is equal dignity of linguistic identities.

292 Elke Cloots and Stefan Sottiaux and respected.21 Between these two poles we find, not surprisingly, theorists who consider both types of language regime morally permissible.22

B. Official Language Policy in (Quasi-)Federal Multinational Member States It is clear that all (quasi-)federal Member States under consideration have designated the idioms of their main sub-State national groups as ‘official’. Official status usually implies that the sub-State nation’s language is the language (or one of the languages) employed by regional public institutions such as the regional parliament, public administration, courts, public schools, etc.23 Another expression of this official recognition is the use of the sub-State nation’s language on public signs and in the public media.24 The question as to whether the languages of national minorities deserve official status in their own region may currently be considered to be more or less settled in these Member States. Nevertheless, also in the (quasi-)federal States under consideration, language remains a highly divisive issue. Not only in the theoretical debate outlined above, but also in practice, disputes often revolve around the status of the dominant State language in the territory of the sub-State nation. As the law stands now, only the use of the regional majority language is generally permitted in public institutions in Flanders and the Åland Islands (ie, Dutch and Swedish, respectively).25 In other self-governing entities, such as Scotland, Wales, the Basque Country, Catalonia, and Trentino–Alto Adige/Südtirol,26 the Member State’s majority language may be employed as well (ie, English, Spanish and Italian, respectively). However, in

21

De Schutter, above n 19, 117. See also Helder De Schutter’s contribution in ch 7 of this volume. See, eg Carens, above n 9, 86; Patten, above n 11, 298. Contrary to Joseph Carens, Alan Patten makes an attempt at determining the conditions under which one or the other regime is preferable. 23 As regards Finland, see Act on the Autonomy of Åland 1991 c VI. As regards Belgium, see especially the Belgian Constitution Arts 4, 30 and 129; Act concerning Language Use in Judicial Proceedings 1935; Act concerning the Linguistic Regime in Education 1963, Arts 4–8; Act concerning Language Use in Administrative Affairs 1966. As regards Spain, see especially the Spanish Constitution, Art 3; Statute of Autonomy of the Basque Country, Arts 6 and 35; Statute of Autonomy of Catalonia, c III; Catalan Linguistic Policy Act, cc I and III. As regards Italy, see especially Special Statute of the Trentino–Alto Adige Region and the Provinces of Trento and Bolzano, Arts 19 and 99–101. As regards the UK, see especially Welsh Language Act 1993, pt III, and Welsh Language (Wales) Measure 2011, pts I and II. 24 See, eg Catalan Linguistic Policy Act, Arts 18, 25 and 27(1); Special Statute of the Trentino–Alto Adige Region and the Provinces of Trento and Bolzano, Art 8; Welsh Language Act 1993, ss 25–27; Flemish Radio and Television Act 2009, Arts 6, 18 and 154–156. 25 Nevertheless, the use of the ‘dominant’ national language is permitted in certain exceptional circumstances. See, eg Belgian Act concerning Language Use in Administrative Affairs 1966, Arts 7–8, 11–15, 24–26 and 28–30; Act on the Autonomy of Åland 1991, ss 37 and 39. 26 Admittedly, the co-official status of the German language in this autonomous Region of the Italian Republic works mainly to the benefit of the German-speaking citizens of the Province of Bolzano, which is one of the two provinces constituting the autonomous Region. See especially Special Statute of the Trentino–Alto Adige Region and the Provinces of Trento and Bolzano, Arts 1, 90–91 and 100–101. 22

EU Law and Language Regulation 293 some of these States, the ongoing struggle of the dominant linguistic group for regaining or retaining language rights in the minority nation’s territory might one day change that picture.27 The question that interests us here is whether EU law, in particular the fundamental freedoms and the prohibition of discrimination on grounds of nationality, should play a role in these domestic processes and influence the design of a multinational Member State’s official language policy. It is to this issue that we now turn.

C. EU Law and Official Language Policy in (Quasi-)Federal Multinational Member States Put bluntly, our answer to the above question is a qualified ‘No’. This section submits that a multinational, (quasi-)federal Member State’s official language policy should, in principle, not be a matter of EU law. More particularly, it is our view that a multinational Member State’s decision that the public institutions of its autonomous regions must operate in the regional majority language, either exclusively or on a par with the State’s dominant language, should fall outside the scope of the Treaty provisions on freedom of movement and non-discrimination. A distinction must be made, however, between the designation of official languages and problems in the margin. As will become clear below, we see no compelling reason why EU law should equally keep its hands off the latter. i. Designation of Official Languages in (Quasi-)Federal Member States In our view, at least two arguments militate in favour of exempting the designation of official languages from the scope of the Treaty provisions on freedom of movement and non-discrimination on grounds of nationality. The first argument relates to institutional choice. The second point we wish to make is a more substantive one. Both arguments can be traced back to the theoretical discussion set out above. It will be clear by now that political philosophers answer the question of how to design a just language policy for a multinational State in a variety of ways. Moreover, within and across the Union’s multinational Member States, opinions strongly differ in this regard. Given the profound disagreement—both in theory and in practice—over the optimal way to determine a multinational State’s official languages, it is doubtful whether the ECJ should take sides in the debate. The sensitive nature of these issues can only reinforce our doubts. The prominence of identity-based arguments on both sides of the theoretical and domestic disputes would allow the ECJ very little room for error. However, the risk of the ECJ 27 Consider, for instance, the discussions about the rights of French-speaking Belgians in the Flemish territory (see, eg Belgian Constitutional Court, No 73/2003, 26 May 2003; No 124/2010, 28 October 2010) and the status of the Castilian language in Catalonia (see, eg Spanish Constitutional Tribunal, No 31/2010, 28 June 2010).

294 Elke Cloots and Stefan Sottiaux making errors here is quite high. In fact, we have ample reasons to distrust the ECJ with the task of screening a Member State’s official language regime. To begin with, the context-specific constitutional accords on which such regimes are based may not fully be understood by a supranational court. In addition, the ECJ’s alleged inclination for individual economic interests and the presence of only one judge per Member State may render the Court biased against the interests of a certain language group, or at least create the perception of such a bias. This high likelihood of far-reaching errors provides a strong argument for not allowing the Court to engage in an open-ended balancing test when monitoring the official language policies of the (quasi-)federal Member States. The Court would do better to restrict its own discretion in this area. One way to do so would be to adopt a doctrinal rule excluding the said policies from the scope of the fundamental freedoms and the prohibition on discrimination on grounds of nationality.28 But there is still another, more substantive argument, which relates to a second conclusion the preceding normative discussion reveals: the considerations which a multinational Member State should take into account when determining its official languages relate to fairness, respect for people’s identity, democratic participation, economic solidarity, etc. The typical EU concerns for freedom of movement and the eradication of nationality discrimination, by contrast, are not among the yardsticks against which these choices should primarily be judged. To be sure, as a matter of fact, some arrangements fit the standards of freedom of movement and non-discrimination on grounds of nationality better than do others. When an officially unilingual Member State decides to extend official status to another, more vulnerable language which immigrants would otherwise have little incentive to learn, it in fact makes moving more difficult. This is especially so if the language in question is designated as the only official language in the sub-State nation’s territory (eg, Dutch in Flanders), or if there is no other EU Member State where the language in question is officially recognised (eg, Basque or Catalan). Thus, on a scale of ‘openness to foreign EU citizens’, a unilingual francophone Belgian State would rank higher than a regime making Dutch the co-official language in Flanders, which would, in turn, rate better than the Dutch monolingualism currently reigning in Flanders.29 A similar thought experiment for Spain would yield an analogous result: from a unilingual Castilian Spanish State, over officially bilingual autonomous communities, down to Catalan, Basque and Galician monolingualism in the regions associated with these national minority languages. However, it should not depend on these types of considerations whether a sub-State national society’s language deserves official status and, if so, whether monolingual or bilingual regional institutions should be preferred. This is, of course, not to say that no moral standards exist against which the policies of the 28 For a more detailed development of this argument, see Elke Cloots’s contribution in ch 14 of this volume. 29 Except from the point of view of the nationals of The Netherlands, of course.

EU Law and Language Regulation 295 Member States in question should be evaluated. The only point we wish to make at this juncture is that the EU norms on freedom of movement and nationality discrimination are not amongst them. It does not therefore seem appropriate to require Member States (or their sub-State political entities) to justify their choices in this regard under the fundamental freedoms and the prohibition of nationality discrimination. Hence, we recommend that the ECJ exclude from the scope of those Treaty provisions questions as to whether a sub-State nation’s institutions should use solely the sub-State nation’s language, solely the State’s dominant language, both languages or even something other than these two languages. As a corollary, the same recommendation goes for the ensuing language proficiency requirements imposed on judges of regional courts, civil servants working for regional administrations or teachers of regional public schools. Linguistic knowledge conditions are an indispensible part of a Member State’s choice to recognise certain languages as official languages. The decision that judicial proceedings in a sub-State nation’s territory must be conducted in that nation’s vernacular, for instance, would be meaningless if the Member State could not require the judges in question to know that vernacular.30 To see what our proposal could mean in practice, it is instructive to take a look at the Groener case.31 In that case, the ECJ was asked to assess an Irish measure requiring that public school teachers be proficient in English as well as the Irish language, both of which are designated as official languages by Irish constitutional law. Since the Court—for the first time—accepted the protection and promotion of a national official language as a legitimate aim of a Member State policy, and engaged in a lenient proportionality review, it should not surprise that the judgment has often been reported as deferential to domestic language policy.32 However, if a similar case were to arise in the context of one of the multinational (quasi-)federal Member States, a different method of adjudication would, in our view, be preferable.33 Imagine, for example, a Portuguese teacher whose application for a job in a Catalan public school is dismissed due to his lack of knowledge

30 It should be noted that this argument does not hold for those staff members of a public body who are not involved in that body’s official communication with ‘the public’, such as concierges, secretaries and cleaning personnel. Requirements regarding these persons’ linguistic proficiency bear no relation to the language(s) in which judicial proceedings are conducted, education is provided, or access to the public administration is available. Obviously, those persons’ language use or knowledge is not what the advocates of official status for national minority languages are concerned about. It does not therefore seem inappropriate to subject linguistic conditions attached to such jobs to the fundamental freedoms and the prohibition of nationality discrimination. 31 Case 379/87 Groener [1989] ECR 3967. 32 See, eg, R Creech, Law and Language in the European Union (Groningen, Europa Law Publishing, 2005) 103; J Gerards, ‘Pluralism, Deference and the Margin of Appreciation Doctrine’ (2011) 17 European Law Journal 80, 96; BME McMahon, ‘Case 379/87, Groener v Minister for Education and the City of Dublin Vocational Education Committee (CDVEC), Judgment of 28 November 1989, not yet reported’ (1990) 27 CML Rev 129, 135 and 139. 33 Whether the alternative mode of adjudication proposed here would also have been preferable in Groener itself is a question beyond this chapter’s scope.

296 Elke Cloots and Stefan Sottiaux of Catalan.34 The teacher is fluent in Castilian though. If the ECJ were to be asked whether the linguistic requirement at issue is compatible with the fundamental freedoms and the ban on nationality discrimination, we think the better approach would be to exclude the measure from the scope of EU law rather than to regard it as an interference with the individual’s freedom of movement to be justified by the Spanish and/or Catalan governments. The fact that the Court, following Groener, would be inclined to apply the proportionality test with light intensity does not alter our conclusion. As a matter of principle, it appears undesirable to regard freedom of movement as a standard which the designation of official languages in a multinational Member State, including the corollary knowledge requirements, should meet in the first place. Moreover, a judicial balancing act would take account of the extent to which the language requirement hinders freedom of movement. Since official protection of ‘weaker’ languages interferes more severely with people’s spontaneous language choice than does the official protection of ‘stronger’ languages, it would be harder to justify the official status of weak languages than that of their stronger counterparts. Such a mode of adjudication sits uneasily with a commitment to equal dignity of the various national groups living in the EU. The significance and reach of the categorical mode of adjudication proposed here should not be overstated, however. We merely suggest that the ECJ adopt a doctrinal rule stating that the determination of the languages to be used by a subState nation’s administration, judiciary, public schools, etc, and—as a corollary— the languages to be known by these institutions’ personnel, is not subject to the Treaty provisions regarding freedom of movement and nationality discrimination. Yet so far nothing has been said about the adjudication of issues that arise at the periphery of the granting of official status to languages in (quasi-)federal Member States but have, in essence, little to do with it. It is to these issues that we now turn. ii. Peripheral Issues A first example of a peripheral issue concerns the proof and the level of language proficiency required for a job in the regional public sector. As far as this type of requirement is concerned, the contention that State neutrality is unattainable does not hold up. There is no inevitability to, for instance, a requirement that the certificate attesting to the job applicant’s proficiency in the official language be obtained in the region or Member State in question, or that that language be the applicant’s mother tongue.35 A State could be perfectly neutral with respect to such issues. In situations where neutrality to a person’s linguistic identity is an option, it seems convenient that the Member State justifies its decision to do 34

See Catalan Linguistic Policy Act, Art 24. See, eg Special Statute of the Trentino–Alto Adige Region and the Provinces of Trento and Bolzano, Art 19(1) (requiring that the mother tongue of teachers in Bolzano is German or Italian). 35

EU Law and Language Regulation 297 otherwise in the face of the EU fundamental freedoms and discrimination ban. In addition, there is little reason to distrust the ECJ with the evaluation of this sort of conditions, which are—in contrast to the aforementioned designation of a State’s official languages—liable to hide an agenda of economic protectionism rather than fairness. The ECJ case law, submitting the contested measures to a proportionality test, is in line with these intuitions.36 Another problem looming in the background of the granting of official status to minority languages is that, in some instances, the facilities attached to that status are open only to the members of the national minority group, not to immigrants in the group’s territory, let alone to occasional visitors. Only persons who reside in that region (and sometimes only those residents who are also nationals of the Member State) are entitled to public services in the regional language. Non-residents or non-nationals, by contrast, are excluded from the scope of these domestic language rights. Nationality and residence criteria for entitlement to a certain benefit granted by domestic law are always viewed with suspicion by the ECJ.37 Again, it cannot be excluded that such nationality and residence conditions, like the aforementioned local certificate requirement, mask a strategy of economic protectionism rather than being grounded in considerations of justice. It seems proper, therefore, for the Court to review such conditions under the fundamental freedoms and the prohibition of nationality discrimination. Is there any possible justification for this type of nationality or residence condition? In Bickel and Franz, the ECJ briefly mentioned the fact that no ‘additional complications or costs’ would be entailed by an obligation to extend the regional language rights to non-residents.38 At issue was whether the right of residents of the Italian Province of Bolzano to require that criminal proceedings be conducted in German must be extended to an Austrian lorry driver and a German tourist travelling in the Province. The Court decided that it must, given the prohibition of discrimination on grounds of nationality enshrined in Article 6 TEC (now Article 18 TFEU).39

36 Groener, above n 31, para 23; Angonese, above n 6, paras 44–45. Admittedly, none of these judgments involved the public regulation of language use in a sub-State nation’s territory. While Groener dealt with a national minority language which had official status in the entire Member State, Angonese concerned a language requirement imposed by a private person. However, it may safely be assumed that the ECJ would apply the same reasoning as regards a sub-State nation’s public institutions. 37 At least if the situation at issue is not considered ‘purely internal’. See, eg Case 212/06 Government of the French Community and Walloon Government v Flemish Government [2008] ECR I-1683, para 38. The fact that the Member State imposed the condition in the exercise of its own powers cannot alleviate suspicion on the part of the Court. See, eg, Bickel and Franz, above n 6, para 17 (the regulation of the language of criminal proceedings); Case C-391/09 Malgožata Runevicˇ-Vardyn and Łukasz Paweł Wardyn (ECJ, 12 May 2011), para 63 (the regulation of the way a person’s name is entered on certificates of civil status). 38 Bickel and Franz, above n 6, para 30. 39 Ibid para 31.

298 Elke Cloots and Stefan Sottiaux Reasoning a contrario, we might expect that a residence condition40 could be justified if the accordance of equal language rights to non-residents would entail excessive financial burdens for the region’s public institutions. However, the likelihood of this risk occurring is rather low. Once a regional public institution, a court for instance, has been organised so as to be able to operate in the national minority language as well, it will usually be capable of communicating at little or no extra cost with foreigners who are proficient in that language. To a certain extent, the good of public services being delivered in the minority language is non-rival: consumption of the good by one individual does not reduce the amount available for others.41 Stated differently, once the good is provided, the cost of providing it to an additional individual is zero or negligible. Especially if the ECJ were only willing to take into account the extra costs entailed by the service delivery to the individual foreign claimant in the case before it, financial considerations would be unlikely to be able to justify the residence requirement. Nonetheless, it is possible to conceive of cases in which it should be legitimate for State authorities to make the provision of public services in the regional language conditional on residence in the region. To understand this, it is instructive to recall why the (quasi-)federal Member States here considered have decided that public institutions (courts, administrative bodies, public schools, etc) should operate in the national minority language(s). As already noted, the very aim of giving official status to a sub-State nation’s language is to accommodate the linguistic interests of the members of that national group, interests they hold not as individuals but as group members. It is, therefore, quite normal for a State to accord the associated language rights to the members of that group (ie, the minority nation) and not to other people. Moreover, several political theorists have argued that only a national group that consists of a sufficient number of territorially concentrated people sharing the same linguistic identity has a justice-based claim to public services in its own language.42 Immigrants, by contrast, would not have the same moral entitlements to access to public institutions in their mother tongue.43 Against this background, it should not be surprising to see a condition of nationality of the Member State and of residence in the region of the minority group attached to certain domestic language rights.44

40 Whether this ground of justification could also be relied on to justify a nationality requirement is less certain, as Art 45(3) TFEU lists the objectives which may be invoked to that end (ie, public policy, public security and public health). 41 See, eg Green, above n 18, 659–60. See also n 118 below. 42 See, eg Carens, above n 9, 82–87; Kymlicka, above n 9, 92–93 and 95–106; Van Parijs, above n 16, 192–93. 43 There is considerable disagreement, though, as to why the moral language rights of immigrants are not as extensive as those of so-called ‘national minorities’. See, eg, Carens, above n 9, 80–87; Kymlicka, above n 9, 95–100; Van Parijs, above n 16, 192–93. 44 See, eg, Special Statute of the Trentino–Alto Adige Region and the Provinces of Trento and Bolzano, Art 100(1); Circular Letter No BA-97/22 concerning the Language Use in Flemish Municipal Authorities 1997 (the so-called ‘Omzendbrief-Peeters’), permitting the inhabitants of certain Flemish municipalities to request that the French Language be used in their dealings with the local authorities.

EU Law and Language Regulation 299 To be sure, even if we agree that immigrants, including EU citizens, do not have a free-standing moral right to receive public services in their mother tongue in their host State, this should not, in principle, prevent EU law from requiring that existing domestic language facilities, which were originally intended to protect national minorities, be extended to foreigners who happen to be fluent in the national minority language. This seems indeed to be the approach taken by the ECJ in Mutsch and Bickel and Franz.45 While not imposing any duty on Member States to grant official status to the languages of immigrants or foreign visitors, the Court does in principle require that those languages that already have official status in a national minority region be open to be used by nationals and residents of other Member States as well. The language rights migrant workers, service providers, tourists, etc can claim, under EU law, vis-à-vis their host State are thus merely parasitic on the rights granted to the national minority group. Nonetheless, we think a residence requirement should be tolerated if the opening up of the linguistic facility in question to non-residents would be likely to threaten the fundamental linguistic interests of the other language group living in the autonomous region. Will Kymlicka would probably consider this condition to be fulfilled as soon as the other national group’s access to its own (viable) societal culture was no longer guaranteed. Philippe Van Parijs would perhaps require that the Member State’s languages and associated identities were no longer treated with ‘equal dignity’. Still others, like Alan Patten, would look at the impact of the language facility’s extension on the conditions of social mobility and deliberative democracy in the regional political community.46 Clearly, whether this ground of justification may successfully be invoked or not will depend on empirical factors such as the relative ‘vulnerability’ and ‘attractiveness’ of the languages involved, the number of immigrants who might be expected to make use of the language facility, the likelihood that the language facility will discourage immigrants (even more) to learn the (other) official language of the regional political community, and so on. An example may clarify the ground of justification we have in mind. The factually dominant language in Belgium is French.47 In order to protect the Flemish national society, a constitutional accord has been reached to apply the territoriality principle to the Flemish and the French-speaking political communities. As a consequence, the Flemish public institutions operate exclusively in Dutch. However, certain exceptions have been allowed to this rule to the benefit of the French-speaking inhabitants of a limited number of municipalities neighbouring Brussels but located in the Flemish territory. These inhabitants are entitled to request the municipal administration 45

Mutsch, above n 6, paras 12 and 18; Bickel and Franz, above n 6, para 31. The latter interest is particularly important when the region in question possesses considerable political autonomy. See Patten, above n 11, 310–13 and 319–20. 47 Although the number of Dutch-speaking Belgians currently outweighs the number of Frenchspeaking Belgians, the spontaneous mechanisms described by Philippe Van Parijs as ‘probabilitysensitive learning’ and ‘maxi-min communication’ work in favour of the French language in Belgium. See Van Parijs, above n 16, 184–85. 46

300 Elke Cloots and Stefan Sottiaux to employ the French language in their communications with them.48 It might be argued that a duty on the part of these municipalities to use French in their dealings with all French-speaking EU citizens, wherever they live, would seriously threaten the interests of the Flemish national group in connection to their language. In any event, such an expansion of the language rights of francophones in Flanders would entail a much more far-reaching exception to the territoriality principle than the Flemish initially agreed to. We suggest, therefore, that the ECJ accommodate this type of residence requirement if extending the language right to non-residents would pose a threat to the fundamental linguistic interests of the other language group living in the autonomous region in question.49 The fundamental linguistic interests envisioned here might include the interests of the language group members in equal dignity of their identity, access to a viable societal culture or ‘context of choice’ centred around their own language, democratic participation, etc. Moreover, we recommend that the ECJ leave the proportionality review to the domestic courts, limiting itself to providing guidance on the parameters to be taken into account. Not only do national courts have better access to the information needed to give estimates of parameters such as language vulnerability and immigration rates, but they also have a better understanding of the constitutional compromises between the national language groups, of these compromises’ background and legitimacy, and of national stability concerns. The ECJ should be aware of the consequences were it to dismiss the ground of justification proposed here. As the case law now stands, multinational Member States have no obligation under EU law to grant official status to the languages of their sub-State national societies. However, if a Member State decides to do so, the Treaty prohibition of nationality discrimination compels the State to extend the linguistic rights flowing from that official status to all foreign EU citizens within the sub-State nation’s jurisdiction, even though these foreigners do not belong to the national group whose interests the Member State initially sought to protect. If the ECJ were to apply this rule without any regard for the particular situation of the linguistic and national groups involved in the case at hand, a wide range of Member States might be put off from conferring any language rights at all upon their sub-State national societies.

48 Belgian Act concerning Language Use in Administrative Affairs 1966, Arts 25–26; the so-called Omzendbrief-Peeters, above n 44. For an account of this language policy in English, see Patten, above n 11, 319–20. See also Council of Europe Parliamentary Assembly, Resolution 1301 (2002) on the protection of minorities in Belgium, available at . 49 This ground of justification may be considered the linguistic policy-orientated counterpart of the grounds the ECJ has accepted in the context of social and fiscal policy. See, eg Case C-120/95 Decker [1998] ECR I-1831, para 39, and Case C-158/96 Kohll [1998] ECR I-1931, para 41 (‘the risk of seriously undermining the financial balance of the social security system’); Case C-204/90 Bachmann [1992] ECR I-249, para 21 (‘the need to preserve the cohesion of the tax system’).

EU Law and Language Regulation 301 This perverse effect may occur especially in post-communist Member States, where various national minorities have a bigger ‘kin-State’ in the EU. One might think of, for example, the ethnic Hungarians in Slovakia and the ethnic Poles in Lithuania. If granting language rights to these national minorities entails that all Hungarian-speaking and all Polish-speaking EU citizens coming within the jurisdiction of Slovakia and Lithuania, respectively, are entitled to the same facilities, these Member States may well prefer not to give language rights to their national minorities. This risk is considered especially high if there is a strong sense of loyalty between the national minority and its kin-State,50 and if the national majority group was dominated by the minority group’s kin-State in the past.51 In sum, to ensure that foreign EU citizens equally join in the rights and benefits accorded to national minorities is an admirable aim; but without attention to the conditions for minority protection to be granted in the first place, the pursuit of this objective will not bring justice much closer.

D. EU Law and Official Language Policy in Unitary Multinational Member States So far, we have discussed the relation between EU law and the language policy of multinational Member States that have devolved powers to their sub-State nations and conferred official status on these nations’ languages. It has been argued, more precisely, that EU law should not prevent Member States from according official status to national minority languages and from requiring the related regional public institutions to operate (exclusively or non-exclusively) in the language of the minority nation. However, the converse question is interesting as well: might a national minority successfully invoke EU law to force its Member State to grant its language official status? The ECJ has recently been faced with this query in a case involving the language rights of the Polish-speaking minority in Lithuania.52 At issue was whether a Lithuanian national belonging to that minority group and her Polish husband were entitled to have their names written in Polish characters in Lithuanian official documents (ie, their birth and marriage certificates). Lithuanian constitutional law denied them that right. The Constitution designates Lithuanian as the official

50 See, eg, the recent accusation by the Lithuanian Ambassador to Poland that ethnic Poles living in Lithuania are disloyal to the Lithuanian State (‘Sikorski “regrets” Lithuanian ambassador’s statement’, Warsaw Business Journal, 21 June 2011, ). 51 Recall the (at least perceived) subordination of Slovaks and Lithuanians in the Polish-Lithuanian Commonwealth (1569–1795), the Austro-Hungarian Empire (1867–1918), the Second Polish Republic (1918–39) and the Soviet Union (1922–91). See more generally on this topic, W Kymlicka, ‘The Internationalization of Minority Rights’ in S Choudhry (ed), Constitutional Design for Divided Societies (New York, Oxford University Press, 2008) 134–35. 52 Malgožata Runevic ˇ-Vardyn and Łukasz Paweł Wardyn, above n 37.

302 Elke Cloots and Stefan Sottiaux State language,53 which implies, amongst other things, that official documents must be written in Lithuanian characters.54 Yet the applicants resorted to their EU rights to equal treatment and freedom of movement, hoping that these rights would entail an obligation for the Lithuanian authorities to spell their names the Polish way.55 The ECJ did not bite, however. The Court held, to begin with, that national rules requiring that the spelling of the official national language be used in a Member State’s official documents do not come within the scope of the EU Race Equality Directive.56 Next, the Court resorted to its settled case law regarding a person’s name to evaluate the claims based on the right to move and reside freely in the EU. In a series of earlier judgments, the Court had pointed out that an obligation to use different names in different Member States, due to national differences between the rules governing names, can cause ‘serious inconvenience’ for the person concerned.57 If, and only if, such a serious inconvenience can be shown to exist will the discrepancy between names amount to a restriction on the freedoms conferred by Article 21 TFEU.58 In the event that such a restriction is found, it must be verified whether the restriction pursues a legitimate objective and is proportionate to its aim.59 Applying its ‘names doctrine’ to the case at hand, the Court found, as regards the spelling of the same names, two discrepancies between the Polish spouse’s birth certificate and the couple’s marriage certificate: the letter ‘w’ had been changed into a ‘v’, and diacritical marks had been omitted. The Court continued that, whereas the latter discrepancy did not cause any serious inconvenience for the person concerned,60 the former potentially did. It was left to the national court, however, to decide whether a serious inconvenience and, hence, a free movement restriction existed and, if so, whether it was justified.61 To this end, the ECJ clarified that the protection of a Member State’s official national language constitutes a legitimate objective.62

53

Lithuanian Constitution, Art 14. See especially Lithuanian Constitutional Court, No 14/98, 21 October 1999, para V.4, accessible in English on the Constitutional Court’s website . 55 That means, basically, using the letter ‘w’ instead of ‘v’ and adding diacritical signs. Neither the letter ‘w’ nor diacritical marks appear in the Lithuanian alphabet. 56 Malgožata Runevic ˇ-Vardyn and Łukasz Paweł Wardyn, above n 37, para 48. See also Council Directive (EC) 2000/43 implementing the principle of equal treatment between persons irrespective of racial or ethnic origin [2000] OJ L180/22. 57 Case C-148/02 Garcia Avello [2003] ECR I-11613, para 36; Case C-353/06 Grunkin and Paul [2008] ECR I-7639, paras 23–28; Case C-208/09 Sayn-Wittgenstein (ECJ, 22 December 2010), paras 55–56 and 66–70. 58 Malgožata Runevic ˇ-Vardyn and Łukasz Paweł Wardyn, above n 37, paras 70, 74–78 and 81–82. 59 Garcia Avello, above n 57, para 39; Grunkin and Paul, above n 57, para 29; Sayn-Wittgenstein, above n 57, para 81; Malgožata Runevicˇ-Vardyn and Łukasz Paweł Wardyn, above n 37, para 83. 60 Malgožata Runevic ˇ-Vardyn and Łukasz Paweł Wardyn, above n 37, paras 81–82. 61 Ibid paras 74–78 and 91. 62 Ibid paras 85–87. 54

EU Law and Language Regulation 303 Importantly, no discrepancy was found to exist between the Lithuanian spouse’s birth certificate and the couple’s marriage certificate. As the woman’s names had invariably been spelled the Lithuanian way, there was no restriction of her free movement rights. The Court’s decision implies that members of the Polish national minority living in Lithuania have no claim under EU law to have their names written in Polish characters in Lithuanian official documents. Thus, whereas Polish citizens—and other foreign nationals—potentially derive certain rights to linguistic accommodation in Lithuania from the judgment, Lithuania’s Polish national minority is without a remedy. Yet the language rights of that national minority are precisely what the controversy between Lithuania and Poland, and between the Lithuanian and Polish language groups, is about.63 To be clear, we do not wish to suggest here that the ECJ should have granted the Lithuanian woman’s request to have her names spelled in Polish characters on the basis of EU law. Quite the contrary is true. Our concern is, rather, that the existing EU norms on freedom of movement and non-discrimination are ill-suited to evaluating claims for official language status made by linguistic minority members against their own Member State. Whether, and to what extent, a sub-State nation’s language ought to be granted official status is simply not, and should not be, a matter of promoting mobility between EU Member States, ensuring equal treatment of other Member States’ nationals and removing ‘discrepancies’ between official documents. Besides, it is not always crystal-clear what domestic language regime is the most beneficial from a ‘mobility’ point of view. Within the confines of one multinational Member State, mobility will likely benefit from there being a single official State language and alphabet. Cross-border mobility, on the other hand, may or may not gain from a system where minority languages have co-official status, depending on whether the minority language in question also enjoys official status in one or more other Member States.64 In a similar vein, we consider the EU norms on free movement and non-discrimination inappropriate for the evaluation of foreign EU citizens’ claims for the use of their own language in official documents issued by another Member State. To be sure, it is beyond doubt that Lithuania in fact hampers the mobility of EU citizens who are used to another alphabet and negatively affects nationals of other Member States by banning the use of foreign characters in its official documents. The point we wish to make, however, is that the languages to be used by a multinational Member State’s public institutions, including in official documents, should be designated on the basis of concerns lying at the heart of the linguistic justice debate, 63 See, eg, ‘Poland and Lithuania. Narcissistic Differences’, The Economist, 21 October 2010, ; ‘Lithuania and Poland. Bad Blood’, The Economist, 20 January 2011, . 64 The official status of German in South Tyrol, for instance, may be assumed to facilitate rather than complicate cross-border mobility in the EU. See, eg, European Commission, The Euromosaic Study. German in Italy, ; Research Centre of Multilingualism, Euromosaic—German in Italy, .

304 Elke Cloots and Stefan Sottiaux such as respect for people’s identity, democracy and redistributive justice concerns. European Union norms on free movement and nationality discrimination, by contrast, should not interfere with this process.65 We believe, therefore, that it would have been more appropriate if the ECJ had exempted Lithuania’s official language policy from the scope of these Treaty provisions, instead of applying these provisions to situations for which they were not designed.66 In our view, such a general exemption for official language policies would have been preferable over the standard-like proportionality review to which the ECJ actually resorted. There are still a number of other reasons why the ECJ’s mode of adjudication in the case of Runevicˇ-Vardyn and Wardyn might come to be be regretted. First, the specific test applied by the Court had so far been employed to review national measures that oblige EU citizens to use materially different names in different Member States. More particularly, the Court’s earlier case law concerned divergences between the Member States’ legal orders as to the composition of surnames67 and the permissibility of titles of nobility.68 However, Runevicˇ-Vardyn and Wardyn involved differences between national alphabets. Putting divergent models of spelling the same name on an equal footing with differences in content between a person’s names may produce odd results. Let us explain this further. In essence, the logic underlying the ECJ’s ‘names doctrine’ is as follows: the greater the discrepancy between the names given to a single person in two different Member States, the more serious the inconvenience for the person concerned, and the weightier the justification required. The same idea also underpins the Court’s judgment in Runevicˇ-Vardyn and Wardyn: whereas the mere omission of diacritical signs is not liable to give rise to any inconvenience, changing the letter ‘w’ into a ‘v’ is so liable, and accordingly needs to be justified. However, while a linear relationship between the degree of the discrepancy between two names, on the one hand, and the strength of the justification required for that discrepancy, on the other hand, makes perfect sense where the content (or pronunciation) of a person’s name is at issue, it is counter-intuitive in the context of different ways of spelling the same name. In the latter type of case, one would expect exactly the opposite proposition to be true: the more significant the differences between two alphabets X and Y, the more self-evident it is that names written in alphabet X need

65 The fundamental right to respect for private and family life does matter. However, since in our view the issue does not come within the scope of application of EU law, it is not for the ECJ to decide on the Lithuanian policy’s compatibility with that fundamental right. See, eg Case C-144/95 Maurin [1996] ECR I-2909, paras 12–13; Case C-299/95 Kremzow [1997] ECR I-2629, para 15. 66 Significantly, the ECJ did employ the decision-making method we put forward here in relation to the applicants’ claims based on the EU Race Equality Directive. The Court held, more precisely, that ‘national rules which provide that a person’s surnames and forenames may be entered on the certificates of civil status of that State only in a form which complies with the rules governing the spelling of the official national language relate to a situation which does not come within the scope of Directive 2000/43’ (Malgožata Runevicˇ-Vardyn and Łukasz Paweł Wardyn, above n 37, para 48). 67 Garcia Avello, above n 57; Grunkin and Paul, above n 57. 68 Sayn-Wittgenstein, above n 57.

EU Law and Language Regulation 305 to be modified for Y-speakers to be capable of reading and correctly pronouncing those names.69 It may surprise that the ECJ overlooked this essential distinction between a name’s content/pronunciation and the alphabet in which a name is written. In effect, the Court had noticed this dichotomy earlier in Konstantinidis, a case concerning the transcription of a Greek name into Roman characters in German official documents. The Court held, more precisely, that ‘[t]here is nothing in the Treaty to preclude the transcription of a Greek name in Roman characters in the registers of civil status of a Member State which uses the Roman alphabet’.70 The Court added that a restriction of the Greek person’s freedom of movement could nonetheless arise from a modification of the pronunciation of his name due to transcription errors.71 The use of a different alphabet as such, however, was exempted from the scope of the Treaty provisions on freedom of movement. Secondly, by stating that the transcription of the letter ‘w’ into a ‘v’ in Lithuanian official documents potentially amounts to a violation of the freedom of movement of Polish nationals, the ECJ may put considerable pressure on the Lithuanian authorities to accord the same facilities to its Polish-speaking national minority. Let us suppose, for a moment, that the referring national court were to find a violation of EU law and that the Lithuanian authorities subsequently allowed Polish and other foreign EU citizens72 to have their names spelled the original way in Lithuanian official certificates. If that language facility were not extended to Lithuanian nationals belonging to the Polish-speaking national minority, a situation of ‘reverse discrimination’ would arise. Even if we leave the difficulties associated with any instance of reverse discrimination aside, this state of affairs seems grossly unjust. Indeed, if there is one category of people that possesses a moral claim against the Lithuanian Government to have its alphabet officially recognised, it is the Lithuanian Polish-speaking minority. Admittedly, this unfair outcome might easily be averted if the Lithuanian State were willing to include the members of the Polish-speaking national minority within the scope of the language facility. In that event, the converse of the Bickel and Franz logic would occur: a national minority would acquire language rights through the extension of the rights conferred by EU law upon nationals of other Member States.73 One may doubt, however, whether this is the most convenient 69 This intuition is supported by Art 2 of ICCS Convention No 14 on the recording of surnames and forenames in civil status registers. The provision requires that surnames and forenames be reproduced without alteration or translation in so far as the name is written in characters that also exist in the language in which the official document is to be made. Diacritical signs, however, must always be reproduced, even if such signs do not exist in the language in which the official document is drafted. 70 Case C-168/91 Konstantinidis [1993] ECR I-1191, para 14. 71 Ibid paras 16–17. 72 Indeed, in that event, we see no reason why the letter ‘w’ in names of, say, English, German or Dutch origin should not be allowed in Lithuanian official documents. 73 In Bickel and Franz (above n 6), residents of other Member States acquired language rights through the extension—imposed by EU law—of the rights conferred by national law upon residents of a certain province of the Member State. See section II.C.ii. above.

306 Elke Cloots and Stefan Sottiaux way of securing justice for national minorities. It may safely be assumed that a negotiated compromise between the language groups living in the multinational Member State in question would be seen as a more legitimate route to justice and stability.74 It might therefore be thought undesirable to employ the Treaty provisions on freedom of movement and nationality discrimination as a means of imposing the use of a national minority language on a multinational Member State’s public institutions through the backdoor. Moreover, proceeding otherwise would place national minorities which lack a kin-State in the EU (or another Member State where their language has official status) at a comparative disadvantage. Indeed, there are no foreign EU citizens who can resort to their EU free movement and non-discrimination rights with a view to indirectly forcing Estonia, for example, to accommodate its ethnic Russian minority. The same holds true for the Catalan, Basque and Galician minorities in Spain, and the Welsh minority in the UK. For these minorities and their languages, the EU guarantees to free movement and equal treatment offer no salvation; quite the contrary.75 Thirdly, a general exemption of the designation of a multinational Member State’s official language policy from the scope of EU law may be important for symbolic reasons. If the referring Lithuanian court were to decide that the fundamental freedoms of Ms Runevicˇ-Vardyn and Mr Wardyn have not been infringed, the Lithuanian Government could truthfully say that its official language policy is sanctioned by EU law. Consequently, the Lithuanian State may be left with little incentive to give official recognition to the Polish alphabet in its official documents. This is a real risk, given that post-communist Member States tend to regard international and European minority protection standards not as minimum requirements but, rather, as a ceiling.76 This risk might have been minimised by excluding a multinational Member State’s selection of official languages from the scope of the Treaty provisions on freedom of movement and non-discrimination. If that were the path the ECJ had chosen, the Lithuanian State could have said, at best, that EU law is not applicable to its official language policy. Now, however, Lithuania may legitimately claim that the way it treats foreign nationals as well as its own Polish-speaking citizens complies with the requirements of EU law.

III. PUBLIC LANGUAGE REGULATION WHERE NEUTRALITY IS POSSIBLE

In the previous section, we stressed the inevitability of the State determining the language in which administrative bodies, courts, public schools, etc operate and in which official documents and signs are written. As a result, it was said, the State

74 W Kymlicka, ‘National Minorities in Postcommunist Europe: The Role of International Norms and European Integration’ in Z Barany and R Moser (eds), Ethnic Politics After Communism (Ithaca, NY, Cornell University Press, 2005) 207. 75 See section II.C.i. above. 76 Kymlicka, above n 74, 217.

EU Law and Language Regulation 307 cannot avoid promoting certain languages. The same reasoning, however, cannot be applied to other types of public language regulation frequently recurring in (quasi-)federal Member States, such as rules restricting language choice in private communication. In effect, a State may specifically decide not to regulate the language in which private practitioners and suppliers advertise, sell their products and services, communicate with their staff, etc. Neither does the said rationale extend to language requirements attached to social benefits that are not delivered through communication but, rather, through a transfer of money or another tangible object (eg, a dwelling). Indeed, authorities are not evidently compelled to reserve advantages such as allowances or social housing for people who are proficient in the local language. Not only is it feasible for a State not to intervene in these matters, it may also be required. Respect for an individual’s fundamental rights, most notably the rights to freedom of expression, to respect for private and family life, and to social and housing assistance, may well imply that private persons must be free to operate and interact in the language of their choice, and may not be excluded from social benefits on grounds of language proficiency.77 Nevertheless, it is not uncommon for the sub-State autonomous polities of the (quasi-)federal Member States to impose this kind of linguistic requirement. A typical example is a regional law requiring private employers to draw up labour contracts and/or employee information in the regional majority language.78 Another widespread obligation concerns the language in which information for consumers or service users is to be provided. Such requirements usually imply that the regional majority language, and sometimes also the national majority language, be employed.79 However, regional regulation of private language use is not limited to the economic sphere. It also affects cultural actors. Language quotas for radio and television programmes, including for those transmitted by private broadcasters, are not uncommon.80 Neither are duties for private radio and television stations to broadcast in the regional language.81 Lastly, there are examples of language requirements in the social sector. In Flanders, for instance, the social housing market is reserved for people who prove their willingness to learn Dutch,82 while in South Tyrol, EU citizens must declare that they belong 77 In this vein, see eg, Human Rights Committee, Singer v Canada, Communication no 455/1991, para 12.2; Canadian Supreme Court, Ford v Quebec (Attorney General) [1988] 2 SCR 712, paras 39–44; Belgian Constitutional Court, No 101/2008, 10 July 2008, B.33.1–B.35. See also Committee on the Elimination of Racial Discrimination, Concluding Observations CERD/C/BEL/CO/15 (Belgium, 2008), para 16. 78 See, eg, Catalan Linguistic Policy Act, Art 36; Flemish Act on language use in employer–employee relations 1973. A reference for a preliminary ruling involving the Flemish law has recently been sent to the ECJ (see n 8 above). 79 See, eg, Catalan Linguistic Policy Act, Art 34; Basque Decreto 123/2008 on the Language Rights of Consumers and Users. 80 See, eg, Catalan Linguistic Policy Act, Art 26; Flemish Radio and Television Act 2009, Arts 138, 154–155, 157 and 186. 81 See, eg, Catalan Linguistic Policy Act, Art 26; Flemish Radio and Television Act 2009, Arts 129, 163, 169 and 174. 82 Flemish Housing Act 1997 (Decreet houdende de Vlaamse Wooncode), Art 95.

308 Elke Cloots and Stefan Sottiaux to—or elect to join—one of the three linguistic groups of South Tyrol in order to be eligible for a provincial housing benefit.83 Why are regional authorities so eager to impose these sorts of language requirements if they are not forced to do so by practical necessity? The protection of employers, consumers and social housing tenants obviously figures among the objectives for limiting the linguistic options of private economic actors or requiring proof of language proficiency.84 Yet the preambles to and travaux préparatoires of some of these language laws also point to interests of a more expressive nature. These laws seek, for example, to promote and encourage the use of the regional language so that it becomes the ‘language of normal use’ in the regional socio-economic sphere.85 The question arises, however, whether a polity is permitted actively to promote the use of a certain language by taking these sorts of measures. As already noted, the situation under consideration here differs significantly from the situation discussed in the previous section. State neutrality is an option towards both language use in communications between private persons and language proficiency of applicants for welfare benefits. The State can avoid privileging certain languages over others in these contexts, just as it can (and should) avoid privileging a given religion or race. If a State or regional polity nonetheless decides not to be neutral but, rather, to intervene with the linguistic choices people would have made spontaneously, this public interference with personal autonomy is in need of justification. This means that the State must prove that the language requirement in question pursues a legitimate aim and is proportionate to that aim.

A. Legitimate Aim What, then, are the objectives that are capable of justifying a language requirement in a context where State neutrality is possible? In its case law concerning the fundamental freedoms, the ECJ has accepted a variety of goals, which may be divided into three categories: facilitating communication, promoting an official language, and maintaining linguistic diversity.

83 Bolzano Provincial Law No 13 on Social Housing 1998 (Ordinamento dell’edilizia abitativa agevolata), Arts 1 and 5. A reference for a preliminary ruling involving this provincial law has recently been sent to the ECJ (see n 8 above). 84 See, eg, the preamble to the Basque Decreto 123/2008 on the Language Rights of Consumers and Users; Flemish Parliament, Travaux Préparatoires 2005–2006, No 824/1, p 5 (). See also Belgian Constitutional Court, No 101/2008, 10 July 2008, paras B.9.1, B.10.2 and B.32.1. 85 See, eg, the preamble to the Basque Decreto 123/2008 on the Language Rights of Consumers and Users; the preamble to and Arts 1, 5, 32(2), 35(2) and 36 of the Catalan Linguistic Policy Act. Spanish Autonomous Communities tend to call this ‘la normalización lingüística’. Although the promotion of the use of the Dutch language was not expressly mentioned as an objective of the language requirement embedded in the Flemish Housing Act, it may be assumed to have been one of the Flemish parliament’s goals.

EU Law and Language Regulation 309 i. Communication The most ‘successful’ objectives are those assigned to the first category. Language is obviously an instrument for communication. Governments often seek to ensure that communication is effective in certain human relationships and mandate, therefore, the use of a specific language. They may wish to facilitate communication between two private actors for a variety of purposes: to protect vulnerable parties such as consumers, workers, patients, attorney clients and social housing tenants,86 to ensure abidance by rules of professional conduct,87 and/or to produce goods like public health, the proper administration of justice or social cohesion.88 ii. Language Promotion The legitimacy of the aim of fostering communication, and of the further aims associated with it, is relatively uncontested. However, as mentioned above, the regional language policies considered here tend to be directed simultaneously towards other objectives. Even though this is not always expressly admitted in the law’s preamble or travaux préparatoires,89 requirements that the regional language be employed in private communications, or known in order to be eligible for a regional social advantage, are often aimed at the promotion of the use of the language in question. The ECJ has already acknowledged that it is legitimate for a Member State to try to defend and promote its official language(s), including in the private sphere.90 In other words, the ECJ allows a polity to promote its official language by requiring the language’s use in certain private communications for the sake of promoting that language. But is this truly a legitimate objective? Is the promotion of a particular language a good worth pursuing in itself? Many (though not all)91 political philosophers would probably deny that.92 They would, rather, point to the important interests people have in relation to their language, such as the interest in their language’s

86 See, eg Colim, above n 6, paras 43–44; Case C-424/97 Haim [2000] ECR I-5123, paras 59–60; Case C-506/04 Wilson [2006] ECR I-8613, para 74; Case C-193/05 Commission v Luxembourg [2006] ECR I-8673, para 44. See also Belgian Constitutional Court, no 101/2008, 10 July 2008, paras B.9.1 and B.32.1. 87 Haim, above n 86, paras 59–60; Wilson, above n 86, para 74; Commission v Luxembourg, above n 86, para 44. 88 See, eg Colim, above n 6, para 33; Belgian Constitutional Court, No 101/2008, 10 July 2008, paras B.9.1 and B.32.1. 89 See n 85 above. 90 UTECA, above n 6, paras 27, 33 and 36. At issue was a legal obligation for public and private television operators to earmark a part of their revenue for the pre-funding of films made in one of the Member State’s languages. 91 See, eg Tamir, above n 9, 74 and 152. 92 See, eg De Schutter, above n 19, 114–16; W Kymlicka, Liberalism, Community and Culture (Oxford, Clarendon Press, 1989) 165; Van Parijs, above n 16, 185–86 and 189; Waldron, above n 12, 781–88.

310 Elke Cloots and Stefan Sottiaux survival over time, in the reassurance that their language group will flourish, in membership in their own societal culture providing them with a context of choice, in their linguistic identities being treated with equal dignity, etc.93 It leaves no doubt that, in pursuing such interests, governments will likely promote the language in question. Nevertheless, the ECJ should be wary of blindly accepting the legitimacy of language promotion as such, without further examination of the public body’s underlying motives. Furthermore, not all of these motives are equally capable of justifying limitations on a private person’s freedom of movement. The imposition of burdens on individuals out of a mere concern for the survival of the language in the long run, for instance, has been regarded with suspicion by several political theorists.94 But even those linguistic interests that may be considered sufficiently fundamental, such as the interest in ‘equal dignity’, will not always be able to outweigh the countervailing interest in personal autonomy. In our view, these fundamental interests could justify constraints on language use between private persons, or the attachment of language proficiency conditions to welfare benefits, only if these interests would be injured in a context of State neutrality. This, however, is a condition to be verified when reviewing the proportionality of the language requirement. Yet before addressing the proportionality issue, we should take a brief look at the third type of rationale that may be accepted by the ECJ to ground a linguistic requirement in the private sphere: respect for linguistic diversity. iii. Linguistic Diversity Although the ECJ has not yet unequivocally accepted respect for linguistic diversity as a legitimate aim to be pursued by domestic law, it may be anticipated that it will do so in the future. To begin with, Article 3(3) TEU and Article 22 of the EU Charter of Fundamental Rights commit the Union to respect its ‘linguistic diversity’. The ECJ referred to these provisions in Runevicˇ-Vardyn and Wardyn,95 a case already discussed above. Even though the case concerned the spelling of names in official documents, it cannot be excluded that the Court will in the future also appeal to ‘linguistic diversity’ when deciding on measures regulating communication between private actors. After all, the ECJ has already confirmed the lawfulness of a legal obligation for cable operators in a bilingual region to broadcast programmes transmitted by private broadcasters belonging to both language groups. It was held in this regard that the law, by seeking to maintain ‘the pluralist nature of the range of television programmes available in the bilingual region’, pursued a legitimate purpose.96

93

See section II.A. above. See, eg De Schutter, above n 19, 114–15; Green, above n 18, 653–58; Van Parijs, above n 16, 185–86. 95 Malgožata Runevic ˇ-Vardyn and Łukasz Paweł Wardyn, above n 37, para 86. 96 United Pan-Europe Communications Belgium SA and Others, above n 6, paras 40–42. 94

EU Law and Language Regulation 311 But is the preservation of linguistic diversity truly an objective worth striving for? Whereas certain authors answer this question in the affirmative,97 others have vigorously contested the value that has been ascribed to linguistic diversity. At best, they say, it is a side-effect caused by the pursuit of other—genuine—values such as Van Parijs’s ‘justice as equal dignity’.98 We recommend the ECJ, therefore, to focus its attention mainly on justifications grounded in respect for the fundamental interests of language groups in relation to their language, as well as in communication concerns.

B. Proportionality Obviously, the sole fact that a language requirement, imposed in a context where State neutrality was equally feasible, pursues a legitimate aim does not render the requirement lawful. For that to be true, the contested measure must also be proportionate to its aim. This means that the measure must be suitable and necessary to attain the aim, and be proportionate stricto sensu. i. Suitability A language requirement is unsuitable if it cannot achieve its objective. To give but one example: the requirement that candidate tenants of social houses located in Flanders must be willing to learn Dutch does not serve the purpose of effective communication if the mother tongue of both the social tenant and the social landlord is French.99 Most legal controversies, however, revolve around the necessity criterion, that is, the question of whether less restrictive but equally effective alternative measures exist. ii. Necessity As noted, we think that State neutrality towards language use in private contexts and towards the linguistic skills of applicants for social benefits should be the default position. When the national minority language has been granted (co-) official status in the sub-State nation’s polity, the inevitable promotion of the Member State’s majority language in the public sphere may be assumed to be offset and linguistic justice for the minority language to be achieved. Consequently, 97 F Grin, ‘Diversity as Paradigm, Analytical Device, and Policy Goal’ in W Kymlicka and A Patten (eds), Language Rights and Political Theory (New York, Oxford University Press, 2003) 171–72; F Grin, ‘On the Costs of Cultural Diversity’ in P Van Parijs (ed), Cultural Diversity versus Economic Solidarity (Brussels, De Boeck, 2004) 201. 98 Van Parijs, above n 11, 33–37. See also Kymlicka, above n 9, 121–23. 99 See the Belgian Council of State’s Opinion on the Flemish Housing Act, Travaux Préparatoires 2005–2006, No 824/1, p 59 (). See also Belgian Constitutional Court, No 101/2008, 10 July 2008, para B.36.4.

312 Elke Cloots and Stefan Sottiaux State intervention in private language use and language proficiency conditions for eligibility to social benefits would go beyond what is necessary for paying due respect to the linguistic interests of the minority language group. The language requirement in question would thus amount to a violation of the fundamental freedoms, unless it is justified by communication concerns. In certain circumstances, however, it might be argued that the use of the regional language by public institutions is not sufficient guarantee for a multinational Member State’s citizens to be treated with equal dignity, or to have access to a viable societal culture of their own, or to be reassured that their language group can flourish. If that is the case, a duty to use the regional language in certain private communications or to learn that language in exchange for social housing may indeed be necessary for the fulfilment of these language groups’ linguistic interests. Indeed, in order to ensure that a viable societal culture remains centred on the regional language and that its native speakers feel respected, immigrants in the region may need to be stimulated to learn the local vernacular. The prerequisite stimuli may be generated by making the language more ‘prestigious’ in the eyes of immigrants.100 For a regional language’s prestige to be sufficiently high, more may be needed than just granting it official status. It may be indispensable to that end that the local vernacular be not only the language of law, politics and education, but also that of economics, culture and community life. In other words, it may be essential that the region’s entire visage linguistique101 be based on the local language. Whether the necessity criterion is met or not in a given case will depend on factual circumstances, such as the immigration rates in the region and the likelihood of immigrants integrating into the local culture without being ‘stimulated’ by the language measure in question. iii. Proportionality Stricto Sensu Even if a regional law prescribing the language to be employed in private communications or making a social benefit conditional on proficiency in the local vernacular is considered suitable and necessary for the achievement of the legitimate aim, it may still turn out to be disproportionate and, therefore, in breach of EU law. Indeed, there is a third threshold to be met: proportionality in the narrow sense. This final stage of the proportionality test implies that the burdens the measure imposes on the individual must not be excessive compared to the benefits the measure entails.102 Although the ECJ usually stops short of verifying whether

100

See, eg Kymlicka, above n 1, 219 and 277–79; Van Parijs, above n 16, 184–85. In this vein, see Ford v Quebec, above n 77, paras 72–73. See also Kymlicka, above n 1, 286. 102 See especially R Alexy, A Theory of Constitutional Rights (J Rivers trans) (Oxford, Oxford University Press, 2002) 401. 101

EU Law and Language Regulation 313 the third element of the proportionality inquiry is met,103 it would be wise to give proportionality stricto sensu a more prominent role in linguistic cases. The added value this third limb may bring can, in our view, best be illustrated with an example. Imagine a regional law stipulating that only people who speak the local language are eligible to buy a house or flat in the region’s territory. The language condition clearly interferes with the freedom of movement of persons as guaranteed by EU law. By imposing this language requirement, the regional legislature seeks to ensure the viability of the regional societal culture. It leaves little doubt that the measure is suitable for the attainment of that objective. Moreover, in regions overwhelmed by immigrants who have virtually no incentive to learn the local language, the measure may also be necessary to achieve the desired end, as less restrictive means would be insufficient to protect the regional societal culture’s viability. If the proportionality inquiry stopped here, the language requirement would be found to comply with the fundamental freedoms. However, the burdens the language condition entails for individual persons are arguably so excessive that they outweigh the potential gains in the protection of the language group’s interests. Compared to the situation of an individual who is free from State interference, an individual who is subject to the language requirement necessarily suffers from one of two burdens: either he must invest money, time, etc in learning the regional language, or he loses the right to gain property in the region. It is obvious that both options are very onerous. Hence, the individual in question cannot avoid suffering a significant impairment when moving into the region. It is only at this third stage of the proportionality inquiry that account is taken of the extent to which the individual’s freedoms are interfered with, and not merely of the fact that they are interfered with.104 We consider this an important element of proportionality review in language cases, as the burdens language requirements impose on individuals may significantly differ. As Will Kymlicka observes: ‘Restricting the language of commercial signs is one thing; restricting the language of newspapers, churches, or private schooling … is another.’105 Although Kymlicka does not claim to have a definite answer to the question of how minority and immigrant protection may best be reconciled, he does convincingly argue that the degree of intrusion into individual freedom matters.106

103

See P Craig, EU Administrative Law (Oxford, Oxford University Press, 2006) 657. Theoretically, the extent of the rights interference might also be considered at the outset of the judicial review, ie before the ECJ starts monitoring the justification for the national measure. The ECJ might, for example, decide to apply a stricter proportionality test to language requirements in the social sphere than to language conditions in the commercial sphere. However, the ECJ seems rather unwilling to diversify at such an early stage of inquiry, and prefers to treat all free movement infringements the same way. 105 Kymlicka, above n 1, 288. 106 Ibid 287–89. 104

314 Elke Cloots and Stefan Sottiaux iv. Intensity Before concluding the present section, we should add a final word about the intensity with which the proportionality test is to be applied. The ECJ case law is not very clear on this point. Nonetheless, some commentators have detected certain factors that seem to influence the intensity of review. One such parameter is the relative importance of the interests invoked by the litigants before the Court. If the applicant claims that one of his fundamental freedoms, that is, one of the cornerstones of EU law, has been violated, the ECJ usually adopts a fairly intensive standard of review.107 The level of intensity generally decreases, however, when the Member State invokes an equally fundamental national interest, such as the protection of a constitutional right or the ‘national identity’, in defence of the contested measure.108 Conversely, one might expect the intensity of review to be heightened when, in addition to an interference with a fundamental freedom, a prima facie violation of a fundamental right has been found, such as the rights to freedom of expression, to respect for private and family life, and to social and housing assistance.109 Another factor that may impact on the rigour with which the ECJ monitors the proportionality of a national measure is the relative expertise and authority of the institutions involved.110 When it comes to the fundamental freedoms, the ECJ obviously has a greater expertise and authority than the national authorities. Again, this in principle warrants a ‘fairly intrusive’ proportionality review in the type of case under consideration. Yet domestic language policy typically lies within the competence of the Member States. Moreover, it touches upon very sensitive national issues, especially in the EU’s multinational Member States. This may weaken the authority of a supranational court like the ECJ to intervene, in particular when domestic constitutional mechanisms have been established for the prevention and resolution of conflicts between the Member State’s language groups, such as a domestic court composed of judges from each

107 Craig, above n 103, 704; G de Búrca, ‘The Principle of Proportionality and its Application in EC Law’ (1993) 13 Yearbook of European Law 105, 126. 108 See, eg United Pan-Europe Communications Belgium SA and Others, above n 6, paras 44–50. See generally, Craig, above n 103, 706–09; Gerards, above n 32, 94–96. Yet also in this type of case, the Court’s review may still be fairly intrusive. See, eg, Case C-438/05 International Transport Workers’ Federation and Finnish Seamen’s Union v Viking Line ABP and OÜ Viking Line Eesti [2007] ECR I-10779, paras 84–90; Case C-341/05 Laval un Partneri Ltd [2007] ECR I-11767, paras 107–11; Case C-271/08 Commission v Germany [2010] ECR I-7087, paras 53–66. 109 Interestingly, in Runevic ˇ-Vardyn and Wardyn, both elements were present. On the one hand, the Lithuanian national identity was at stake (above n 37, paras 86 and 91). On the other hand, not only the applicants’ fundamental freedoms, but also their right to respect for their private and family life was restricted (ibid paras 66 and 89–91). The ECJ decided to leave the ultimate decision as to the proportionality of the Lithuanian language law in the hands of the national court (ibid paras 77–78, 83 and 91). Yet it did not hesitate to provide detailed guidance (see especially ibid paras 92–93). 110 de Búrca, above n 107, 111 and 127; Gerards, above n 32, 118.

EU Law and Language Regulation 315 linguistic community.111 Furthermore, in the (quasi-)federal Member States here considered, language regulation is often based on more comprehensive and contextcontingent constitutional accords regarding the recognition of sub-State nations. It may be assumed that Member State authorities have more specialised expertise in these affairs than the ECJ. In particular when identity-related arguments (as opposed to communication concerns) are relied on in order to justify the regional language policy, the ECJ’s authority and expertise will be relatively poor. In that event, the Court should conduct its proportionality review with due deference to the domestic institutions’ determinations. More precisely, it would be wise for the ECJ to restrict its inquiry to: (a) examining whether the ground invoked is genuine, that is, not being used as a pretext for economic protectionism; and (b) setting the parameters the referring court should consider when monitoring the language measure’s suitability, necessity and strict proportionality (eg, the vulnerability of the language in question, immigration rates, etc).

IV. PRIVATE LANGUAGE REGULATION

A final type of language regulation we should like to explore in this chapter emerges not from State authorities, but from private persons. Consider, for example, a private employer requiring his workers to be proficient in the local vernacular, or to speak that language in the workplace. Or think of a private school, crèche or sports club imposing quotas limiting the admission of children whose parents do not speak the local language, or encouraging parents to communicate with their children in the local language. Given the intimate link between language and nationality, these language requirements once again seem to pose difficulties in the light of the Treaty provisions on freedom of movement and non-discrimination. The fact that these restrictive measures have not been imposed by State authorities but by private persons does not exclude them from the scope of the relevant Treaty provisions. The ECJ’s landmark judgment in Angonese, where the freedom of movement of persons and the ban on nationality discrimination were for the first time applied to an individual undertaking, removed any lingering doubt on this issue.112 Yet in that very judgment it was also

111 See, eg, Belgian Constitution, Arts 141–143; Belgian Special Law on the Constitutional Court, Arts 31, 33 and 55–56; Special Statute of the Trentino–Alto Adige Region and the Provinces of Trento and Bolzano, Arts 90–91, 93 and 97–98. 112 Angonese, above n 6, para 36. The applicability of the Treaty provision on sex discrimination to private undertakings had already been established at that time. See Case 43/75 Defrenne I [1976] ECR 455. See also R Lane and N Nic Shuibhne, ‘Case C-281/98, Roman Angonese v Cassa di Risparmio di Bolzano SpA, Judgment of 6 June 2000, not yet reported’ (2000) 37 CML Rev 1237, 1240 and 1243–44.

316 Elke Cloots and Stefan Sottiaux recognised that private employers may ‘require an applicant for a post to have a certain level of linguistic knowledge’.113 The ECJ did not specify, however, on what grounds a private person might be allowed to subject other individuals to linguistic requirements.114 Although private language measures are often directed at ensuring effective communication at work, at school, etc, they may also have another rationale, especially in multinational Member States. Indeed, some of these requirements result from individual people’s desire to protect their own culture and language from evils such as disrespect, extinction, etc. Although we concluded above that public authorities may rely on the latter type of considerations to justify the language requirements they impose, it seems far more controversial for a private person to invoke the fundamental linguistic interests of the language group to which he belongs. How are we to account for this intuition? What could be wrong about seeking to promote the interests of the group to which one feels attached? And is it not, after all, also in the interest of the individual that his linguistic community is able to thrive? Why would an individual not be permitted to promote his own interests, while the State or regional polity he lives in may legitimately do so? Since these questions have hardly been considered by courts and have received little attention in legal doctrine, it is instructive to see what moral and political philosophy may teach us. The work of Joseph Raz may be a good starting point here. It is largely uncontested that an individual may defend his own interests, including the interests he holds as a group member.115 Yet it is one thing to say that an individual is allowed to promote his interests, but quite another to claim that his interests are sufficient for holding others to be bound by duties. To be sure, some individual interests are a sufficient reason for holding others to be under a duty; these interests are what Raz calls ‘rights’.116 But not all individual interests are rights. According to Raz, an individual may have very important interests which nonetheless do not justify holding another person to be dutybound. Thus, on Raz’s account, ‘one may … doubt the possibility of a justification for a fundamental individual right to a collective good.’117 By a collective good, Raz understands a special type of public good. A good is a public good in a given society ‘if and only if the distribution of its benefits in that society is not subject to voluntary control by anyone other than each potential

113

Angonese, above n 6, para 44. The Advocate General, by contrast, did briefly touch upon the issue. He stated more particularly that ‘it is obvious, in the light of the linguistic regime in the province of Bolzano and of the linguistic make-up of its population, that the defendant was entitled to require its potential employees to give evidence of bilingualism’. See Angonese, above n 6, Opinion of AG Fennelly, para 42. 115 See De Schutter, above n 19, 116. 116 J Raz, The Morality of Freedom (Oxford, Clarendon Press, 1986) 166. 117 Ibid 208. See also ibid 202–03; L Green, ‘Two Views of Collective Rights’ (1991) 4 Canadian Journal of Law and Jurisprudence 315, 321–24; N MacCormick, Legal Right and Social Democracy (Oxford, Clarendon Press, 1982) 143; D Réaume, ‘Individuals, Groups, and Rights to Public Goods’ (1988) 38 University of Toronto Law Journal 1. 114

EU Law and Language Regulation 317 beneficiary controlling his share of the benefits.’118 In other words, members of that society cannot be involuntarily excluded from the enjoyment of the good. Whereas the non-excludability from certain public goods is contingent, on technological limitations for instance, other goods are inherently non-excludable. It is only the latter category of public goods that Raz characterises as ‘collective goods’.119 Examples of collective goods are the general beneficial features of a society, such as living in a beautiful environment and in a prosperous, cultured and tolerant society,120 but also the self-determination of the community to which one belongs.121 Individual persons clearly have an interest in these collective goods, interests arising from their membership in a group. However, no single group member’s interest is, in and of itself, sufficient for holding another person to be under a duty to satisfy that interest.122 Importantly, Raz’s account of collective goods should not be interpreted as a denial of the existence of any rights to these goods. The bearer of the right, however, is not an individual group member but the group as such.123 Although several theorists agree with Raz’s assumption that no individual rights to collective goods exist,124 there is considerable controversy as to why this would be the case. According to Raz himself, the satisfaction of collective goods such as self-determination makes enormous demands on a great number of people. In his view, demands of such a far-reaching nature cannot possibly be justified by a single individual’s interest in the collective good. As a result, individual persons cannot have a right to collective goods.125 Denise Réaume, however, provides an alternative explanation. Her account of the conceptual impossibility of individual rights to (certain) collective goods is based on the nature of these goods and the way they are enjoyed.126 Réaume characterises the collective goods Raz has in mind (eg, living in a cultured society) as participatory goods, that is, as goods that cannot be enjoyed individually, even though they are enjoyed by individuals. The individual needs others to enjoy the good in question.127 Moreover, the necessary 118 Raz, above n 116, 198. But see Green, above n 117, 321; Green, above n 18, 659–60; Réaume, above n 117, 3, fn 10. According to Leslie Green and Denise Réaume, public goods are not only inexcludable, but also non-rival in consumption, meaning that consumption of the good by one individual does not prevent simultaneous consumption by other individuals. 119 Raz, above n 116, 199. Green, by contrast, distinguishes between ‘contingent’ and ‘shared’ public goods. The public aspect of shared goods is ‘not merely a contingent feature of their production but partly constitutes what is valuable about them’. See Green, above n 117, 321. Green’s ‘shared’ goods and Raz’s ‘collective’ goods do not neatly coincide though. Whereas friendship, for instance, is considered a shared good, Raz does not characterise it as collective. Compare Green, above n 117, 321, and Raz, above n 116, 199. 120 Raz, above n 116, 201–03. 121 Ibid 207–09. 122 Ibid 208. 123 Ibid 208. See also Réaume, above n 117, 17 ff. 124 For an exception, see J Morauta, ‘Rights and Participatory Goods’ (2002) 22 Oxford Journal of Legal Studies 91. 125 Raz, above n 116, 207. 126 Réaume, above n 117, 7. 127 Ibid 8–9.

318 Elke Cloots and Stefan Sottiaux participation of many is constitutive of the good’s value.128 Because a participatory good cannot be enjoyed by an individual as an individual, an individual cannot have a right to this type of good.129 Still others, like Michael McDonald, contend that collective goods serve the interests of the group rather than its individual members. Since rights are generally most properly exercised by their beneficiaries, the community itself, not an individual member, should exercise the right to a collective good.130 If these theorists are right in that a single individual’s interest in a collective or participatory good is not sufficient justification for imposing duties on others, the question arises as to whether the fundamental interests individuals have in relation to their language constitute interests in such goods. If that question is answered in the affirmative, we may safely conclude that an individual cannot successfully invoke these interests as a ground of justification for imposing duties on other individuals. Many theorists would probably agree that living in a society where one’s mother tongue and associated linguistic identity are treated with respect, where one’s language group can flourish, where one may freely express and develop one’s cultural and linguistic identity, etc, are collective or participatory goods.131 For one thing, the good may be considered collective in the Razian sense of the word. It might indeed be assumed that it is inherently impossible to exclude any member of the language group in question from the diffuse benefits living in such a society provides. Réaume, for her part, would likely identify it as a participatory good.132 The way she describes an aspect of the value people attach to their language supports this assumption: For the group as a whole, its language is a collective accomplishment. An individual member’s use of the language is at once a participation in this accomplishment and an expression of belonging to the community that has produced it. Because this participation has intrinsic value, members of a language community identify with that language … 133

In other words, the good related to the use of one’s language Réaume envisages here is the participation in one’s linguistic community. If it is true that the good in question is collective and participatory, and that it, therefore, cannot ground individual rights, the intuition we expressed at the outset 128

Ibid 10. Ibid 7 and 10–11 and 13. For a critique, see Morauta, above n 124. 130 M McDonald, ‘Should Communities Have Rights? Reflections on Liberal Individualism’ (1991) 4 Canadian Journal of Law and Jurisprudence 217, 232. For a critique, see Green, above n 117, 324–25 (criticising McDonald’s confusion between holding and exercising a right). 131 Raz, above n 116, 207. 132 Réaume, above n 117, 23; DG Réaume, ‘The Group Right to Linguistic Security: Whose Rights, What Duties?’ in J Baker (ed), Group Rights (Toronto, University of Toronto Press, 1994); DG Réaume, ‘Official Language Rights: Intrinsic Value and the Protection of Difference’ in W Kymlicka and W Norman (eds), Citizenship in Diverse Societies (New York, Oxford University Press, 2000) 266. Green prefers the term ‘shared’ good to describe a public good one cannot enjoy individually but only jointly with others (Green, above n 18, 659–60). 133 Réaume, ‘Official Language Rights’, above n 132, 251. See also Réaume, ‘Individuals’, above n 117, 10 and 17-26. 129

EU Law and Language Regulation 319 may be confirmed: a private person is not permitted to restrict other individuals’ personal autonomy for the sake of ensuring the satisfaction of his interest in living in a societal culture based on his own language. Consequently, the ECJ has good reasons for dismissing the arguments invoked by a private employer or school director in defence of the linguistic requirements he imposes on his employees and pupils in so far as these arguments are based on the employer’s and director’s interests as members of their language group. Besides, the mode of adjudication proposed here also seems to be the approach followed in the American case law and scholarly literature. It is commonly considered unlawful for private employers in the US to oblige their workers to speak English with a view to promoting the national language. The pursuit of a public policy objective like this is considered insufficiently related to the operation of a business and cannot therefore be successfully relied upon by an employer.134 A more promising ground of justification would be the private employer’s interest in effective communication between his employees and the firm’s (potential) clients, and between the employees themselves. Similarly, it may be considered legitimate for a school director to purport to facilitate communication between the teachers and the pupils of his school. Admittedly, it may be objected that communication is also a participatory good. Communication is a good which can be neither created nor enjoyed individually. Therefore, a single individual’s interest in communication cannot justify others being subject to duties. Whilst we do not pretend that it is conceptually flawed to regard communication as a participatory activity or to characterise an undertaking or school as a group, it is essential to recall that it is the head of an undertaking/school we are considering here. Contrary to his individual workers or pupils, the head may try to serve the interests of his business or school by imposing duties on his workers and pupils, just as a State’s government, but not its individual citizens, may try to realise public goods by imposing duties on State nationals.

V. CONCLUSION

This chapter offers a preliminary proposal of a mode of adjudication the ECJ could employ when called upon to decide on a language regulation adopted by public or private persons in federalised multinational Member States. The proposal is built around two major distinctions. A first distinction is based on whether the (public) language regulation occurs in a context where State neutrality to linguistic identity was possible or not. In the event State neutrality is impossible, considerations relating 134 See, eg Gutierrez v Municipal Court of Southeast Judicial District, Los Angeles County, 838 F 2d 1031, 1042–44 (CA9 Cal 1988); Garcia v Gloor, 618 F 2d 264, 267 (CA Tex 1980); J Leonard, ‘Bilingualism and Equality: Title VII Claims for Language Discrimination in the Workplace’ (2004) 38 University of Michigan Journal of Law Reform 57, 135; JF Perea, ‘English-Only Rules and the Right to Speak One’s Primary Language in the Workplace’ (1990) 23 University of Michigan Journal of Law Reform 265, 306–07.

320 Elke Cloots and Stefan Sottiaux to economic mobility and equal treatment of distinct nationalities should be left out of the story. Hence, it has been submitted that the designation of a Member State’s official languages should be excluded from the scope of the EU norms on freedom of movement and nationality discrimination. The rationale underlying our proposal is that a State has to determine the languages to be employed by its public institutions and therefore cannot avoid promoting the use of certain languages over others. In that context, it is of paramount importance that the EU’s multinational Member States are free to design their official language policy in a fair way, unconstrained by concerns of economic mobility across Member States. This rationale cannot be extended to all public language regulation, however. Neither laws regulating the language to be used in communication between private persons, nor language proficiency conditions attached to eligibility for non-communicative social benefits are covered by it. In these situations, State neutrality clearly is an option. If State authorities nonetheless decide to interfere with private persons’ autonomy, they should justify this, including in the light of EU law. We have argued that concerns of communication and, to a lesser extent, the need to respect the fundamental interests of language groups in relation to their language may serve as legitimate grounds of justification in this regard. This brings us to the second distinction, that is, the division between public and private language regulation. It has been argued that whereas public authorities may rely on fundamental linguistic interests in order to justify a prima facie violation of the fundamental freedoms and the non-discrimination prohibition, private persons cannot. We are well aware that, in numerous respects, the research conducted in this chapter is too limited to be perceived as a full-fledged analysis of the normative questions surrounding domestic language policy. For one thing, boundaries between public and private language use are far from clear-cut. What about, for example, the many privately-run but publicly-funded schools, hospitals, railway companies, etc?135 Secondly, our study is restricted to the Treaty provisions on freedom of movement and nationality discrimination. The compatibility of the Member States’ language policies with fundamental rights, by contrast, has not been assessed. Moreover, an abundance of relevant secondary legislation, such as the Race Equality Directive,136 the Product Labelling Directive137 and the Migrant Workers Regulation,138 has not been touched upon. In several of these acts,

135 See also N Nic Shuibhne, EC Law and Minority Language Policy (The Hague, Kluwer Law International, 2002) 2. 136 See n 56 above. 137 European Parliament and Council Directive (EC) 2000/13 on the approximation of the laws of the Member States relating to the labelling, presentation and advertising of foodstuffs [2000] OJ L109/29, Art 16. 138 Council Regulation (EEC) 1612/68 on freedom of movement for workers within the Community [1968] OJ L257/2, Art 3(1). This Regulation was codified by European Parliament and Council Regulation (EU) 492/2011 on freedom of movement for workers within the Union [2011] OJ L141/1.

EU Law and Language Regulation 321 the balancing exercise between linguistic interests and freedom of movement/ non-discrimination has already been conducted by the EU legislature, such that the ECJ’s room for manoeuvre is more limited in these fields. Lastly, the focus of this chapter is on the impact of EU law on language policies in Member States which have gone through a process of devolution. It would be interesting to investigate the relevance of our findings to language laws enacted in other Member States. We may think, for example, of Member States with a unitary structure which have granted co-official status to a language spoken by a national minority. The minority in question sometimes consists of a territorially concentrated group of citizens, as is the case in Slovenia,139 but its members may also be geographically dispersed, as in Ireland.140 In addition, further research is needed on the relationship between EU law and the language policies of Member States where sub-State national societies enjoy neither political autonomy nor official language status. This chapter should, therefore, be regarded as a limited contribution to a wider and continuing debate, not as a full and definitive account.

139 See Art 11 of the Slovenian Constitution, granting Italian and Hungarian regional co-official status, along with Slovenian. See also Arts 62 and 64. 140 See Art 8 of the Irish Constitution, designating Irish and English as the State’s official languages.

14 The European Court of Justice and Member State Federalism: Balancing or Categorisation? ELKE CLOOTS*

To the Sirens first shalt thou come, who beguile all men whosoever comes to them. … But if thou thyself hast a will to listen, let them bind thee in the swift ship hand and foot upright in the step of the mast, and let the ropes be made fast at the ends to the mast itself, that with delight thou mayest listen to the voice of the two Sirens. And if thou shalt implore and bid thy comrades to loose thee, then let them bind thee with yet more bonds. (Homer, Odyssey, Book 12 (AT Murray trans, Loeb Classical Library, Cambridge, Mass, Harvard University Press, 1919))

I. INTRODUCTION

C

ONSTITUTIONS TEND TO cover a tremendous variety of topics, ranging from the name of the first King (if any), to the colours of the national flag and the defence of the Fatherland, to the country’s economic development. Common to most constitutional orders, however, is that they include rules and principles concerning human rights and State structure (unitary, federal or something in between).1 That is no different for the constitutional traditions of the 27 EU Member States, albeit that the actual content of those rights and State structures varies greatly from Member State to Member State.

* PhD Fellow of the Research Foundation—Flanders (FWO), Institute for Constitutional Law, Katholieke Universiteit Leuven. I am most grateful to the workshop participants for helpful comments and criticisms. Special thanks go to Pavlos Eleftheriadis, Andreas Føllesdal, Koen Lenaerts, Giuseppe Martinico, Emmanuel Slautsky and Stefan Sottiaux. All mistakes are of course my own. 1 In the context of this contribution, the terms ‘(governance) structure’ or ‘structural’ refer exclusively to the unitary or (quasi-)federal character of a Member State. They do not concern any other sort of constitutional provisions which might be labelled ‘structural’, such as provisions relating to the separation of powers or the republican status of a Member State.

The ECJ and Member State Federalism 323 Constitutionally entrenched rights and governance structures often embody fundamental societal values and/or reflect important aspects of national identity. Although the Union institutions are now under a Treaty obligation to respect the ‘national identities [of the Member States], inherent in their fundamental structures, political and constitutional’,2 it is not difficult to imagine that the need to observe constitutional rights and structural provisions may be in tension with the Treaty provisions on freedom of movement. Two examples suffice to illustrate this. It is obvious, for instance, that a trade union exercising its constitutionally protected right to take collective action may hinder a company’s pursuit of economic activities in another Member State. And it is equally plausible that internal federalisation processes re-erect within a single Member State barriers to freedom of movement which have been abolished among the Member States. One might expect the European Court of Justice (ECJ) to have developed a uniform method for adjudicating these sorts of clashes between national constitutional provisions and the Treaty provisions on freedom of movement. Yet a cursory glance at the Court’s case law readily reveals that this assumption is false. When the ECJ finds a restriction of freedom of movement, it permits the defendant in question to rely on the need to protect or observe fundamental rights to justify the measure. In contrast, the Court refuses to regard a Member State’s system of power division as a ground of justification. Thus, while the internal market freedoms are to be balanced against constitutional rights on a case-by-case basis, they always outweigh requirements flowing from a Member State’s governance structure. The observation that the ECJ treats a Member State’s constitutional rights and structural provisions differently is the starting point of this chapter. In section II., we dig into the roots of this distinction. The concerns underlying the Court’s seemingly preferential treatment of constitutional rights will be explored. In section III., it is argued that the ECJ’s refusal to balance the fundamental freedoms against a Member State’s governance structure at the justification stage does not mean that the Court is insensitive to the latter type of constitutional provisions. More precisely, this contribution demonstrates that the ECJ generally takes account of a Member State’s governance structure by limiting the scope of application of EU law. The Court’s approach to State structure cases is characterised as ‘categorical’ and ‘rulebased’, and examined in the light of the balancing/categorisation and standards/ rules debates prevalent in US constitutional law. The merits and pitfalls of the ECJ’s categorical and rule-based approach are the subject of section IV. of this chapter.

II. DISTINCT TREATMENT OF CONSTITUTIONAL RIGHTS AND FEDERALISM PROVISIONS

The paradigm situation with which we are concerned here is as follows. The Treaty rights to freedom of movement are prima facie violated, and the defendant 2

Art 4(2) TEU.

324 Elke Cloots before the ECJ proffers arguments based on its Member State’s constitutional order to justify the interference. Does this defence have any chance of success? The answer to this question will clearly differ depending on whether the State’s constitutional rights or its (quasi-)federal structure are being invoked. In this section, we examine the ECJ’s approach in both types of case and attempt to lay bare the rationale that underlies the distinction in the case law.

A. Constitution-based Limitations of Freedom of Movement i. Constitutional Rights The line of cases in which the ECJ has formally accepted that the need to protect the fundamental rights embodied in a Member State’s constitutional order may be invoked to justify an obstacle to freedom of movement is well-known and has often been discussed. In Schmidberger, Omega, Viking, Laval, Dynamic Medien and, more recently, Commission v Germany and Sayn-Wittgenstein, the ECJ regarded the protection of fundamental rights as a ‘legitimate interest’ which, in principle, justifies a restriction on the fundamental freedoms. According to the Court, the protection or exercise of the fundamental rights at issue must be ‘reconciled with’ the right to free movement along the lines of the proportionality principle.3 For the proportionality test to be met it is not, however, required that the domestic rights conception at issue is shared by all other Member States. In Omega, for instance, the particular interpretation of the principle of respect for human dignity prevalent in German constitutional law was upheld against the internal market argument.4 Likewise, in Dynamic Medien, the Court left room for the specific German view of the rights of the child when assessing the proportionality of a Member State measure restricting the free movement of goods.5 Finally, in Sayn-Wittgenstein it was held that Austria’s distinctive implementation of the principle of equal treatment may legitimate a limitation of the freedom to move to and reside in another Member State conferred upon citizens of the Union.6

3 Case C-112/00 Schmidberger [2003] ECR I-5659, paras 74, 77, 81 and 82; Case C-36/02 Omega [2004] ECR I-9609, paras 35–36; Case C-438/05 International Transport Workers’ Federation and Finnish Seamen’s Union v Viking Line ABP and OÜ Viking Line Eesti [2007] ECR I-10779, paras 45–46 and 77; Case C-341/05 Laval un Partneri Ltd [2007] ECR I-11767, paras 93–94 and 103; Case C-244/06 Dynamic Medien Vertriebs GmbH [2008] ECR I-505, para 42; Case C-271/08 Commission v Germany [2010] ECR I-7087, para 44; Case C-208/09 Sayn-Wittgenstein (ECJ, 22 December 2010), paras 83–85 and 90. See also Case C-81/09 Idryma Typou (ECJ, 21 October 2010), Opinion of AG Trstenjak, paras 80, 83, 86–88 and 120. 4 Omega, above n 3, paras 37–38. 5 Dynamic Medien, above n 3, paras 44 and 49. 6 Sayn-Wittgenstein, above n 3, para 91.

The ECJ and Member State Federalism 325 It is clear that the search for ‘reconciliation’7 or a ‘fair balance’8 between competing interests, and the belief that that objective can be achieved through proportionality review,9 are central to the judgments under consideration. ii. Constitutional State Structure Provisions We now turn our attention to the relationship between structural constitutional provisions and the freedom of movement guaranteed by EU law, where a rather different picture emerges. Indeed, the ECJ refuses to regard the need to respect a Member State’s federal or devolved structure as a legitimate interest to which a Member State (or its regions) may resort in order to justify an interference with the freedom of movement. A notable example of the ECJ’s doctrine in this respect may be found in Government of the French Community and Walloon Government v Flemish Government (the ‘Flemish Care Insurance Scheme Case’). In that judgment, the ECJ held that a residence requirement for entitlement to a regional social benefit—the Flemish care insurance scheme—constituted an obstacle to freedom of movement for workers and freedom of establishment.10 In an attempt to justify this obstacle, the Flemish government invoked the requirements inherent in the division of powers within the Belgian federal structure.11 Under Belgian constitutional law, the autonomous sub-State entities have, as far as ‘aid to persons’ is concerned, power over the people who reside (as opposed to work) in their respective territory.12 Yet the Court readily dismissed the Flemish defence argument by recalling its long-established case law according to which ‘a Member State cannot plead provisions, practices or situations prevailing in its domestic legal order, including those resulting from the constitutional organization of that State, to justify the failure to observe obligations arising under Community law’.13

7 Schmidberger, above n 3, para 77; Viking, above n 3, para 46; Laval, above n 3, para 94; Commission v Germany, above n 3, paras 44, 51, 55, 58 and 66. 8 Schmidberger, above n 3, paras 81 and 90; Viking, above n 3, para 79; Laval, above n 3, para 105; Commission v Germany, above n 3, para 52; Sayn-Wittgenstein, above n 3, para 83. See also Idryma Typou, above n 3, Opinion of AG Trstenjak, paras 86, 107 and 124. 9 Schmidberger, above n 3, paras 79, 80 and 82; Omega, above n 3, para 36; Viking, above n 3, para 46; Laval, above n 3, para 94; Dynamic Medien, above n 3, paras 42 and 46; Commission v Germany, above n 3, para 44; Sayn-Wittgenstein, above n 3, para 90. See also Idryma Typou, above n 3, Opinion of AG Trstenjak, paras 86–107 and 121–24. 10 Case C-212/06 Government of the French Community and Walloon Government v Flemish Government [2008] ECR I-1683, para 54. 11 Ibid para 57. 12 See, eg, J Velaers and J Vanpraet, ‘De materiële en territoriale bevoegdheidsverdeling inzake sociale zekerheid en sociale bijstand (II)’ (2009) Tijdschrift voor Bestuurswetenschappen en Publiekrecht 195, 199–200; H Verschueren, ‘La régionalisation de la sécurité sociale en Belgique à la lumière de l’arrêt de la Cour de Justice européenne portant sur l’assurance-soins flamande’ (2008) 2 Revue Belge de Sécurité Sociale 173, 216–19. 13 Government of the French Community and Walloon Government, above n 10, para 58.

326 Elke Cloots The crux of the decision was not that the free movement restriction was disproportionate but, rather, that State structure-related considerations were not among the reasons that might be invoked to justify the residence requirement.14 The Belgian federal structure was simply ruled out as a legitimate excuse. The language of balancing, reconciliation and proportionality, so prominent in the constitutional rights cases, was abandoned. iii. Conclusion At first sight, the above outline of the ECJ case law seems to justify the conclusion that the Court is more sensitive to constitutional rights protection than to the Member States’ structural constitutional provisions, at least in the context of EU free movement law. This is of course not to say that the ECJ’s approach to Member States’ constitutional rights traditions would be immune from criticism. Quite the contrary—the Court’s decisions have also been heavily criticised for being insufficiently protective of constitutional rights. So, the ECJ has been reproached for hiding a hierarchical view of the relationship between fundamental freedoms and constitutional rights behind its balancing talk. The Court’s hierarchical vision is derived, more precisely, from the way in which the balancing test is designed and administered. First, the terminology used by the ECJ to denote the distinct interests in the balance is telling. The Court presents the protection of fundamental rights as a mere ‘legitimate interest’ or ‘public policy’ reason which must be balanced against the ‘fundamental’ economic freedoms.15 Secondly, the structure of the Court’s balancing act too seems to suggest a hierarchy of rights, with internal market rights at the top. So, in order to determine whether a fair balance was struck between both interests, the Court assesses whether the exercise of the fundamental right in question constitutes a suitable and necessary limitation of freedom of movement. Yet the converse question of whether the exercise of the Treaty rights is a proportionate restriction on fundamental rights is not posed, let alone verified.16 Thirdly, the Court sometimes replaces the interest of the exercise of a given fundamental right with related but more ordinary policy interests. So, although the ECJ qualified the right to take collective action and the right to bargain collectively as fundamental rights,17 the public interest which was balanced against the fundamental freedoms was not the exercise as such of the fundamental social

14 Cp, eg, S Tsakyrakis, ‘Proportionality: An Assault on Human Rights?’ (2009) 7 International Journal of Constitutional Law 468, 493. 15 See, eg, E Spaventa, ‘Federalisation Versus Centralisation: Tensions in Fundamental Rights Discourse in the EU’ in M Dougan and S Currie (eds), 50 Years of the European Treaties (Oxford, Hart Publishing, 2009) 353–63. 16 For a critique of this approach see, eg, Commission v Germany, above n 3, Opinion of AG Trstenjak, paras 81, 84, 188, 190–95 and 199; Idryma Typou, above n 3, Opinion of AG Trstenjak, para 86. See also Spaventa, above n 15, 356–58. 17 Viking, above n 3, para 44; Laval, above n 3, para 91; Commission v Germany, above n 3, para 37.

The ECJ and Member State Federalism 327 right but, rather, the protection of workers.18 According to Advocate General Trstenjak, the requirement of an additional reason of public interest supplementing the rights-based justification suggests that the Court ranks fundamental freedoms above fundamental rights.19 Lastly, in the type of case under consideration, the Court tends to apply the proportionality test with considerable intensity. True, in several judgments, the Member State in question was expressly granted a ‘margin of discretion’ in deciding what constitutes a fair balance between its constitutional rights and the fundamental freedoms.20 Moreover, the necessity of a Member State’s rights protection system was not rejected for the sole reason that other Member States adhere to a less rigorous protection system.21 Nonetheless, the Court’s actual assessment of the suitability and necessity of a particular measure or action implementing a constitutional right often remains fairly intrusive, especially when social rights are involved.22 It should not surprise, therefore, that, in social rights cases, the scales have consistently been tipped towards the fundamental freedoms.23 Taken together, these concerns suggest that the difference between the Court’s approach to constitutional rights, on the one hand, and State structure, on the other hand, may well be less significant than initially thought. Indeed, it might be asserted that the Court ranks freedom of movement more highly than a Member State’s particular constitutional traditions, whether they concern State structure or fundamental rights. Nonetheless, we do not believe that the distinction drawn at the beginning of this contribution has entirely lost its relevance. Even if one accepts the soundness of the criticisms outlined above, the Court’s attitude remains importantly dissimilar in at least two respects.

18 Viking, above n 3, paras 77, 80, 81and 90; Laval, above n 3, paras 102, 103 and 107–08; Commission v Germany, above n 3, paras 51–53 and 66. See also Commission v Germany, above n 3, Opinion of AG Trstenjak, paras 179–82. But see Schmidberger, above n 3, paras 65–69. 19 Commission v Germany, above n 3, Opinion of AG Trstenjak, paras 183–85. 20 Schmidberger, above n 3, paras 82, 89 and 93; Omega, above n 3, para 31; Dynamic Medien, above n 3, para 44; Sayn-Wittgenstein, above n 3, para 87. 21 Omega, above n 3, paras 37–38; Dynamic Medien, above n 3, para 49; Sayn-Wittgenstein, above n 3, para 91. 22 See especially Schmidberger, above n 3, paras 82–94; Viking, above n 3, paras 84–90; Laval, above n 3, paras 107–11; Commission v Germany, above n 3, paras 53–66. Notable counter-examples are Omega (above n 3, paras 39–40), Sayn-Wittgenstein (above n 3, paras 92–94) and, to a lesser extent, Dynamic Medien (above n 3, paras 46–51). On the intensity of the ECJ’s proportionality review of Member State action, see far more extensively, P Craig, EU Administrative Law (Oxford, Oxford University Press, 2006) 706–10; G de Búrca, ‘The Principle of Proportionality and its Application in EC Law’ (1993) 13 Yearbook of European Law 105, 126–49; J Gerards, ‘Pluralism, Deference and the Margin of Appreciation Doctrine’ (2011) 17 European Law Journal 80, 94–96. 23 See the decisions in Viking, Laval and Commission v Germany (above n 3). Admittedly, in Viking, the ultimate proportionality inquiry was formally left to the referring national court (paras 80, 83–85 and 87). However, the detailed ‘guidance’ provided by the ECJ (para 85) left the national court with little choice other than to declare the collective action disproportionate. See also C Barnard, ‘Fifty Years of Avoiding Social Dumping? The EU’s Economic and Not So Economic Constitution’ in Dougan and Currie (eds), above n 15, 315–16 and 341–42; P Syrpis and T Novitz, ‘Economic and Social Rights in Conflict: Political and Judicial Approaches to their Reconciliation’ (2008) 33 EL Rev 411, 425.

328 Elke Cloots First, from a formal point of view, it cannot be denied that the Court’s line of reasoning varies according to the constitutional tradition in issue. The balancing and proportionality talk that pervades the judgments involving a clash of fundamental freedoms and constitutional rights, however spuriously used it may be, is totally absent in the Court’s decisions concerning a Member State’s (quasi-)federal structure. Thus, it might be said that, in the latter type of case, the ECJ does not even attempt to mask its hierarchical agenda. Secondly, it is quite obvious that constitutional rights have at least a theoretical chance of prevailing over fundamental freedoms in a particular case. In contrast, a justification based on the Member State’s structure is doomed to fail. For the time being, we stick therefore to the hypothesis posited at the outset of this chapter: the ECJ is more sensitive to constitutional rights than it is to constitutional State structure provisions. Let us now consider what the reasons for this differential appraisal of national constitutional traditions might be.

B. The Rationale behind the Distinction The difference in treatment of constitutional traditions concerning rights protection, on the one hand, and division of powers, on the other hand, may have multiple theoretical underpinnings. To begin with, there might be textual reasons for according more significance to rights than to State structure. A second motivation might possibly be found in the nature of the Union’s objectives. Thirdly, it is often contended that human rights are intrinsically more valuable than federalism. Lastly, it is conceivable that the ECJ has developed alternative, more appropriate decision-making techniques for showing sensitivity to the Member States’ governance structures. In what follows, we shall take a closer look at each of these candidate reasons. i. Textual Arguments Let us start our exploration with a glance at the Treaties. Do they contain a provision mandating the Court to accord a special position to constitutional rights protection? A first provision worth considering is Article 4(2) TEU, which stipulates that the Union shall respect the Member States’ ‘national identities, inherent in their fundamental structures, political and constitutional’. Yet it becomes immediately apparent that this Treaty provision cannot take us very far. Indeed, there is no indication that Article 4(2) solely requires respect for national rights. Quite the contrary—an amendment by the Treaty of Lisbon made clear that ‘regional and local self-government’ belong to the national identities which ought to be respected.24 Moreover, Advocate General Kokott had already referred

24 See generally N MacCormick, ‘The European Constitutional Convention and the Stateless Nations’ (2004) 18 International Relations 331, 337.

The ECJ and Member State Federalism 329 to the Treaty obligation to respect national identities in a case involving regional autonomy even before the Lisbon Treaty removed all doubt on this point.25 A more plausible textual reason for privileging constitutional rights may be found, however, in the EU Charter of Fundamental Rights (‘the Charter’). The Charter contains a ‘Bill of Rights’ for the European Union26 and is legally binding since the entry into force of the Lisbon Treaty.27 The Charter is primarily addressed to the Union institutions.28 This means, in the first place, that the EU legislature must comply with the Charter when enacting secondary legislation. However, it has been argued that the Union courts, being Union institutions as well, are equally subject to the Charter.29 Hence, they should interpret and apply EU law, including the Treaties, in accordance with the Charter. Arguably, a reading of the Treaty provisions on freedom of movement in line with the Charter implies that a Member State’s observance of the free movement requirements should not obstruct compliance with the Charter. As a consequence, a Member State should be allowed to rely on the Charter rights to justify its free movement restrictions. Since the Charter does not serve to protect a Member State’s governance structure or a nation’s right to self-government, there is no analogous Charter-based duty for the ECJ to accept these objectives as grounds of justification. Yet there are at least two reasons why we believe that the contrast found in the ECJ case law cannot be fully accounted for on the basis of the Charter. First, even if one accepts that Article 51(1) of the Charter obliges the ECJ to respect the Charter rights when interpreting the Treaty provisions on freedom of movement, it is questionable whether that obligation, in and of itself, requires the ECJ to balance freedom of movement against a Member State’s own, distinctive understanding of rights protection.30 Yet the Court has, on several occasions, done exactly that. True, it must be admitted that, in recent judgments, the ECJ supplemented Member State arguments based on constitutional rights with references to the corresponding Charter provisions, apparently to put the legitimacy of the national action’s objective beyond any possible doubt.31 However, these references to the Charter do not alter the fact that the Court, when determining the content of the fundamental 25 Joined Cases C-428/06 to C-434/06 UGT-Rioja and Others [2008] ECR I-6747, Opinion of AG Kokott, para 54. 26 See, eg K Lenaerts and E De Smijter, ‘A “Bill of Rights” for the European Union’ (2001) 38 CML Rev 273. 27 Art 6(1) TEU. 28 Art 51(1) Charter. 29 See, eg, Craig, above n 22, 501. 30 Arguably, that requirement could be implied from the reference to the principle of subsidiarity in Art 51(1) of the Charter. According to Paolo Carozza, the application of the subsidiarity principle to the Charter might permit the ECJ to take into account the distinct domestic rights conceptions in its interpretation and application of the Charter. However, Carozza also admits that ‘it is far from selfevident how subsidiarity should relate concretely to the implementation of the Charter’. See P Carozza, ‘The Member States’ in S Peers and A Ward (eds), The EU Charter of Fundamental Rights: Politics, Law and Policy (Oxford, Hart Publishing, 2004) 56–57. 31 See, eg Viking, above n 3, paras 43–44 (Art 28 Charter); Laval, above n 3, paras 90–91 (Art 28 Charter); Dynamic Medien, above n 3, para 41 (Art 24 Charter); Commission v Germany, above n 3, paras 37–38 (Art 28 Charter); Sayn-Wittgenstein, above n 3, para 89 (Art 20 Charter).

330 Elke Cloots right, paid deference to the national constitutional tradition in question and did not expect a rights conception common to the Member States.32 Secondly, as a matter of fact, the ECJ already used to take a more generous approach towards the protection of constitutional rights than to a Member State’s structural arrangements long before the Charter had acquired legally-binding force. The landmark judgments in this regard, that is, Schmidberger and the Flemish Care Insurance Scheme Case, all date from before the Lisbon Treaty’s entry into force. On these grounds it may be assumed that, prior to and perhaps in addition to the Charter, another motivation lies behind the Court’s relatively receptive attitude towards domestic rights. ii. A Duty of Respect on the Part of the Union To identify the ECJ’s motives, we need to dig deeper into the relevant case law. Most notably, it ought to be examined what leads the Court to qualify the protection of constitutional rights as a ‘legitimate interest’. The wording of the Court’s decisions is illuminating. Thus, the ECJ is willing to accept the protection of fundamental rights as a legitimate interest which may justify free movement restrictions ‘since both the Community and its Member States are required to respect fundamental rights’.33 More precisely, fundamental rights are considered to be ‘an integral part of the general principles of law the observance of which the Court ensures’.34 Hence, the ECJ shows respect for national constitutional rights not because they are of constitutional rank35 but, rather, because the Member States as well as the EU itself must respect fundamental rights. Using a contrario reasoning, the dismissal of a Member State’s structural arrangements as a ground of justification for free movement restrictions might be explained as follows: contrary to fundamental rights protection, power devolution in the Member States is not worthy of respect from the Union as such. Might it be possible that this proposition is true? It is indeed plausible to assume that the ECJ does not see it as its role to promote internal devolution of powers. The EU’s primary aim being to advance European integration, it seems almost natural that the ECJ is sceptical about internal forces of disintegration. In addition, contrary to fundamental rights protection, a federal structure is not a universal requirement for all States. The very question of whether

32 Laval, above n 3, para 92; Dynamic Medien, above n 3, paras 44–45; Commission v Germany, above n 3, para 43; Sayn-Wittgenstein, above n 3, paras 25, 82–84, 88 and 91–92. See also D Chalmers, G Davies and G Monti, European Union Law: Texts and Materials (Cambridge, Cambridge University Press, 2010) 246–48. 33 Schmidberger, above n 3, para 74; Omega, above n 3, para 35. 34 Schmidberger, above n 3, para 71; Omega, above n 3, paras 33–34; Viking, above n 3, para 44; Laval, above n 3, para 91; Dynamic Medien, above n 3, para 39; Sayn-Wittgenstein, above n 3, para 89. 35 Admittedly, Sayn-Wittgenstein may mark a turn in the attitude of the ECJ. Here, the Court stated that ‘[i]t must also be noted that, in accordance with Article 4(2) TEU, the European Union is to respect the national identities of its Member States’ (above n 3, para 92; see also para 83).

The ECJ and Member State Federalism 331 a State should adopt a federal form of government is often highly controversial and contingent. This is not even to mention issues such as which regions are to be empowered and what the boundaries of their powers should be. As a consequence, it should not surprise that the ECJ prefers to take a hands-off approach towards the Member States’ internal power patterns. In many instances, this will even be what the Member State in question wants the ECJ to do.36 However, it is one thing to be understanding of the ECJ’s hesitation actively to promote internal federalisation processes, but quite another to say that the ECJ is not required to respect existing federalism arrangements in the Member States.37 Nowadays, there are at least three compelling reasons for the Court to show respect for domestic constitutional provisions concerning federalisation and devolution. The first rationale has been touched upon already: Article 4(2) TEU, as amended by the Lisbon Treaty, expressly requires the Union to respect not only the Member States’ national identities in general, but also regional and local self-government in particular. Admittedly, this Treaty amendment was considered to be only a textual recognition of a practical reality.38 Nevertheless, it cannot be denied that the demand of respect for regional autonomy39 now has a firm legal basis such that its satisfaction should no longer be merely a matter of judicial goodwill. Respect for regional autonomy has become a legal principle to which the ECJ must have regard when deciding on the interpretation and validity of EU law. Secondly, it should be noted that an influential strand in political thought advocates that sub-State national societies are entitled to institutional accommodation within their host State, which may imply the federalisation of the State.40 This right to accommodation is said to be justified not so much because sub-State nations would have any intrinsic value but, rather, because they are considered

36 See, eg Joined Cases 51/71 to 54/71 International Fruit Company NV and Others [1971] ECR 1107, para 4; Case C-156/91 Hansa Fleisch Ernst Mundt GmbH & Co KG [1992] ECR I-5567, para 23; Case C-428/07 Horvath [2009] ECR I-6355, paras 49–51. 37 See also S Weatherill, ‘The Challenge of the Regional Dimension in the European Union’ in S Weatherill and U Bernitz (eds), The Role of Regions and Sub-National Actors in Europe (Oxford, Hart Publishing, 2005) 29. 38 See, eg, The Regional and Local Dimension in Europe, CONV 518/03, 29 January 2003, 2–3. 39 Whether the same degree of respect should be accorded to the administrative autonomy of a municipality as to the constitutionally-enshrined legislative powers of a region claiming national identity is a question falling outside the scope of this contribution. 40 See, eg, W Kymlicka, Multicultural Citizenship (Oxford, Oxford University Press, 1995) 107–30; W Kymlicka, Politics in the Vernacular (New York, Oxford University Press, 2001) 69–90; N MacCormick, ‘A Kind of Nationalism’ in N MacCormick, Questioning Sovereignty (Oxford, Oxford University Press, 1999) 173–74; Y Tamir, Liberal Nationalism (Princeton, NJ, Princeton University Press, 1993) 57–77; S Tierney, Constitutional Law and National Pluralism (Oxford, Oxford University Press, 2004) 125–29. However, a slightly different account is taken by Avishai Margalit and Joseph Raz, who regard the right to self-determination not as a substantive right to have self-government when it is right that one should have it but, rather, as a right to determine that a territory shall be self-governing (even if the case for self-government is not established). See A Margalit and J Raz, ‘National Self-Determination’ in J Raz, Ethics in the Public Domain (Oxford, Oxford University Press, 1994) 139–40.

332 Elke Cloots valuable for their individual members.41 Indeed, the nation to which a person feels connected is supposed to fulfil a major role in the formation of a person’s identity. The basic assumption is that a person is not an atomic individual but a member of various social groups, of which ‘the nation’ is one. Since communal ties shape a person’s identity and self-esteem, social groups like nations deserve to be respected and protected.42 An important mechanism for ensuring the prosperity of a national society which belongs to a larger polity is by granting it self-governing powers. It might be fairly argued that a duty to respect those self-governing powers rests not only on the host State, but also on supra-State institutions. Thirdly, it is important to realise that devolution of power is often a necessary step to ensure the peaceful coexistence of a Member State’s constituent national societies. Indeed, leaving sub-State national groups devoid of constitutionallyprotected autonomy will likely not yield a more homogeneous political community at the State level but, rather, provoke much less desired responses such as secession or terrorism.43 Seen from this perspective, European integration and internal devolution may be regarded as two different means to the same end, that is, the peaceful coexistence of different nations in a larger polity.44 Although these methods may come into tension with each other, it is fair to say that both are inextricably linked to the peace we currently enjoy in Europe. It is therefore crucial that the Court interpret and apply EU law in a manner which neither precludes a Member State from adopting a (quasi-)federal structure nor endangers European integration. Importantly, our argument for a duty on the part of the ECJ to respect not only constitutional rights, but also a Member State’s governance structure does not necessarily imply that ‘respect’ should take the same form in both types of situations. There may be more proper ways of respecting a State’s structure than allowing internal power division arrangements as a ground of justification for free movement restrictions. This point will be further developed in section III. What matters at this juncture, however, is that the ECJ has normative reasons to respect not only fundamental rights as enshrined in the Member States’ constitutions, but also a Member State’s (quasi-)federal structure, especially when multinational Member States are concerned.

41 Kymlicka, Multicultural Citizenship, above n 40, 75–106; W Kymlicka, ‘Territorial Boundaries. A Liberal Egalitarian Perspective’ in D Miller and SH Hashmi (eds), Boundaries and Justice (Princeton, NJ, Princeton University Press, 2001) 260; Margalit and Raz, above n 40, 135; Tamir, above n 40, 83–86 and 96–99; Tierney, Constitutional Law and National Pluralism, above n 40, 48–51. 42 MM Feeley and E Rubin, Federalism. Political Identity and Tragic Compromise (Ann Arbor, University of Michigan Press, 2008) 7–12; Kymlicka, ‘Territorial Boundaries’, above n 41, 261; MacCormick, ‘A Kind of Nationalism’, above n 40, 174–82 and 186–87; Margalit and Raz, ‘National Self-Determination’, above n 40, 131–36; Tamir, above n 40, 73–74. 43 See, eg, Feeley and Rubin, above n 42, 45. For an example of this radicalising effect, see C Colino, ‘Constitutional Change without Constitutional Reform: Spanish Federalism and the Revision of Catalonia’s Statute of Autonomy’ (2009) 39 Publius 262, 266–68 and 277. 44 See generally J McGarry, B O’Leary and R Simeon, ‘Integration or Accommodation? The Enduring Debate in Conflict Regulation’ in S Choudhry (ed), Constitutional Design for Divided Societies (New York, Oxford University Press, 2008) 41–88.

The ECJ and Member State Federalism 333 iii. Intrinsic Qualitative Differences A third possible rationale behind the Court’s comparatively protective attitude towards constitutional rights might be that the ECJ regards individual rights as intrinsically more valuable and, therefore, more worthy of respect than structural arrangements. In order to be able to judge this argument on its merits, it is instructive to take a look at the American scholarly debate. Contrary to what is the case in (quasi-)federal European countries, judicial review of federalism issues is surrounded by great controversy in the United States.45 One of the arguments advanced by the critics of judicial review is that federalism is of far less importance than constitutional rights, such that the US Supreme Court should not waste its institutional capital on it.46 The reason why federalism is considered less critical than rights is well explained by Jesse Choper. According to Choper, federalism determines nothing more than the level of government which has power to act. In contrast, an individual right implies that a given action may not be undertaken by any governmental body, whether it be at the central or the regional level. Whereas federalism turns merely on ‘the relative competence of different levels of government to deal with societal problems’, that is, on issues of practicability, individual rights are matters of principle which protect the dignity of the individual.47 Other strident opponents of federalism in the United States are Malcolm Feeley and Edward Rubin. Their key point is that the benefits generally attributed to US federalism, such as increased opportunities for political participation and policy experimentation, may equally be gained through mere decentralisation.48 In Feeley and Rubin’s opinion, federalism brings no added value, at least not in States with a single national identity, like the US. Although these authors do acknowledge that federalism may remedy political conflict in States where a common national identity is lacking, they stress that federalism nonetheless remains, from the participants’ perspective, a sub-optimal ‘tragic compromise’ rather than an optimal strategy.49 However, at the other end of the spectrum, it is argued that federalism is as important as, if not more important than, individual rights. Robert Nagel, for instance, stresses that, in the Framers’ eyes, the principle of federalism rather than the Bill of Rights was the guarantor of individual liberty.50 The division of powers

45 See, eg, United States v Lopez, 514 US 549, 575 (1995) (Kennedy J and O’Connor J concurring): ‘Of the various structural elements in the Constitution, separation of powers, checks and balances, judicial review, and federalism, only concerning the last does there seem to be much uncertainty respecting the existence, and the content, of standards that allow the Judiciary to play a significant role in maintaining the design contemplated by the Framers.’ 46 See, eg JH Choper, ‘The Scope of National Power vis-à-vis the States: The Dispensability of Judicial Review’ (1977) 86 Yale Law Journal 1552, 1554, 1577 and 1579–80. 47 Ibid 1154–56. 48 Feeley and Rubin, above n 42, 20–29. 49 Ibid 38–68 and 150–53. 50 RF Nagel, Constitutional Cultures. The Mentality and Consequences of Judicial Review (Berkeley, University of California Press, 1989) 65–66. See also New York v United States, 505 US 144, 181 (1992); United States v Lopez, 514 US 549, 576–77 (1995) (Kennedy J and O’Connor J concurring).

334 Elke Cloots among state and national governments was considered the ultimate safeguard against tyranny. Another proponent of judicial protection of the principle of federalism is Steven Calabresi, who puts it as follows: There is nothing in the US Constitution that is more important or that has done more to promote peace, prosperity, and freedom than the federal structure of that great document. There is nothing in the US Constitution that should absorb more completely the attention of the US Supreme Court.51

Obviously, Calabresi does not deny the importance of constitutional bills of rights. Yet he is convinced that individual rights alone cannot face the challenges posed by national diversity, challenges which may carry the germs of secession, violence, civil war and even genocide. In his view, the creation of federations, at the State as well as the supra-State level, offers the most promising avenue to peace among nations.52 As Calabresi convincingly demonstrates, federalism is a most ingenious invention which reduces the risk of majority tyranny in a socially heterogeneous State, either by incorporating the dominant national group into a larger, supra-State federation or by devolving powers to sub-State geographical entities.53 Without taking a stance on the question of whether the federal structure is still the proper one for the United States of America, it is clear that even those scholars who answer this question in the negative do not deny the value of federalism for Europe. On Calabresi’s as well as Feeley and Rubin’s account, federalism is a useful response to political conflict in various EU Member States such as Belgium, Spain and the United Kingdom. This being said, is the (quasi-)federal form of government characterising these polities nonetheless less important than constitutional rights protection? Obviously, arguments could be proffered in support of this thesis. Even if we do not regard federalism merely as a matter of administrative organisation but also, and primarily, as a device for granting self-governing powers, and hence protection, to distinct national identities within a single polity, it remains fair to say that federalism takes the form of a State-contingent and pragmatic political compromise.54 Thus, in contrast to the constitutional entrenchment of fundamental rights, the concrete implementation of federalism is, to a great extent, a matter of politics rather than principle.55 51 SG Calabresi, ‘“A Government of Limited and Enumerated Powers”: In Defense of United States v Lopez’ (1995) 94 Michigan Law Review 752, 770. See also ibid 754 and 756–90. 52 Ibid 762 and 789–90. 53 Ibid 761–70. 54 See, eg, VC Jackson, ‘Federalism and the Uses and Limits of Law: Printz and Principle’ (1998) 111 Harvard Law Review 2180, 2228; VC Jackson, ‘Comparative Constitutional Federalism and Transnational Judicial Discourse’ (2004) 2 International Journal of Constitutional Law 91, 95–96 and 105–08; JF Manning, ‘Federalism and the Generality Problem in Constitutional Interpretation’ (2009) 122 Harvard Law Review 2003, 2040–61 and 2068–69; RH Pildes, ‘Ethnic Identity and Democratic Institutions: A Dynamic Perspective’ in S Choudhry (ed), Constitutional Design for Divided Societies (New York, Oxford University Press, 2008) 200; Tamir, above n 40, 75. 55 See, eg, VC Jackson, ‘Narratives of Federalism: Of Continuities and Comparative Constitutional Experience’ (2001) 51 Duke Law Journal 223, 273–74. For an analogous reasoning, involving language

The ECJ and Member State Federalism 335 Yet it is not straightforward that federalism’s inextricable link with political compromise makes structural arrangements less imperative than constitutional rights. Indeed, a few cautionary remarks should be made. First, the fact that a Member State’s federal structure results from a political compromise does not, per se, preclude its being fundamental in nature. Leslie Green, for instance, rebuts that presumption, holding that ‘a legal right is a fundamental right if and only if it is at least partly justified by the fact that it protects a moral right’.56 According to Green, a right cannot lose its fundamental character solely because it also originates from other, less noble motives.57 Per analogiam, it may be contended that a Member State’s (quasi-)federal structure is fundamental if at least one of its justifications rests on moral grounds. Several distinguished political theorists regard this condition to be satisfied, at least in multinational federal States. According to Will Kymlicka, for example, fairness requires that the good of cultural membership be equally protected for the members of all national groups. In order to avoid the members of minority nations being systematically disadvantaged by those belonging to the majority nation, they should be protected through language rights and territorial selfgovernment.58 In Kymlicka’s view, human rights protection alone is not sufficient for ensuring ethnocultural justice, and may even aggravate injustice to minorities.59 In a similar though not identical way, Helder De Schutter presents federalism in a multinational State as an expression of fairness. De Schutter starts from the observations that, in multinational States, people often belong simultaneously to federal and sub-State national identity contexts, and that groups tend to be internally divided as to the context which is the most important one.60 De Schutter’s federalism-as-fairness thesis then implies, first, that fairness requires equal political recognition of national identities and, secondly, that federalism is the only governance structure which gives equal recognition to federal and sub-State national identities. Unitarism and secession, by contrast, neglect the

rights instead of federalism, see Canadian Supreme Court, La Société des Acadiens du NouveauBrunswick Inc et al v Association of Parents for Fairness in Education et al [1986] 1 SCR 549, paras 23–25: ‘Unlike language rights which are based on political compromise, legal rights tend to be seminal in nature because they are rooted in principle. … Language rights, on the other hand, although some of them have been enlarged and incorporated into the Charter, remain nonetheless founded on political compromise. This essential difference between the two types of rights dictates a distinct judicial approach with respect to each. More particularly, the Courts should pause before they decide to act as instruments of change with respect to language rights. This is not to say that language rights provisions are cast in stone and should remain immune altogether from judicial interpretation. But, in my opinion, the Courts should approach them with more restraint than they would in construing legal rights.’ 56 57 58 59 60

L Green, ‘Are Language Rights Fundamental?’ (1987) 25 Osgoode Hall Law Journal 639, 647. Ibid 645–46. Kymlicka, Multicultural Citizenship, above n 40, 113. Kymlicka, Politics in the Vernacular, above n 40, ch 4. H De Schutter, ‘Federalism as Fairness’ (2010) 18 Journal of Political Philosophy 1, 2 and 7–9.

336 Elke Cloots identities of those individuals who belong primarily to either the sub-State or the federal nation, respectively.61 Secondly, it should be noted that individual rights, as enshrined in a constitution, may also be the result of a constitutional accord, which often embodies Statespecific value judgments.62 Thus, it is not uncommon for the EU Member States to adhere to different conceptions of a human right, depending on the country’s specific history and culture. This was very well illustrated in, for example, Omega and Sayn-Wittgenstein, where the German understanding of the right to respect for human dignity and the Austrian implementation of the principle of equality, respectively, were at issue. Yet the fact that other Member States might construe the rights involved in a different way was no reason for the ECJ to deny the fundamental nature of the German and Austrian constitutional rights.63 Thirdly, even though it is true that federalism is given concrete form by a political compromise and that it is hard to determine abstract criteria for how best to structure a federal system, this by no means implies that everything is permissible. On the contrary, whether federalism can actually succeed in protecting distinct national identities is dependent on objective factors such as the territorial limits of a State’s subdivisions and the nature of the competences accorded to them.64 It is, therefore, to be welcomed that several interesting attempts have recently been undertaken to develop more general rules for the design of multinational federal States.65 Lastly, and independently of whether constitutional rights are more important than federalism from a normative point of view, (quasi-)federal Member States do not always seem to consider rights more fundamental than structure. In Belgium, for instance, the Court of Arbitration (now the Constitutional Court) was established as a special judicial body with jurisdiction to monitor acts of the federal and sub-State parliaments in the light of constitutional and semi-constitutional provisions concerning the Belgian State structure. However, no judicial review— neither by the Court of Arbitration nor by any other court—was available to challenge the compatibility of parliamentary acts with the fundamental rights enshrined in the Belgian Constitution. Admittedly, in the course of time, the Court of Arbitration’s jurisdiction was extended such that the Court now also has the authority to supervise the legislatures’ observance of constitutional rights.66 Nonetheless, the Court still has the habit of adjudicating first on federalism issues,

61

Ibid 2–3 and 13–17. See also Helder De Schutter’s contribution in ch 7 of this volume. See also Green, above n 56, 645–46; Manning, above n 54, 2042–43. 63 Omega, above n 3, paras 30, 32 and 37; Sayn-Wittgenstein, above n 3, paras 86, 91 and 93. 64 Kymlicka, Politics in the Vernacular, above n 40, 101–12; Kymlicka, ‘Territorial Boundaries’, above n 41, 267. See also N Stojanovic´, ‘Mononational and Multinational States. A Valid Dichotomy?’, EUI Working Paper, available at , 15–16. 65 See, eg, De Schutter, above n 60, 17–20; Kymlicka, Politics in the Vernacular, above n 40, chs 5 and 11. 66 See Belgian Constitution, Art 142; Special Law on the Constitutional Court (1989), Arts 1 and 26. 62

The ECJ and Member State Federalism 337 proceeding to the rights-based arguments only if the act has successfully passed the federalism review.67 Before concluding this section on the comparative value of constitutional rights and federalism, a final point should be made. If the ECJ were to consider Member State federalism insufficiently important to be included in the list of grounds of justification, the greatest fallacy in the Court’s reasoning would not be that governance structure is ranked lower than constitutional rights. Indeed, a far more severe error would come from the fact that a Member State’s federal structure is given a lower priority than public interests such as the protection of public health, consumer protection, road safety and environmental protection. Contrary to federalism, policies of this type do not serve to ensure a person’s (individual or collective) moral rights but, rather, aggregations of individual interests.68 Nonetheless, the ECJ regards them as legitimate reasons for restricting fundamental freedoms, while denying that status to a Member State’s (quasi-)federal structure.69 iv. Alternative (Not Less) Protection So far, we have examined three grounds upon which the ECJ might have based its preferential treatment of constitutional rights as opposed to State structure provisions: the Charter; the Union’s duty to respect fundamental rights; and the idea that fundamental rights are, by their very nature, more important than federalism. While not denying that these factors may explain the ECJ case law, we have tried to demonstrate that their justificatory force is not uncontroversial. However, before we are able to conclude with confidence that the ECJ treats Member State federalism less favourably than constitutional rights without giving a fully satisfactory justification, a final possibility must be considered. The question arises as to whether the acceptance of constitutional rights and the dismissal of a Member State’s (quasi-)federal structure as a ground of justification should necessarily be presented in terms of distinct levels of respect. More precisely, might it be that both types of constitutional traditions are actually being respected, albeit in different manners, which cannot adequately be described as ‘more’ or ‘less’ respectful?70 To test whether this hypothesis is true, we need to delve deeper into the Flemish Care Insurance Scheme Case. The Court’s rejection of the Flemish government’s

67 See, eg, Belgian Constitutional Court, No 29/96, 15 May 1996, paras B.2 and B.9; No 42/97, 14 July 1997, para B.7.2; No 151/2007, 12 December 2007, para B.11; No 187/2008, 18 December 2008, para B.2. 68 Green, above n 56, 648–49. 69 See, eg, Case 96/85 Commission v France [1986] ECR 1475, para 10; Case 27/80 Fietje [1980] ECR 3839, paras 10–11; Case C-246/00 Commission v Netherlands [2003] ECR I-7485, para 67; Case C-302/86, Commission v Denmark [1988] ECR 4607, para 9. 70 For a comparable argument, see F Schauer, ‘Categories and the First Amendment: A Play in Three Acts’ (1981) 34 Vanderbilt Law Review 265, 287.

338 Elke Cloots defence based on the Belgian federal structure is, in effect, only one part of the picture. Equally noteworthy is the preceding piece of the judgment, in which the scope of application of the Treaty provisions on freedom of movement was defined. In essence, the Court excluded situations which are ‘purely internal’ to the Belgian State from the scope of EU law.71 In that type of situation, the Belgian federal structure applied in full, not being subject to any requirements of EU law. The other side of the same coin, however, was that, in situations which were not classified as ‘purely internal’, the Belgian State structure could not alleviate the Member State’s obligations under EU law. Here, EU free movement law applied with full force. Hence, ‘the requirements inherent in the division of powers within the Belgian federal structure’ were not capable of justifying the free movement obstacle at issue.72 Taking this broader view of the Court’s decision in the Flemish Care Insurance Scheme Case, we can see that the Belgian State structure was not left without a remedy, even though no balancing approach was taken. But if it is not balancing, what is it that the ECJ is doing? Once again, US scholarly literature is revealing. American scholars tend to contrast ‘balancing’ with ‘categorisation’. In what follows, it will be argued that the latter label is accurate for the ECJ’s mode of adjudication in cases involving a restriction on the freedom of movement flowing from a Member State’s structural arrangements. Before coming to that argument, however, it is useful to clarify the distinction between balancing and categorisation as it is commonly presented in US legal scholarship. The next section will therefore start with a brief sketch of the origins of the balancing/categorisation divide in American constitutional theory. Subsequently, our attention will shift to the divide’s manifestations in the ECJ case law.

III. BALANCING VERSUS CATEGORISATION

A. In American Constitutional Theory The search for the most appropriate mode of rights adjudication has occupied generations of legal scholars in the US. The debate mainly revolves around two competing methods for reviewing rights limitations, that is, balancing and categorisation. The choice between both approaches, in essence, turns on the degree of discretion a legal system wishes to leave to the judge called upon to apply a given constitutional right to a particular case.73 If a balancing approach 71

Government of the French Community and Walloon Government, above n 10, paras 36–38. Ibid paras 57–58. 73 See, eg, VC Jackson, ‘Being Proportional about Proportionality’ (2004) 21 Constitutional Commentary 803, 836–38; S Sottiaux, Terrorism and the Limitation of Rights. The ECHR and the US Constitution (Oxford, Hart Publishing, 2008) 37–40; A Stone, ‘The Limits of Constitutional Text and Structure: Standards of Review and the Freedom of Political Communication’ (1999) 23 Melbourne University Law Review 668, 686–87. 72

The ECJ and Member State Federalism 339 is preferred, the judge is invited to consider all interests implicated in the case presented to him. The judge puts the interests which he thinks relevant on a common scale, and weighs them against one other.74 By contrast, when using a categorical approach, the judge is much more constrained in his ability to ponder the relevant factors in an individual case. The only task which is conferred upon him is to verify whether the facts of the case may be classified within a certain legal category or concept. Once this definitional question has been decided, the outcome follows automatically. The essence of the distinction between a balancing and a categorical approach was captured succinctly by Kathleen Sullivan: ‘For the categorizer, legal questions turn on differences in kind; for the balancer, they are matters of degree.’75 The balancing/categorisation dichotomy is often mentioned in connection with a separate but related distinction running through American constitutional theory, ie the standards/rules divide. Both standards and rules are forms of legal directives designed to bridge the gap between substantive ‘background justifications’ (eg, fairness, democracy, etc) and a factual situation.76 Under perfect conditions, these legal directives help judges resolve concrete cases in accordance with the background justifications. In parallel with the balancing/categorisation debate, the difference between standards and rules is said to lie in the degree of discretion which is left to the judge. The vaguer and, thus, the more standardlike a legal directive is, the more discretion is given to the judge.77 In the extreme case, this means that judicial decision-making is ‘collapse[d] … back into the direct application of the background principle or policy to a fact situation’.78 Adjudication based on bright-line rules, by contrast, implies that certain specified facts trigger a determinate judicial response. It is neither necessary nor possible for the judge to consider the background justifications underpinning the rule. On the contrary, the rule must be applied even if that means that the outcome mandated by the direct application of the background justification is not obtained.79 The balancing/standards approach has at least one significant merit in comparison with the categorisation/rules method. Conferring a considerable degree 74 See, eg, TA Aleinikoff, ‘Constitutional Law in the Age of Balancing’ (1987) 96 Yale Law Journal 943, 945; Stone, above n 73, 686–87; KM Sullivan, ‘Post-Liberal Judging: The Roles of Categorization and Balancing’ (1992) 63 University of Colorado Law Review 293, 293–94. 75 KM Sullivan, ‘Foreword: The Justices of Rules and Standards’ (1991) 106 Harvard Law Review 22, 59. For a trenchant critique, see PJ Schlag, ‘An Attack on Categorical Approaches to Freedom of Speech’ (1983) 30 UCLA Law Review 671, 733–34. 76 Sullivan, above n 75, 57. As regards rules, see also F Schauer, Playing by the Rules: A Philosophical Examination of Rule-Based Decision-Making in Law and in Life (Oxford, Oxford University Press, 1991) 104. 77 Jackson, above n 73, 838; Sullivan, above n 75, 58–59. It should be noted, however, that some scholars contest the commonly-held view that there is an inextricable link between a legal directive’s specificity and the degree of discretion the directive confers upon the decision-maker. See especially F Schauer, ‘The Convergence of Rules and Standards’ (2003) New Zealand Law Review 303. 78 Sullivan, above n 75, 58. See also Stone, above n 73, 689. 79 See, eg, Schauer, above n 76, 100–02 and 135; Stone, above n 73, 688–90; Sullivan, above n 75, 58.

340 Elke Cloots of judicial discretion, it bestows upon the judge the necessary flexibility to take full account of the particular circumstances of each individual case.80 Rules, by contrast, may hinder such a particularised decision-making by screening off certain relevant factors from judicial consideration. Since it seems unreasonable and unfair not to look at all relevant particularities of a case, what reason could there be for favouring rules over standards, categorisation over balancing?81 Why should we risk opening the door for sub-optimal judicial decisions? Although a whole range of arguments has been proffered in defence of rules and categorical rights adjudication,82 we shall confine ourselves here to what is, in our view, the weightiest reason in the particular context of the ECJ case law concerning Member State federalism. Frederick Schauer explained that reason as follows: ‘When decision-makers not shackled by rules are thus empowered to inquire into every factor that could lead to the best decision for a particular case, they simply might not make that best decision.’83 Indeed, a judge who is neither guided nor constrained by rules may run into several traps. Lack of knowledge and understanding of certain factors, as well as personal and/or structural bias, may lead decision-makers to misapply background principles.84 Standards, a proportionality test for instance, do not always suffice to neutralise potential judicial misunderstanding and prejudice. Sometimes, stricter rules, such as bright-line definitions of legal concepts, are needed to ensure that due consideration is given to a factor which the court would otherwise either over- or under-value. Put differently, if there is a significant risk that the court will not value a certain interest as it should, be it for cognitive or ideological reasons, there is a good case for rule-based decision-making and a categorical approach.85

B. In the ECJ Case Law The abundance of literature on categorisation and balancing, and rules and standards in US legal scholarship stands in sharp contrast to the complete absence of this debate in EU legal doctrine.86 How are we to account for this disparity? Could 80

Aleinikoff, above n 74, 961; Stone, above n 73, 689–90. See, eg, Aleinikoff, above n 74, 962, 992 and 1001–02; Schauer, above n 76, 135–37; Schlag, above n 75, 734–35. 82 The most comprehensive account may be found in Schauer, above n 76, 135–66. 83 Schauer, above n 76, 149–50. For an analogous argument, see L Alexander and E Sherwin, The Rule of Rules (Durham, NC, Duke University Press, 2001) 55; MB Nimmer, ‘The Right to Speak from Times to Time: First Amendment Theory Applied to Libel and Misapplied to Privacy’ (1968) 56 California Law Review 935, 945–47. 84 Schauer, above n 76, 150–51. 85 Ibid 152–55. See, in a similar vein, JH Ely, ‘Flag Desecration: A Case Study in the Roles of Categorization and Balancing in First Amendment Analysis’ (1975) 88 Harvard Law Review 1482, 1500–01; Jackson, above n 73, 856–57; M Tushnet, ‘Defending Korematsu?: Reflections on Civil Liberties in Wartime’ (2003) Wisconsin Law Review 273, 281–82. 86 However, the case law of the European Court of Human Rights has occasionally been studied in the light of the rules/standards debate. See, eg, S Sottiaux and G van der Schyff, ‘Methods of 81

The ECJ and Member State Federalism 341 it be that there is simply no need to repeat the American discussion here, because the European Treaties—unlike the US Constitution87—have already decided the issue for the judiciary and academia? Theoretically, it is, of course, possible that the European Treaties prescribe in sufficient detail what method of adjudication is to be employed in a particular type of situation. As far as freedom of movement is concerned, it must, moreover, be admitted that several Treaty provisions actually do contain limitation clauses.88 Nevertheless, it leaves no doubt that the definition and limitation of freedom of movement (and other core European principles) are, to a great extent, matters of judge-made law.89 Moreover, as was demonstrated above and will be further explored in what follows, the balancing/categorisation divide is equally present in the ECJ case law. Let us now take a closer look at the distinction drawn by the Court. As noted in section II., balancing rhetoric is ubiquitous in cases where the protection or exercise of a constitutional right entails a restriction on freedom of movement. Moreover, the ECJ has recently expressly rejected an alternative, categorical approach to this type of case. In Viking, Laval and Commission v Germany, the argument was advanced that the exercise of the right to take collective action and the right to bargain collectively, respectively, falls outside the scope of the Treaty provisions on freedom of movement. Yet this suggestion was readily dismissed by the Court.90 The Court decided that the exercise of fundamental rights is not, in and of itself, excluded from the scope of EU law. Instead, the Court reaffirmed its traditional balancing approach: the exercise of a fundamental right must be reconciled with the free movement rights guaranteed by the Treaty in accordance with the principle of proportionality.91 Importantly, we do not mean to imply here that a

International Human Rights Adjudication: Towards a More Structured Decision-Making Process for the European Court of Human Rights’ (2008) 31 Hastings International & Comparative Law Review 115. See also J Gerards and H Senden, ‘The Structure of Fundamental Rights and the European Court of Human Rights’ (2009) 7 International Journal of Constitutional Law 619. 87 As Stefan Sottiaux amply demonstrates, the US Constitution formulates rights in absolute terms and contains neither general nor rights-specific clauses concerning their limitation. See Sottiaux, above n 73, 56–57. 88 See especially Arts 36, 45, 52 and 65 TFEU. 89 See, amongst many other landmark judgments, Case 8/74 Dassonville [1974] ECR 837; Case 120/78 Rewe-Zentral AG [1979] ECR 649; Joined Cases C-267/91 and C-268/91 Keck and Mithouard [1993] ECR I-6097; Case C-67/96 Albany International [1999] ECR I-5751; Schmidberger, above n 3; Case C-88/03 Portugal v Commission [2006] ECR I-7115; Government of the French Community and Walloon Government, above n 10. 90 Viking, above n 3, paras 37, 46–47 and 55; Laval, above n 3, paras 94–95; Commission v Germany, above n 3, paras 41–42 and 50. 91 Viking, above n 3, para 46; Laval, above n 3, para 94; Commission v Germany, above n 3, para 44. One may of course question the equation of proportionality review with a balancing exercise. Much has been said about that and we do not intend to re-open the debate here. It suffices to say that, on a continuum, the proportionality test applied by the ECJ to constitutional rights cases is more ‘standardlike’ or less ‘rule-like’ than the Court’s approach to federalism cases. We suppose that even those who deny that a proportionality inquiry necessarily involves a balancing of interests could agree with that. Moreover, the ECJ itself is not hesitant to depict its proportionality assessment as a balancing act (see nn 7–9 above).

342 Elke Cloots conflict between a fundamental freedom and a constitutional right calls for balancing as a matter of logical necessity. We merely note the empirical fact that the Court tends to engage in a balancing act when confronted with this type of dispute.92 A rather different mode of adjudication may be observed in decisions on free movement restrictions caused by a Member State’s (quasi-)federal structure. Here, the Court distinguishes between two ‘kinds’ or ‘categories’ of situations:93 purely internal situations, and situations which have a link with EU law. To the extent that the regional measure affects situations which are ‘purely internal’ to the Member State, it is not subject to EU law.94 Hence, within that category, the Member State’s governance structure can flourish in full. However, in so far as the measure impinges on situations which do have a link with EU law, EU law applies with full force. This means that if the measure constitutes a hindrance to free movement, it cannot be justified by relying on the federal structure of the Member State.95 It may be clear, by now, that the ECJ’s analysis in the Flemish Care Insurance Scheme Case constitutes a paradigm example of a categorical, rule-based style of decision-making. The Court carved out two categories of situations and considered to what extent the facts of the case were caught by categories x and y, respectively. Once this classification exercise had been completed, the Court proceeded to undertake a balancing act in neither of the two categories.96 Instead, it allotted the State’s federal structure to the first category, and EU free movement law to the second. One may of course object that the exemption of purely internal situations from the scope of EU free movement law is not the exclusive province of cases involving Member State federalism. Intra-State situations are not covered by the 92 There are some exceptions, though. For instance in Grogan, the ECJ did apply a categorical approach to a clash between a constitutional right (ie, the right to life of the unborn as enshrined in the Irish Constitution) and freedom of movement (ie, the freedom to provide services). At issue was an Irish legal prohibition on the distribution of information on abortions carried out abroad. Instead of verifying whether the Irish measure, designed to protect the right to life of the unborn, constituted a proportionate restriction of the freedom to provide services, the Court ruled that the measure did not amount to a restriction in the first place (Case C-159/90 Grogan [1991] ECR I-4685, paras 24, 27 and 29). The Advocate General, by contrast, did regard the Irish measure as falling within the scope of the freedom to provide services and subjected it to proportionality review (Opinion of AG Van Gerven, paras 21 and 24–29). Another exception to the balancing approach may occur when the constitutional right at stake is considered ‘absolute’ in nature (eg, the right not to be tortured; see Schmidberger, above n 3, para 80). 93 Government of the French Community and Walloon Government, above n 10, para 36. See also para 42: ‘So far as that second category of worker is concerned …’ (emphasis added) 94 Ibid paras 38–39. 95 Ibid paras 41–42, 54 and 56–58. 96 True, as regards the category of inter-State situations, the Court did embark on the balancing path (para 55). However, the Court retraced its steps as soon as it realised that only the Belgian federal State structure was invoked as a ground of justification (paras 56–58). Instead of reconciling State structure and free movement, the Court screened out federalism issues as irrelevant considerations. Obviously, if ordinary policy interests rather than federalism had been relied on, the Court would have balanced the competing interests, but that is not what concerns us here (see, eg Case C-169/08 Presidente del Consiglio dei Ministri v Regione Sardegna [2009] ECR I-10821, paras 40–50). The point we wish to make is that the ECJ does not actually balance Member State federalism against freedom of movement.

The ECJ and Member State Federalism 343 Treaty provisions on freedom of movement, whatever the interest the defendant invokes, including constitutional rights protection. This is a totally fair point. However, it does not refute our basic argument that the Court, though unwilling to weigh and balance fundamental freedoms against State structure-related considerations on a case-by-case basis, nonetheless accommodates the latter, and that the Court’s preferred tool for doing so is to delineate the scope of EU law. Contrary to what is usual in constitutional rights cases, the ‘hard work’ is done at the outset of the inquiry, at the stage where legal concepts must be defined. It will be submitted below that the ECJ has good reasons to favour that approach over balancing in federalism cases. However, before proceeding to the next section, it may be instructive to examine some other judgments involving the (quasi-)federal structure of Member States. It will become clear immediately that a categorical approach is not only employed to decide on regional limitations of freedom of movement, but also to adjudicate State structure-related challenges to the Treaty prohibitions on State aid and discrimination. The landmark case as regards Member State federalism and State aid is Portugal v Commission. There, the Court had to decide whether a tax benefit accorded by a region in the exercise of its constitutionally-entrenched fiscal autonomy constituted, for the sole reason of its limited geographical scope, a ‘selective’ advantage. Since Article 87(1) TEC (now Article 107(1) TFEU) prohibits State aid which is selective in nature,97 the finding that regional tax reductions were necessarily selective, even if adopted by an autonomous region, would entail that devolution of tax powers opened the door to prohibited State aid. In order to avoid this unfortunate consequence of the prohibition of State aid, the Court modified its interpretation of the concept of ‘selectivity’ and, hence, ‘State aid’. More precisely, the Court held that a measure could not be regarded as selective ‘on the sole ground that it is applicable only in a limited geographical area of a Member State’.98 If a regional authority which was ‘sufficiently autonomous’ in relation to the Member State’s central government adopted a tax rate lower than the national rate to the benefit of undertakings present in the region’s territory, the measure’s limited geographical scope could not, in and of itself, render the benefit selective.99 In order to determine whether the regional authority adopted the fiscal measure in a ‘sufficiently autonomous’ way, the Court laid down three essential conditions. First, the regional authority must, from a constitutional point of view, possess a political and administrative status distinct from that of the central government

97 More precisely, Art 87(1) TEC (now Art 107(1) TFEU) is targeted at ‘any aid granted by a Member State or through State resources … favouring certain undertakings or the production of certain goods’ (emphasis added). 98 Portugal v Commission, above n 89, paras 57 and 60. See also UGT-Rioja and Others, above n 25, para 47. 99 Portugal v Commission, above n 89, paras 58 and 62–66. See also UGT-Rioja and Others, above n 25, paras 48–50. Obviously, it is possible that the regional fiscal benefit is nonetheless selective because it does not apply to all undertakings established in or all production of goods on the region’s territory.

344 Elke Cloots (institutional autonomy). Secondly, the central government must not be able to intervene directly as regards the content of the tax measure (procedural autonomy). Lastly, the financial consequences of the regional tax reduction for undertakings present in the region must not be compensated for by aid or subsidies from other regions or the central government (economic and financial autonomy).100 Establishing this threefold autonomy test, the Court designed a bright-line rule redrawing the boundaries of the legal concept of State aid. The outcome of a particular case will depend on whether or not the test is met on the facts of the case, not on whether, all things considered, the regional fiscal benefit is justified.101 Thus, once again, the Court left room for devolution of powers in the Member States by adopting a categorical, rulelike mode of adjudication rather than a balancing, standard-like approach. The ECJ’s analysis in Horvath confirms this trend. In this case, the ECJ was confronted with a claim that the territorially-differentiated implementation of a EU Regulation in the UK, resulting from the devolution of powers in that State, ran counter to the EU law prohibition of discrimination.102 The Court dismissed the claim, stating that where it is the devolved administrations of a Member State which have the power to implement a Regulation, ‘divergences between the measures provided for by the various administrations cannot, alone, constitute discrimination’.103 That conclusion was drawn from an analogy with the case law on differences in implementation in different Member States. The Court recalled that the prohibition on discrimination is, in principle, ‘not concerned with’ any disparities in treatment which may result from divergences existing between the legislation of the various Member States.104 The same reasoning was applied, mutatis mutandis, to devolved administrations constitutionally invested with legislative powers.105 In other words, differences in implementation in different autonomous polities, be it Member States or sub-State entities, are not captured by the definition of discrimination. It may thus be said that, in this type of case too, a categorical approach saved the (quasi-)federal structure of a Member State from the integration logic of EU law. The Court seemed to prefer that mode of adjudication over the balancing exercise suggested by some of the parties to the case.106 Instead of inquiring into whether the differential implementation of the Regulation was ‘objectively justified’ by the British system of power division,

100

Portugal v Commission, above n 89, para 67; UGT-Rioja and Others, above n 25, para 51. See especially UGT-Rioja and Others, above n 25, para 141. 102 Horvath, above n 36, paras 18 and 47. For a more detailed analysis of this case, see J Hunt, ‘Devolution and Differentiation: Regional Variation in EU Law’ (2010) 30 Legal Studies 421, 434–36. 103 Horvath, above n 36, para 57. 104 Ibid para 55. Cp, eg, Case 14/68 Wilhelm and Others [1969] ECR 1, para 13; Case C-137/00 Milk Marque and National Farmers’ Union [2003] ECR I-7975, para 124; Case C-403/03 Schempp [2005] ECR I-6421, para 34. 105 Horvath, above n 36, paras 57–58. 106 See Horvath, above n 36, Opinion of AG Trstenjak, paras 30 and 35. 101

The ECJ and Member State Federalism 345 the Court simply chose to keep this sort of divergence out of the EU concept of discrimination.107

IV. A CATEGORICAL APPROACH TO MEMBER STATE FEDERALISM: MERITS AND PITFALLS

We may infer from this case study that the ECJ tends to pursue a categorical and rule-based strategy for showing sensitivity to a Member State’s governance structure. But it is one thing to describe the Court’s case law; it is another to assess judicial decisions from a normative point of view. In this section, the merits and drawbacks of the Court’s categorical approach as regards Member State federalism will be discussed.

A. Merits i. Minimising Decision-maker Errors As noted above, the difference between balancing and categorisation, and between standards and rules, in essence turns on the extent of the discretion conferred upon the decision-maker. As Kathleen Sullivan has amply demonstrated, actual methods of adjudication may be put on a continuum from rule-like to standard-like, from categorical to balancing, depending on the margin of discretion the court retains.108 Since we observed that the ECJ employs a categorical and rule-like style of decisionmaking in federalism cases, in contrast to a balancing and standard-like mode of adjudication in rights cases, the question arises why the Court restricts its own discretion more in the former type of case than it does in the latter. Put differently, what reason might the ECJ have to submit itself to bright-line rules when Member

107 It should be noted that the Horvath case concerned the implementation of a Regulation in the area of the Common Agricultural Policy (CAP). Art 40(2) TFEU contains a specific anti-discrimination provision with respect to the agricultural sector. As has been well documented, the ECJ’s evaluation of the justification for a different treatment of comparable situations in the agricultural sector tends to be very minimalistic. In a host of cases, the Court considers only whether the difference in treatment is ‘arbitrary’. It is not always possible clearly to distinguish this inquiry from the comparability question. Moreover, the Court does not easily find situations sufficiently comparable, such that it often does not even get to the justification stage. See, eg, R Barents, ‘The Significance of the Non-Discrimination Principle for the Common Agricultural Policy: Between Competition and Intervention’ in D Curtin and T Heukels (eds), Institutional Dynamics of European Integration, vol 2 (Dordrecht, Martinus Nijhoff Publishers, 1994) 536–37, 543 and 548; JH Gerards, Rechterlijke toetsing aan het gelijkheidsbeginsel (The Hague, Sdu Uitgevers, 2002) 238–43, 251–58, 275–78 and 289–98; T Tridimas, The General Principles of EU Law (Oxford, Oxford University Press, 2006) 78–84. The Court’s particular approach to the CAP non-discrimination principle urges us to be careful not to attach too much importance to the fact that the ECJ in Horvath did not properly test whether the difference in treatment was ‘objectively justified’. Nonetheless, we believe that this insight does not undermine our analysis of the Court’s decision, since the Horvath rule would presumably apply as well to other, non agriculture-related manifestations of the non-discrimination principle (see Horvath, above n 36, para 55). 108 Sullivan, above n 75, 61. See also Sottiaux, above n 73, 38.

346 Elke Cloots State federalism and core principles of EU law are involved? Here it is instructive to recall the advantage of rule-based decision-making which was already highlighted in the previous section: rules reduce the risk of decision-maker errors. Since there are reasons to believe that the ECJ will err on conflicts between Member State federalism and core EU principles, and that these errors may have serious consequences, a strong argument may be made in favour of rule-based decision-making in this type of case. What, then, are these reasons? 1. Imperfect Knowledge A first source of distrust in the ECJ is the Court’s presumed lack of understanding of domestic federalism issues. As noted above, a State’s governance structure is the product of a political compromise, which is, in turn, contingent on a wide variety of parameters, such as geography, history, the number and size of the national minorities concerned, the strength of the respective national identities, etc.109 It depends on these and other factors whether or not all territorial subdivisions of a State are granted an equal degree of autonomy, and how the competences of these subdivisions and the central governmental level are delineated. Against this background, one may easily understand that it is difficult for a supranational institution like the ECJ fully to grasp the particularities and sensitivities of a Member State’s (quasi-)federal system.110 This interpretation difficulty is compounded by the fact that federalism provisions tend to be embedded in a larger ‘package’ of structural arrangements, such that their interpretation is often affected by other parts of the whole.111 Moreover, as a result of their contingent and embedded character, similar constitutional provisions concerning federalism may be interpreted differently in different Member States.112 In such circumstances, it seems almost unfeasible for the ECJ to decide issues of Member State federalism correctly. 2. Bias Secondly, decision-maker errors may be caused by bias, on the part of the ECJ, against devolution of powers in the Member States and in favour of the individual right to freedom of movement. The prejudice against domestic devolution 109 See, eg Jackson, above n 55, 226, 270 and 272–74; Jackson, ‘Comparative Constitutional Federalism’, above n 54, 95–96 and 105–08; R Simeon, ‘Constitutional Design and Change in Federal Systems: Issues and Questions’ (2009) 39 Publius 241, 242, 248–49 and 257–58. See also LH Tribe, American Constitutional Law, vol 1 (New York, Foundation Press, 2000) 127–30. 110 See, in a similar vein, Sejdic´ and Finci v Bosnia and Herzegovina App no 27996/06 and no 34836/06 (ECHR, 22 December 2009), Bonello J dissenting: ‘I doubt that the [European] Court [of Human Rights] is better placed than the national authorities to assess the point in time when previous fractures consolidate, when historical resentments quell and when generational discords harmonise.’ 111 Jackson, above n 55, 273–74; Jackson, ‘Comparative Constitutional Federalism’, above n 54, 95 and 102–05. 112 See, eg, the principle of federal loyalty, which is embedded in the German and Belgian Constitutions but does not have exactly the same meaning in both constitutional orders.

The ECJ and Member State Federalism 347 processes may be both institutional and personal in nature. There are, to begin with, various reasons why the ECJ, as an institution, might be biased against, and therefore undervalue, Member State federalism. A first reason relates to the manner in which the type of case under consideration tends to appear before the ECJ. In most cases, there is, on the one hand, a claim to an individual right or freedom, for instance the right to freedom of movement or the right not to be discriminated against, and, on the other hand, a defence based on a structural provision which may be captured in the following general formula: ‘regional authority x has (or does not have) the power to do y’.113 Courts are likely to feel pressure to decide in favour of the individual in this kind of litigation setting. Robert Nagel convincingly explains why: These concepts [like federalism] are important, but their importance is based in large measure on theoretical considerations. These considerations do have real world consequences but often only in the long run and only in a diffuse or systemic way. In lawsuits, these structural considerations come into conflict with highly individualized claims of right, and at each such juncture, it is likely that the structure will seem basically secure and (in any event) rather an abstract matter, while the individual’s interest is likely to seem concrete, immediate, and in jeopardy.114

Put differently, courts are reluctant to tip the balance in favour of long-term structural interests when this directly harms an individual’s rights or freedoms. A second reason why courts tend to undervalue structural elements such as federalism and separation of powers, as opposed to individual interests, has been pointed out by Alexander Aleinikoff and Vicki Jackson. These authors identify the root of the problem in the inadequate representation of structural interests in judicial proceedings, which may be due to the fact that only individual parties are being involved.115 While it is true that Member States are always entitled to intervene in litigation before the ECJ, this does not allow us to conclude that issues of Member State federalism are well represented. In Portugal v Commission, for instance, the primary stakeholders in the case, that is, sub-State territorial entities with genuine fiscal power (eg, Scotland, Flanders and the Basque Country), were not permitted to intervene in the proceedings.116 Although it must be admitted that in the case under consideration, the Scottish and Basque interests were defended by the UK and Spanish Governments, respectively, this cannot be considered self-evident. The fact

113 See, eg, UGT-Rioja and Others, above n 25 (free competition versus the Basque authorities’ powers over certain tax matters); Government of the French Community and Walloon Government, above n 10 (freedom of movement for persons versus the Flemish Community’s lack of power—in the field of ‘aid to persons’—over persons not living in Flanders); Horvath, above n 36 (the right not to be discriminated against versus the power of the UK’s devolved administrations to implement Regulations concerning the CAP). 114 RF Nagel, ‘Liberals and Balancing’ (1992) 63 University of Colorado Law Review 319, 322–23. 115 Aleinikoff, above n 74, 978–79; Jackson, above n 73, 845. 116 According to Art 40 of the Statute of the Court of Justice of the European Union, natural or legal persons, including autonomous regions, ‘shall not intervene in cases between … Member States and institutions of the Union’.

348 Elke Cloots that the Belgian Government did not intervene to secure the defence of the Flemish interests proves this. Moreover, the interests of the central State cannot be assumed to coincide with those of sub-State nations. Hence, one may wonder whether the central government is the right institution to voice the concerns of autonomous regions regarding issues of domestic federalism. A third source of institutional bias is particularly connected with the ECJ’s capacity as a Union institution. Whereas the EU seeks to promote economic and political integration, devolution may be regarded as a form of disintegration. Seen in this light, devolution seems to run counter to the very purpose of the Union. If the ECJ is called upon to settle a conflict between domestic devolution of powers and principles appertaining to the essence of European integration, such as the internal market freedoms and the prohibition on State aid, it should not be surprising that the Court will likely sympathise more with the latter than with the former. In addition to these institutional types of bias, there may also be personal prejudice against Member State federalism in ECJ proceedings. Since there is only one ECJ judge for each Member State, the views of the national society to which a multinational Member State’s judge belongs will probably be better represented in the Court than those of other national groups. True, it may be objected that the underrepresentation of the concerns of certain autonomous national groups is alleviated if the Member State in question ensures that each territorial subdivision has, on alternating terms, a judge in the ECJ.117 That, however, will offer little consolation to autonomous regions which have a stake in a particular case but lack representation among the ECJ judges at the moment the decision is made. Moreover, it should be noted that the establishment of a rotation system is dependent on the goodwill of each Member State, and that not all (quasi-)federal States seem prepared to appoint ECJ judges in a way reflecting their multinational character.118 3. Serious Consequences The fact that the ECJ, for the reasons set out above, runs a great risk of making errors of judgement is compounded by the seriousness of these types of errors.119 Both the goals pursued by multinational federalism and the process through which devolution of powers is implemented suggest that any misunderstanding or underestimation of the significance of a multinational State’s (quasi-)federal structure could be devastating. The noble aims of multinational federalism were set out earlier in this contribution: federalism, at least in multinational Member States, is a means of accommodating distinct national identities and achieving

117 This is, for example, the case in Belgium, where there is a convention that Flemish and Frenchspeaking judges (and Advocates General) rotate. 118 See, eg, the Spanish appointment system (J Bengoetxea, ‘The Participation of Infra-State Entities in European Union Affairs in Spain: the Basque Case’ in S Weatherill and U Bernitz (eds), The Role of Regions and Sub-National Actors in Europe (Oxford, Hart Publishing, 2005) 59). 119 See Schauer, above n 76, 152.

The ECJ and Member State Federalism 349 social peace. It is fair to argue that the ECJ should not risk thwarting the pursuit of these goals by interpreting domestic federalism issues erroneously. As to the process, it should be borne in mind that the compromises giving concrete shape to devolution of powers in a multinational State often emerge from highly legitimate political processes. Thus, federal arrangements invariably have a (semi-)constitutional status. Although the details of the amendment procedures differ from Member State to Member State, federalism arrangements must generally be approved by multiple political stakeholders. In addition, there is often a constitutional requirement of a special majority vote and/or a referendum.120 This firm democratic basis underlying Member State federalism is further supplemented by the provision for review by the respective constitutional courts.121 In sum, rigorous democratic and judicial checks built into the constitutional orders of multinational Member States lend particular legitimacy to these States’ governance structures. The ECJ should, therefore, be wary of disturbing the (quasi-) federal arrangements agreed in these States. 4. Conclusion Given the inherent risks of imperfect knowledge and bias on the part of the ECJ, and given the potentially severe impact of mistaken decisions in matters of Member State federalism, a good case may be made for a rule-based mode of adjudication. In contrast to a particularistic balancing approach, rules can provide the Court with the necessary guidance and prevent it from being misled by ignorance and prejudice. It is interesting to note that a comparable argument for rules has been raised elsewhere in EU law. Indeed, not only in the context of Member State federalism, but also in other situations where the ECJ is distrusted, salvation is expected from rule-based and categorical decision-making. Thus, the European Trade Union Confederation (ETUC) advocates the insertion in the Treaties of a clause establishing, amongst other things, the prevalence of fundamental social rights over internal market freedoms.122 If the Court were subject to this bright-line rule, the 120 See, eg, as regards Belgium—Belgian Constitution, Arts 4(3), 39, 127–130 and 134; as regards Spain—Spanish Constitution, Arts 146–147; Catalan Statute of Autonomy, Arts 222–223; Statute of Autonomy of the Basque Country, Arts 46–47; as regards Italy—Italian Constitution, Art 123. It must be admitted that, in the UK, institutional reform is less formalised and more pragmatic in nature. See, eg, C Jeffery, ‘Devolution in the United Kingdom: Problems of a Piecemeal Approach to Constitutional Change’ (2009) 39 Publius 289; S Tierney, ‘Giving with One Hand: Scottish Devolution within a Unitary State’ in S Choudhry (ed), Constitutional Design for Divided Societies (New York, Oxford University Press, 2008) 439. 121 Provisions concerning federalism or devolution have been successfully challenged, as was exemplified by the following decisions: Spanish Constitutional Tribunal, sentencia 31/2010, 28 June 2010; Belgian Constitutional Court, No 35/2003, 25 March 2003; Italian Constitutional Court, sentenza 378/2004 and sentenza 379/2004, 29 November 2004; sentenza 12/2006, 11 January 2006. 122 ETUC Response to ECJ Judgements Viking and Laval, Resolution adopted by the Executive Committee of the ETUC, Brussels, 4 March 2008, available at , 4. See also Barnard, above n 23, 316; Syrpis and Novitz, above n 23, 416–22.

350 Elke Cloots risk of social rights being undervalued would be minimised. The proportionality assessment currently used by the Court to settle conflicts between social rights and fundamental freedoms obviously does not give this guarantee. Since the Court has, so far, repeatedly tipped the scales in favour of the fundamental freedoms,123 it should not surprise that the ETUC would like to substitute a rigid and easily applicable rule for the Court’s malleable and multifaceted proportionality test. ii. The Rationales behind the Treaty Provisions A second ground upon which the ECJ’s rule-based, categorical approach to issues of Member State federalism may be defended relates to the Treaty provisions under consideration and the rationales behind them. As far as the prohibitions of State aid and discrimination are concerned, the underlying rationale is—roughly speaking—to avoid public authorities intervening to the benefit of certain economic actors, thereby distorting competition in the EU.124 Obviously, there are occasions when this rationale, and hence the prohibition flowing from it, will conflict with and even be outweighed by other interests. Thus, Article 107(2)–(3) TFEU lists a number of countervailing policy objectives (eg, the restoration of damage caused by natural disasters, or the promotion of culture) which do or may legitimise the grant of State aid. Similarly, the ECJ has identified grounds, most notably the CAP objectives laid down in Article 39(1) TFEU, which are capable of justifying different treatment of comparable situations in the agricultural sector.125 Yet Portugal v Commission, UGT-Rioja and Horvath confronted the Court with quite another problem. The issue here was not that a countervailing interest (ie, the need to respect the Member State’s system of power division) was, in the particular circumstances of the case, weightier than the rationale behind the prohibitions of State aid and unequal treatment, such that the Member State in question was permitted to provide aid and to treat comparable situations differently, respectively. The crux was that the rationale underlying these prohibitions was not involved in the first place.126 The autonomous adoption of diverse tax regimes by distinct sub-State polities is simply not what the Member States sought to forbid when inserting a ban on State aid into the Treaty. Likewise, it is fair to say that the different treatment of similar situations by different lawmakers does not implicate

123 See especially Laval, above n 3, para 111; Commission v Germany, above n 3, para 105. In Viking, the proportionality assessment was formally left to the national court (above n 3, paras 80, 83–85 and 87). Yet the ECJ did not hesitate to provide the referring court with detailed ‘guidance’ (para 85). In fact, the national court was left with little choice other than to declare the collective action disproportionate. See, eg Barnard, above n 23, 316. 124 Art 107(1) TFEU; Barents, above n 107, 530–32; Tridimas, above n 107, 79. 125 In the context of the CAP, see, eg, Case 139/77 Denkavit [1978] ECR 1317, paras 15–17; Case 166/78 Italy v Council [1979] ECR 2575, paras 18–20; Case C-292/97 Karlsson [2000] ECR I-2737, paras 45–47. See also Barents, above n 107, 532–34 and 548; Tridimas, above n 107, 83–84. 126 For a more elaborate argument to this effect, see F Schauer, ‘Freedom of Expression Adjudication in Europe and the United States: A Case Study in Comparative Constitutional Architecture’ in G Nolte (ed), European and US Constitutionalism (Cambridge, Cambridge University Press, 2005) 67–69.

The ECJ and Member State Federalism 351 the rationale of the non-discrimination principle. Against this background, the ECJ’s categorical or conceptual approach seems most appropriate. Indeed, in order to ensure that the Treaty prohibitions under consideration remained in line with their purpose, the best option was to fine-tune their scope of application. In a similar vein, it might be argued that intra-State border transgressions do not belong to what the Treaty provisions on the freedom of movement purport to protect. Although such a view of the rationale underlying these Treaty provisions is not universally shared,127 we believe that there are good reasons why the (relatively stringent) EU standard of freedom of movement protection should not be extended to apply to intra-State situations. Limitations of space prevent us from developing a full-fledged theoretical analysis here, but the argument runs roughly as follows. Federal multinational Member States as well as the EU are in quest for an equilibrium between, on the one hand, the unity of the polity as a whole and, on the other hand, the autonomy of their territorial subdivisions. The most satisfactory state of equilibrium may vary between the Union and the Member States, across Member States themselves, and even across regions within a single Member State. It is also likely to fluctuate over time. Where the equilibrium lies for a given multinational polity at a given point in time depends on political compromises and on a wide range of factual circumstances such as history and geography. For this reason, we think it unwise to impose the equilibrium achieved in a particular polity (eg, the EU) on another polity (eg, Spain). Since the standard of free movement protection applicable within a given polity is an essential part of that polity’s unity/autonomy equilibrium, it should not be imposed imperatively on another polity. If it is true that transactions over Member State borders and transactions over regional borders in a single multinational Member State may (and should) be subject to distinct free movement standards, it is sensible, first, to exempt intra-State situations from the scope of the Treaty provisions on freedom of movement and, secondly, to dismiss State structure as a ground of justification for an interference with the EU free movement standard. A categorical or rule-based approach is, therefore, the most suitable mode of adjudication where the Union free movement standard clashes with the standard of a (quasi-)federal multinational State. The alternative is to include intra-State situations in the realm of the Treaty provisions on free movement, and to demand multinational Member States to justify constitutional provisions pertaining to their own internal equilibrium in the light of the EU standard of free movement. This style of analysis seems analytically far less appealing. iii. The Structure of Federalism Provisions A third reason why the ECJ is right to prefer a categorical approach over balancing to adjudicate conflicts between Member State federalism and EU law is connected 127 Cp Government of the French Community and Walloon Government, above n 10, Opinion of AG Sharpston, paras 116–18.

352 Elke Cloots to the structure of federalism provisions. It is useful to recall at this juncture Robert Alexy’s theory of the structure of norms. On Alexy’s account, two types of norms exist: ‘rules’ and ‘principles’. Rules are norms which are always either fulfilled or not. Principles, by contrast, are ‘optimisation requirements’, which means that they require that something be realised to the greatest extent legally and factually possible.128 Interestingly, conflicts of norms are resolved differently depending on whether either rules or principles are at issue. According to Alexy, a conflict between two rules may be resolved only by creating an exception to one of the rules, or by invalidating at least one of them.129 The solution to a conflict of principles, by contrast, lies in the weighing and balancing of competing interests.130 It appears that constitutional norms concerning federalism, either as such or as a result of judicial interpretation, often take the form of ‘rules’ as defined by Alexy.131 They determine which level of government is responsible for a particular subject-matter area (eg, social security, agriculture, education), or define the group of persons or situations falling within a government’s jurisdiction (eg, the residents of the region). Obviously, it is hard to balance these sorts of legal provisions against other rule-type norms on a case-by-case basis.132 The federalism rule’s demand is either satisfied or not satisfied at all. A way out of such a norm conflict lies, therefore, not in the balancing of the conflicting rules but, rather, in the establishment of a hierarchy of norms, the insertion of an exception or the allocation of a separate scope of application to each of the two rules. While the hierarchical approach is necessarily an all-or-nothing affair, the carving out of an exception or distinct niches lets both rules co-exist. Hence, if we care for both Member State federalism and European integration, the last two solutions are likely to be more satisfactory. The unsuitability of balancing was particularly evident in the Flemish Care Insurance Scheme Case. As noted above, this case concerned a rule pertaining to the Belgian federal structure, which established that the Flemish Community’s powers in the field of care insurance were restricted to persons residing in the Flemish territory. Against this rule, EU legal provisions regarding the free movement of persons were invoked. According to these provisions, a person’s social security status is, as a general rule, determined by the laws of the Member State where he 128 R Alexy, A Theory of Constitutional Rights (J Rivers trans) (Oxford, Oxford University Press, 2002) 44–48. Note that Robert Alexy has a different conception of a ‘rule’ than do Frederick Schauer and Kathleen Sullivan. Whereas Alexy regards rules as the opposite of ‘principles’ or ‘optimisation requirements’, Schauer and Sullivan hold the view that rules are to be distinguished from ‘standards’ in that the latter are relatively vague as compared to rules. Moreover, Sullivan regards balancing as a sort of standard. For Alexy, by contrast, balancing is not a particular manifestation of a principle but, rather, a tool for resolving conflicts between principles. 129 Ibid 49–50. 130 Ibid 50–56. 131 However, this is not to say that federalism norms are always ‘rules’, as the principles of subsidiarity, federal loyalty, etc show. 132 See also Jackson, above n 73, 843–44.

The ECJ and Member State Federalism 353 works rather than resides. In other words, the Belgian lex loci domicilii rule and the European lex loci laboris rule were in competition with each other. It is fair to say that balancing would not have provided much assistance in resolving this conflict of rules. The solution lay, instead, in a more categorical mode of adjudication. The Court’s decision to distinguish between two categories of situations—internal and inter-State ones—and allot each of both rules to a category is perfectly defensible in this context.

B. Pitfalls Although the benefits of categories, definitions, limitations of scope, etc for deciding conflicts between Member State federalism and EU law are now clear, one should not lose sight of the pitfalls inherent in such an approach. A first pitfall is that of the unavoidable under- and over-inclusive character of rules. A second recurrent objection against rule-based, categorical decision-making is that it does not significantly differ from judicial analysis based on standards and balancing. In what follows, we address each of these criticisms. i. Under- and Over-inclusiveness One of the major difficulties with rules is that any rule necessarily captures its underlying background justification in an imperfect way. Any rule will usually be over-inclusive, which means that it hits factual situations which do not fit the rule’s background justification. At the same time, rules are likely to be under-inclusive, that is, to exclude some facts which bear relevance to the rule’s justification. Why a rule and its justification do not neatly coincide has been explained by Frederick Schauer. As Schauer points out, rules may be expressed in the following formula: ‘if X, then you may or shall do Y’, where X is the rule’s factual predicate or hypothesis, and Y its consequent. It is characteristic of rules that the factual predicate consists of one or more generalisations, which makes the rule applicable to all instances of a certain category (eg, ‘movement’, ‘goods’, ‘selective aid’, etc). The generalisations making up a rule’s factual predicate are considered to be in a causal relationship to the rule’s justification. However, in practice, generalisations will often also encompass instances which are not causally relevant to the rule’s rationale or, conversely, be insensitive to factors which would serve the rule’s rationale.133 This risk of over- and under-inclusiveness is not merely imaginary. Quite the contrary—its exploration has been ubiquitous in the academic debate surrounding the ECJ’s judgments on conflicts between Member State federalism and key

133 Schauer, above n 76, 23–37. See also Alexander and Sherwin, above n 83, 34–36 and 53–61. See also section IV.A.ii. of this chapter.

354 Elke Cloots principles of EU law.134 Also the Advocates General have devoted ample attention to the question of where the boundaries of the Treaty concepts of ‘purely internal situations’, ‘selective aid’ and ‘discrimination’ ought to be drawn in order correctly to capture the Treaty provision’s background justification.135 It would be beyond the scope of the present inquiry to determine how each of these notions should be defined. Nonetheless, two general remarks seem appropriate here. 1. Importance A first point relates to the importance of drawing conceptual boundaries in a correct manner. If concepts are defined either too broadly or too narrowly, a rule’s factual predicate will not perfectly match the principle or policy which justifies the rule. Consequently, situations which are comparable in the light of the rule’s background justification will be treated differently, and situations which are not will be treated alike. Since a rule’s factual predicate (ie, its X conditions) consists of generalisations, the risk of a rule producing an arbitrary result in an individual case cannot be completely avoided.136 However, that does not absolve the ECJ (or any other rule-making body) from the responsibility of fine-tuning the rule’s factual predicate so as to make it fit the rule’s justification as closely as possible. The arbitrariness flowing from carelessly constructed rules was most obvious in the Flemish Care Insurance Scheme Case. The ECJ rightly excluded ‘purely internal situations’ from the scope of EU free movement law in order to accommodate the Belgian federal structure. However, the Court construed the category of ‘purely internal situations’ extremely narrowly and used an arbitrary criterion to demarcate this category from situations which did have a link with EU law.137 This arbitrary distinction might have been blurred if the referring court (ie, the Belgian Constitutional Court) had applied the ECJ’s decision regarding situations which have a link with EU law to purely internal situations as well. Yet the Belgian Constitutional Court

134 See, eg, E Cloots, ‘Germs of Pluralist Judicial Adjudication: Advocaten voor de Wereld and other References from the Belgian Constitutional Court’ (2010) 47 CML Rev 645, 662–63; C Dautricourt and S Thomas, ‘Reverse Discrimination and Free Movement of Persons under Community Law: All for Ulysses, Nothing for Penelope?’ (2009) 34 EL Rev 433, 444–54; K Lenaerts and N Cambien, ‘Regions and the European Court: Giving Shape to the Regional Dimension of the Member States’ (2010) 35 EL Rev 609, 629–35; T Vandamme, ‘Case C-212/06, Government of the French Community and the Walloon Government v Flemish Government, judgment of the Court of Justice (Grand Chamber) of 1 April 2008’ (2009) 46 CML Rev 287, 295–97; P Van Elsuwege and S Adam, ‘The Limits of Constitutional Dialogue for the Prevention of Reverse Discrimination’ (2009) 5 European Constitutional Law Review 327, 332–39. 135 See especially Portugal v Commission, above n 89, Opinion of AG Geelhoed, paras 41–79; UGT-Rioja and Others, above n 25, Opinion of AG Kokott, paras 39–116; Government of the French Community and Walloon Government, above n 10, Opinion of AG Sharpston, paras 112–57; Horvath, above n 36, Opinion of AG Trstenjak, paras 102–18. 136 Schauer, above n 76, 135–37. 137 This point will be further developed in section IV.B.ii.2. below. See also Cloots, above n 134, 662–63.

The ECJ and Member State Federalism 355 was unwilling to do so, since that would have been in violation of the requirements inherent in Belgium’s federal structure.138 Given this unsatisfactory state of affairs, it is hardly surprising that numerous commentators now advocate the abolition of the distinction between purely internal and other situations, at least in certain circumstances, such that EU law would also cover a multitude of intra-State situations.139 In that event, the best we might hope for is that the ECJ would be willing to balance the need to protect domestic structural provisions against the fundamental freedoms.140 However, as demonstrated above, there are major difficulties with the use of a balancing approach for adjudicating conflicts between Member State federalism and freedom of movement. It is submitted, therefore, that it would be more appropriate if the ECJ continued exempting purely internal situations from the scope of EU free movement law, but employed a more relevant demarcation criterion than it does currently. 2. Methodology Our second general observation concerns the methodology to be used to define concepts appearing in the factual predicate of a rule. As already noted, the rulemaker should ensure that the background justification of the rule is well captured in the factual predicate. Consequently, one might reasonably expect that, in the type of case under consideration, the boundaries of a legal concept will depend on—and vary according to—the elements of European integration and domestic federalism the rule-maker wishes to reconcile. Seen in this light, it is remarkable to observe that several Advocates General and commentators recommend that regional measures interfering with the freedom of movement, the prohibition on State aid or the non-discrimination principle be subject to a uniform test.141 Two examples, however, suffice to demonstrate the need for different tests in different situations. To begin with, there are good reasons to believe that the threefold autonomy test designed by the Court to verify whether a regional tax reduction constitutes ‘selective’ State aid would render the Treaty prohibition of discrimination overinclusive if it were equally applied to check whether divergent regional implementations of an EU measure are ‘discriminatory’. Put differently, the autonomy test 138 Belgian Constitutional Court, No 11/2009, 21 January 2009, paras B.11.1–B.17. See also Stef Feyen’s contribution in ch 17 of this volume. 139 See, eg, Dautricourt and Thomas, above n 134, 444 ff; A Tryfonidou, ‘The Outer Limits of Article 28 EC: Purely Internal Situations and the Development of the Court’s Approach through the Years’ in C Barnard and O Odudu (eds), The Outer Limits of European Union Law (Oxford, Hart Publishing, 2009) 218–23. See also Case C-34/09 Ruiz Zambrano (ECJ, 8 March 2011), Opinion of AG Sharpston, paras 69–97, 139–50 and 156–77. 140 For a proposal in this direction, see Vandamme, above n 134, 297–99. 141 Government of the French Community and Walloon Government, above n 10, Opinion of AG Sharpston, paras 117–18; Horvath, above n 36, Opinion of AG Trstenjak, paras 102–08; Lenaerts and Cambien, above n 134, 634; Van Elsuwege and Adam, above n 134, 339.

356 Elke Cloots which matches the rationale behind the Treaty prohibition of State aid would be excessively burdensome for a Member State trying to prove that its territoriallydifferentiated implementation of EU law does not amount to discrimination. It would be unreasonable, for instance, to require that regional implementation measures be taken without the central government being able to intervene directly as regards its content (second autonomy criterion). In many (quasi-)federal Member States, the central level retains the power to intervene where necessary in order to ensure compliance with EU law obligations.142 Furthermore, it makes little sense to hold that a regional implementation measure has not been adopted in an autonomous way merely because the regional authority in question does not bear the full financial burden of its own policy choices (third autonomy criterion). If that were the case, numerous regional implementation measures would be in violation of the EU non-discrimination principle. Indeed, many regions fund the exercise of their self-governing powers in the field of, for instance, agriculture with the help of the central government and/or other regions. Moreover, as a matter of principle, it is hard to maintain that a Member State discriminates amongst its own citizens by giving regional authorities not only the power but also the money to make their own policy choices.143 The inappropriateness of simply copying the test developed in the State aid judgments to other contexts becomes even clearer when we turn our attention to the Flemish Care Insurance Scheme Case, which involved the freedom of movement. The Advocate General in this case argued that ‘the entities between which [free movement] barriers actually need to be abolished are not necessarily always the Member States, but the entities that have the relevant regulatory authority (be these Member States or decentralised authorities within a single Member State)’.144 The Advocate General stressed the analogy between her proposal and the ECJ’s analysis in the State aid judgments, where the Court was ‘prepared to examine the impact of such aid by specific reference to the area in which an autonomous infra-State body exercised its competences’.145 However, the analogy drawn by the Advocate General is misleading. Applying a Treaty provision to an autonomous region instead of the Member State as such may have a totally different impact depending on the EU measure and the

142 See, eg, Scotland Act 1998, ss 35 and 57(1); Memorandum of Understanding between the United Kingdom Government, the Scottish Ministers, the Welsh Ministers, and the Northern Ireland Executive Committee 2010, Art 21, and the attached Concordat on Co-ordination of European Union Policy; Belgian Special Law on Institutional Reform 1980, Art 16(3); Spanish Constitution, Arts 149(3), 150(3) and 155. Moreover, there is a practice in Spain that the Autonomous Communities wait for the central Government to adopt a general framework law for the implementation of EU legislation before issuing their own implementing laws. See Bengoetxea, above n 118, 58–59. 143 See also Hunt, above n 102, 440. 144 Government of the French Community and Walloon Government, above n 10, Opinion of AG Sharpston, para 117. 145 Ibid fn 67. See also Dautricourt and Thomas, above n 134, 453–54; Van Elsuwege and Adam, above n 134, 337–39.

The ECJ and Member State Federalism 357 domestic structural arrangement being at issue. Whereas this method is useful for reconciling the EU’s interest in the eradication of State aid and discrimination, respectively, and the domestic interest in devolution, it would entail an extension of the hegemony of EU free movement law to purely internal situations and, consequently, make residence-based devolution of social powers next to impossible. Once again, this example reveals that similar-looking tests and categories may facilitate the reconciliation of European integration and Member State federalism in one context but fail to advance (or even undermine) that objective in other areas of the law. ii. A Genuine Alternative to Balancing? A second critique of a rule-based, categorical decision-making style is that it does not provide a genuine alternative to balancing. This critique has two versions. One version holds that what appears, at first sight, to be a categorical approach will often also involve a tacit balancing inquiry. According to these critics, decisionmakers should be more honest and do their balancing openly.146 A second version of the critique acknowledges that rules and standards do differ, but opens our eyes to the fact that rule-subjects tend to employ ‘rule-avoiding devices’. Despite their dissimilarities, the essence of both versions of the critique may be captured as follows: categories and rules are not capable of restricting a decision-maker’s discretion as much as their proponents would like to believe. Hence, we should not give up the distrust we feel of a given judicial body solely because that body is subject to rules and categories. 1. First Version As far as the first version of the objection is concerned, indeed it cannot be excluded that a kind of balancing exercise is involved in the demarcation of concepts which belong to a rule’s factual predicate. Advocate General Kokott, for instance, saw in the Court’s reading of the notion of ‘selective aid’ in Portugal v Commission the expression of a ‘reasonable balance’ between the Union’s duty to respect the national identities of its Member States, on the one hand, and the prohibition on the Member States hiding behind their constitutional order to justify an interference with EU law, on the other hand.147 In a similar vein, it might be argued that the exemption of ‘purely internal situations’ from the scope of EU free movement law in the Flemish Care Insurance Scheme Case and the Court’s interpretation of the concept of ‘discrimination’ in Horvath are the result of an implicit balancing act between respect for Member State federalism and European integration. 146

For a discussion of this objection, see Schauer, above n 126, 56–57 and 64. UGT-Rioja and Others, above n 25, Opinion of AG Kokott, paras 54–58. See also Lenaerts and Cambien, above n 134, 634. 147

358 Elke Cloots Nonetheless, it is fair to say that the tacit balancing discernible in these federalism judgments differs in kind from the Court’s balancing approach to conflicts between constitutional rights and fundamental freedoms. Balancing is exercised at a more abstract level in federalism judgments than it is in constitutional rights decisions. In the latter type of decision, the Court verifies whether, in the particular circumstances of the case before it, the protection or exercise of the constitutional right at issue entails a proportionate interference with the right to freedom of movement invoked by the individual claimant. This proportionality inquiry is conducted from scratch in each individual case. However, in decisions on Member State federalism, balancing takes place at a higher level of generalisation. It is the more abstract interest in an internal market where movement and competition are free and discrimination is prohibited which is weighed against the equally abstract interest in devolution of powers in multinational Member States. The product of this rather theoretical balancing exercise is, then, a rule, which may easily be applied in subsequent cases without the need for further balancing.148 Such a rule presumably offers a better guarantee than case-specific balancing that Member State federalism will be valued properly.149 Melville Nimmer has famously described this kind of abstract, rule-generating balancing as definitional balancing.150 The three-step autonomy test developed by the ECJ to delineate selective State aid, the distinction drawn between purely internal and other situations in free movement cases, and the exclusion of territorial differentiation in (quasi-)federal Member States from the scope of the discrimination prohibition may all be rephrased in these terms. Any of those rules, tests and definitions emerged from a preceding abstract and implicit balancing act, and make, at the same time, balancing in future individual cases superfluous. As a consequence, it may be assumed that the ECJ’s discretion will be more constrained in federalism cases than in constitutional rights cases. 2. Second Version Advocates of the use of a rule-based style for adjudicating conflicts between Member State federalism and values of European integration should, however, be more concerned about the second version of the critique. According to this version, decision-makers will likely transform the rules to which they are subject into standards. The leading proponent of this view is Frederick Schauer. To be sure, Schauer does not contest that rules and standards are different in principle. Whereas rules are (comparatively) precise, standards are (comparatively) 148 See, eg, NT Deutsch, ‘Professor Nimmer Meets Professor Schauer (and Others): An Analysis of “Definitional Balancing” as a Methodology for Determining the “Visible Boundaries of the First Amendment”’ (2006) 39 Akron Law Review 483, 484–500; DL Faigman, ‘Reconciling Individual Rights and Government Interests: Madisonian Principles versus Supreme Court Practice’ (1992) 78 Vanderbilt Law Review 1521, 1536; Nimmer, above n 83, 944–45. 149 Cp Schauer, above n 126, 64. 150 Nimmer, above n 83, 942–48.

The ECJ and Member State Federalism 359 vague.151 However, Schauer challenges the widespread assumption that a varying degree of norm specificity necessarily entails a varying degree of discretion for the body which must interpret and apply the norm in a concrete fact situation. According to conventional wisdom, (specific) rules constrain decision-maker discretion more than do (vague) standards. Yet, as Schauer convincingly shows, this hypothesis fails to take account of, amongst other things, the wide range of techniques rule-subjects employ in order to avoid rules.152 A court bound by a certain rule may, for example, create an exception in a particular case, thus making the rule more standard-like. Hence, the discretion-constraining effect of a rule will often be less pronounced in practice than the precise wording of the rule might suggest. The rule-avoiding behaviour described by Schauer may clearly be identified in the ECJ judgments concerning freedom of movement and Member State federalism. Indeed, the Court has developed multiple devices which enable it to get round the rule that purely internal situations are not covered by the Treaty provisions on freedom of movement. To begin with, in the Flemish Care Insurance Scheme Case, the concept of ‘purely internal situations’ was interpreted in a very narrow and capricious way. If a Belgian national who worked and lived in Belgium had ever exercised his right to free movement in the EU, the ECJ regarded him as falling within the scope of EU law.153 The Court did not require that person’s earlier movement to have any relevant connection to the benefit he presently claimed. The line between situations which were covered by EU free movement law and those which were not was, therefore, fraught with arbitrariness. In addition, the Court pointed out that the exclusion of purely internal situations from the scope of EU law did not mean that the ECJ’s interpretation of EU law could not be useful to the national court when deciding on internal situations.154 This suggestion, in combination with the arbitrary nature of the distinction drawn by the ECJ, left the referring court with little choice other than to extend the ECJ’s decision on inter-State situations to intra-State situations.155 Lastly, it should be noted that the ECJ does not always confine its decisions to situations which have a cross-border character. In Presidente del Consiglio dei Ministri v Regione Sardegna, for instance, a case involving the compatibility of a regional tax law with the fundamental freedoms, the Court merely verified whether the facts of the case contained an inter-State element. Once such an 151

Schauer, above n 77, 304–11. Ibid 312–15. Schauer shows that the conventional view equally fails to acknowledge the opposite tendency of courts, that is, the tendency to convert standards into rules (ibid 315–18). The combination of both adaptive tendencies yields the ‘convergence hypothesis’ (ibid 319–28). See also Stone, above n 73, 694. 153 Government of the French Community and Walloon Government, above n 10, paras 34, 37 and 41–42. 154 Ibid para 40. 155 This was confirmed by the Opinion of AG Sharpston in Ruiz Zambrano, above n 139, fn 118. The Belgian Constitutional Court did not succumb to the European pressure, however. See Belgian Constitutional Court, No 11/2009, 21 January 2009, paras B.11.1–B.17. 152

360 Elke Cloots element had been found, the ECJ decided the case as a whole, without strictly limiting the reach of its judgment to inter-State situations.156 In other judgments, the Court went even further by explicitly including intra-State trade within the scope of EU law.157 Although it must be admitted that the regional or local measures at issue were qualified as customs duties (or charges having equivalent effect) rather than obstacles to freedom of movement stricto sensu, it might be only a matter of time before the scope of EU free movement law will be extended so as to cover intra-State situations.158 This brief case study shows that the rule establishing the exemption of purely internal situations from the scope of EU law is riddled with judge-made exceptions and qualifications. The rule is, therefore, not capable of constraining the ECJ’s discretion as much as originally envisaged. Is this a reason for abandoning the rule altogether and opting for a truly standard-based style of adjudication? We do not think so. If we want the ECJ to take Member State federalism seriously in the face of the fundamental freedoms, little is to be expected from a particularised balancing exercise. A more effective alternative may be to put greater effort into constraining the ECJ’s discretion in this type of case, for instance by making the rule more precise and/or entrenching it in the Treaties.159

V. CONCLUSION

This chapter started from a fairly simple observation: whereas the ECJ is prepared to reconcile and balance the potentially competing interests in freedom of movement and constitutional rights protection, it regards a Member State’s (quasi-)federal arrangements as necessarily outweighed by the internal market freedoms. This difference in treatment of constitutional rights and structural provisions might indicate that the ECJ considers the latter kind of constitutional norms less important 156 See, eg, Presidente del Consiglio dei Ministri v Regione Sardegna, above n 96, paras 28, 30–32 and 50. In other cases, the ECJ ruling did seem restricted to inter-State situations, but was nonetheless extended to intra-State situations by a national court. See, eg, Case C-464/05 Geurts and Vogten [2007] ECR I-9325, para 29 and Belgian Constitutional Court, No 83/2010, 8 July 2010, para B.9. 157 See, eg, Case C-163/90 Legros and Others [1992] ECR I-4625, paras 16–18; Joined Cases C-363/93, C-407/93, C-408/93, C-409/93, C-410/93 and C-411/93 Lancry and Others [1994] ECR I-3957, paras 25–32; Joined Cases C-485/93 and C-486/93 Simitzi [1995] ECR I-2655, paras 24–27; Case C-72/03 Carbonati Apuani [2004] ECR I-8027, paras 22–26; Case C-293/02 Jersey Produce Marketing Organisation [2005] ECR I-9543, paras 61–67. 158 See, eg, Government of the French Community and Walloon Government, above n 10, Opinion of AG Sharpston, paras 122–32. 159 See, in a similar vein, Schauer, above n 77, 328. See also, in the context of the First Amendment to the US Constitution, JH Ely, Democracy and Distrust: A Theory of Judicial Review (Cambridge, Mass, Harvard University Press, 1980) 112: ‘One doesn’t have to be much of a lawyer to recognize that even the clearest verbal formula can be manipulated. But it’s a very bad lawyer who supposes that manipulability and infinite manipulability are the same thing. An “unprotected messages” approach [being a categorical approach] cannot guarantee liberty—nothing can—but it’s the surest hedge against judicial capitulation that humans have available.’ See also Nimmer, above n 83, 945; Stone, above n 73, 694–96.

The ECJ and Member State Federalism 361 than constitutional rights. However, this contribution puts forward a normatively more satisfactory justification for the distinction. It argues, more precisely, that the ECJ’s refusal to put a Member State’s structural arrangements on the balancing scales does not mean that these arrangements are left defenceless against the fundamental freedoms but, rather, that they are protected in another, more suitable way. This alternative mode of adjudication may aptly be described as ‘categorical’ and ‘rule-based’. In other areas of EU law too, categorisation and rule-based decisionmaking have proved useful tools for reconciling EU interests and domestic devolution of powers. One of the most powerful arguments in favour of the ECJ’s categorical and rule-based approach in domestic federalism cases is that rules are relatively precise, and therefore restrict judicial discretion more than do standards. Why is a limitation of the Court’s room for manoeuvre in this type of case to be welcomed? The reason is that the ECJ, due to insufficient understanding and bias, is likely to undervalue a Member State’s federal arrangements when balancing them against core Treaty provisions. In addition, this undervaluation may have very serious consequences for the State. By constraining the ECJ’s discretion, bright-line rules reduce the risk of such potentially disastrous judicial mistakes. They can help the ECJ resist the temptation of prioritising European integration, just as the ropes that tied Ulysses to the mast of his ship prevented him from succumbing to the enchanting song of the Sirens.160 Although a rule-based, categorical approach is thus preferable over balancing, this does not mean that Member State federalism is saved as soon as the Court is bound by rules. To begin with, a rule-based and categorical mode of adjudication does not, in and of itself, assure that conceptual boundaries are drawn properly. Moreover, like Ulysses, courts will try to loosen their ties, for example by qualifying the interpretation of and adding exceptions to the rule. The ECJ’s decision on the Flemish care insurance scheme amply illustrates that these are major challenges for rule-based adjudication. However, these challenges are not insurmountable, as the judgments regarding the prohibition on State aid show. In Portugal v Commission, definitions were carefully crafted, expressing an appropriate equilibrium between Member State federalism and European integration. Moreover, in UGT-Rioja and Others—the rule’s first test case—the Court did not attempt to get round the rule but construed it in a way that is truly respectful of Member State federalism. The island of the Sirens was safely navigated …

160 I borrow this metaphor from Kathleen Sullivan. See Sullivan, above n 75, 73; Sullivan, above n 74, 296.

15 The Impact of ‘Regional Blindness’ on the Italian Regional State GIUSEPPE MARTINICO*

I. INTRODUCTION

T

O WHAT EXTENT do Italian courts adapt the national legal instruments regarding State structure to the requirements of EU law? This chapter aims to answer this question by providing an overview of the most emblematic ‘adaptations’ made by the Italian courts in order to ensure respect for EU law. In doing so, a distinction is made between two kinds of principles/rules (ie, norms) and techniques (ie, operative emanations of norms), which were originally conceived for other goals but were in a second phase used by the Italian courts in order to guarantee compliance with the EU requirements. The first group is composed of those norms and techniques expressly conceived for governing the relationship between the Regions and the Italian State, which are more substantive in nature and have been reshaped over the years by the Italian Constitutional Court. The second group consists of norms and techniques characterising the proceedings before the Italian Constitutional Court, which also serves as final arbiter in conflicts between the Regions and the State. The latter norms and techniques are genuinely procedural. Over the years, the Italian Constitutional Court has used them in order to guarantee first of all the supremacy of the Constitution and, episodically, the primacy of EU law. These norms and techniques do not specifically relate to the relationship between the State and the Regions but they can, of course, be used in order to ensure respect for the Constitution and EU law by regional legislatures. Keeping this in mind, this contribution is structured as follows. First, I shall explain the reasons why research like this is ‘difficult’; then I shall turn to the analysis of the two groups of legal instruments mentioned above. Some final * García Pelayo Fellow, Centro de Estudios Politicos y Constitucionales, Madrid; Max Weber Fellow, European University Institute, Florence (2010–11). Researcher at the Centre for Studies on Federalism, Turin; Adjunct Professor, Scuola Superiore Sant’Anna, Pisa. I would like to thank Elke Cloots, Geert De Baere, Giacomo Delledonne, Filippo Fontanelli, Marco Mazzarella, Francesco Palermo and Stefan Sottiaux for their comments and suggestions. The usual disclaimers apply.

Impact of ‘Regional Blindness’ on Italian Regional State 363 remarks will be presented at the end of the chapter. Generally speaking, my main argument is that EU law has had a certain impact on the relationship between the State and the Regions in Italy. The evolving meaning of the principle of conferred powers enshrined in Italian constitutional law is a first example of that impact. The principle has been understood more and more as referring to the idea of ‘legislative preference’ rather than the existence of a ‘legislative reserved domain’.1 Other evidence of the influence of EU law may be found in the extension of the use of instruments born in Italian administrative law to the sphere of legislative acts. This cross-fertilisation may be explained in the light of the fungible character of administrative and legislative acts with regard to the implementation of EU law measures, this being a leitmotif of the case law of the European Court of Justice (ECJ), which can be traced back to the original absence of a hierarchy of legal sources in the EC/EU law context.2 Given Italy’s unitary judicial system and given the particular role of arbiter between the State and the Regions attributed to the Italian Constitutional Court, I shall focus on the Constitutional Court’s case law when developing my reasoning in these pages. Thus, special attention will be paid to the constitutional principaliter proceedings. Principaliter or direct action proceedings are the arena of the legislative conflicts occurring between the State and the Regions. Obviously, this does not mean that ‘common’ courts (ie, ordinary and administrative courts) have not adjusted legal concepts under pressure from EU law. Quite the contrary is true, as has been amply shown in the domestic literature.3 A preliminary caveat should be issued before proceeding with the analysis of the norms and techniques reinvented by the Italian courts to guarantee that the principles of EU law are respected. One should keep in mind the strongly dogmatic nature of the Italian legal system: while the ECJ has been inspired by a strong pragmatism, which gave its case law a certain flexibility, the Italian Constitutional Court is the guardian of a set of principles imbued with the dogmatic flavour of its legal tradition, very refined from a theoretical point of view, but also resistant to being reshaped. Bearing this in mind, one can appreciate the efforts made by both courts in their attempt to carry out a sort of convergence, renouncing, at least partly, their original positions (ie, dualism for the Italian Constitutional Court and monism for the ECJ). Lastly, it should be noted at the outset that at least two factors further complicate this research: the well-known principle of ‘territorial blindness’ of the EU, and the less-studied ‘procedural impermeability’ that characterises the case law 1

L Paladin, Le fonti del diritto italiano (Bologna, Il Mulino, 2000) 93 ff. G Gaja, ‘Fonti comunitarie’ in Digesto delle Discipline Pubblicistiche (Torino, VI, UTET, 1991) 436 ff. 3 F Cortese, ‘ECJ and Administrative Courts in EU Member States: Towards a Common Judicial Reasoning?’ in F Fontanelli, G Martinico and P Carozza (eds), Shaping Rule of Law through Dialogue. International and Supranational Experiences (Groningen, Europa Law Publishing, 2010) 257–71; M Gnes, ‘Giudice amministrativo e diritto comunitario’ (1999) Rivista trimestrale di diritto pubblico 331; N Pignatelli, ‘L’illegittimità comunitaria dell’atto amministrativo’ (2008) Giurisprudenza Costituzionale 3635. 2

364 Giuseppe Martinico of the Italian Constitutional Court. Territorial blindness refers to the fact that, traditionally, from a ‘formalistic’ point of view, the sub-State level has been neglected in the EU law context. To express such a situation, German constitutional lawyers use the term ‘Landesblindheit’.4 This blindness was confirmed in the Treaties (especially in the former Article 10 EC Treaty), which designate the Member States as the subjects of the EU legal order. It might well be argued that this ‘regional neglect’ constitutes just one element of the democratic deficit of the EU. Taking a broad conception of the democratic gap,5 we may conceive the absence of a strong legal status for the regions as one of the most important ‘constitutional wounds’ of the EU, even after the coming into force of the Lisbon Treaty. There are a number of reasons for that: the internationalist origin of the European enterprise; the heterogeneous composition of the Committee of Regions; and the lack of specific remedies for violations of regional prerogatives. However, despite their unclear status under EU law, regions do play a fundamental role in the implementation of EU policies. Procedural impermeability, on the other hand, refers to the general reluctance of the Italian Constitutional Court to get involved in interpretative questions dealing with EU law. As we shall see in section III. of this chapter, the origin of this doctrine may be traced back to Granital.6 In that judgment, the Constitutional Court decided to attribute the role of natural guardians of the ECJ’s Simmenthal doctrine to the ordinary judges.7

II. SUBSTANTIVE NORMS AND TECHNIQUES CONCERNING THE RELATIONSHIP BETWEEN STATE AND REGIONS

In this section, three important principles/techniques governing the relationship between the Italian State and the Regions will be discussed: the principle of conferred powers; the cedevolezza technique; and the principle of subsidiarity. It will be explored, more particularly, in what ways the Italian Constitutional Court’s reading of these legal instruments has been influenced by EU law. 4 HP Ipsen, ‘Als Bundesstaat in der Gemeinschaft’ in E von Caemmerer, HJ Schlochauer and E Steindorff (eds), Probleme des europäischen Rechts: Festschrift für Walter Hallstein (Frankfurt am Main, Klostermann, 1966) 248. 5 R Bellamy, ‘Still in Deficit: Rights, Regulation and Democracy in the EU’ (2006) 12 European Law Journal 725; B Crum, ‘Tailoring Representative Democracy to the European Union: Does the European Constitution Reduce the Democratic Deficit?’ (2005) 11 European Law Journal 452; A Føllesdal and S Hix, ‘Why There is a Democratic Deficit in the EU: A Response to Majone and Moravcsik’ (2006) Journal of Common Market Studies 533; G Majone, ‘Europe’s “Democratic Deficit”: The Questions of Standards’ (1998) 4 European Law Journal 5; A Moravcsik, ‘In Defence of the “Democratic Deficit”: Reassessing Legitimacy in the European Union’ (2002) Journal of Common Market Studies 603; JHH Weiler, ‘Bread and Circus: The State of European Union’ (1998) Columbia Journal of European Law, 223–24. 6 Corte Costituzionale, sentenza No 170/1984, Granital, 8 June 1984. See also G Gaja, annotation of Granital (1984) 21 CML Rev 756; M Cartabia, Principi inviolabili e integrazione europea (Milan, Giuffrè, 1995). 7 Case 106/77 Simmenthal [1977] ECR I-62.

Impact of ‘Regional Blindness’ on Italian Regional State 365 A. The Principle of Conferred Powers According to the principle of conferred powers,8 the State and the Regions may act only within the limits of the competences conferred upon them by the Constitution to attain the objectives set out in the fundamental charter. In Italian constitutional law, there are three types of competences.9 The first type includes competences exclusive to the State. These are listed in Article 117(2) of the Italian Constitution. The second category consists of shared (or concurring) competences, which are equally expressly enumerated in the Constitution (Article 117(3)). In this area, the State and the Regions co-legislate. While it is for the State to set out the fundamental principles in so-called ‘framework laws’ (leggequadro), each Region is free to implement the national framework through more detailed provisions. This model of ‘shared competence’ is evidently different from the German system of konkurrierende Gesetzgebung.10 When the German federal State and its Länder share a competence, the Land ‘shall have power to legislate so long as and to the extent that the Federation has not exercised its legislative power by enacting a law’.11 According to the Italian model, by contrast, the activities of the two actors are conceived of as ‘complementary’. Thirdly, Article 117(4) of the Italian Constitution proclaims that the Regions have legislative powers in all subject matter areas that are not expressly covered by State legislation, ie that are not included in the lists of powers laid down in Article 117. These regional powers should be understood as exclusive to the Regions. The transformation of the Regions into ‘residual legislators’ was the most important novelty of the 2001 Italian constitutional reform. It might be seen as a sort of Copernican revolution. At least on paper, it looks as if the Regions have been vested with considerable autonomy. In reality, however, a very different picture seems to emerge. This is due to some judgments of the Italian Constitutional Court, as well as to so-called ‘transversal matters’ or clauses allowing the State to intervene in different areas.12 An example of such a transversal matter is the 8 On the principle of competence in the Italian legal system, see G Zanobini, ‘Gerarchia e parità tra le fonti’ in Studi in onore di S Romano, vol 1 (Padua, Cedam, 1939) 303; V Crisafulli, ‘Gerarchia e competenza nel sistema costituzionale delle fonti’ (1960) Rivista trimestrale di diritto pubblico 775; V Crisafulli, Lezioni di diritto costituzionale: L’ordinamento costituzionale italiano (le fonti normative) (Padua, Cedam, 1993) 234 ff; A Moscarini, Competenza e sussidiarietà nel sistema delle fonti (Padua, Cedam, 2003); Paladin, above n 1, 93 ff. 9 A D’Atena, ‘Materie legislative e tipologia delle competenze’ (2003) Quaderni Costituzionali 15. 10 On this distinction, see A Anzon, ‘La delimitazione delle competenze dell’Unione Europea’, available at . 11 German Basic Law, Art 72. 12 F Benelli, La ‘smaterializzazione delle materie’. Problemi teorici ed applicativi del nuovo Titolo V della Costituzione (Milan, Giuffrè, 2007); F Benelli, ‘L’Ambiente tra “smaterializzazione” della materia e sussidiarietà legislativa’ available at ; R Bin, ‘Materie e interessi: tecniche di individuazione delle competenze dopo la riforma del Titolo V’, paper presented at the workshop organised by the IDAIC—Siena, 25–26 November 2005, 5; P Carrozza, ‘Le “materie”: uso delle tecniche di enumerazione materiale delle competenze e modelli di riferimento del “regionalismo di esecuzione” previsto dal nuovo Titolo V della Costituzione’ in GF Ferrari and G Parodi (eds), La

366 Giuseppe Martinico legislation on the ‘determination of the basic level of benefits relating to civil and social entitlements to be guaranteed throughout the national territory’. This kind of legislation may obviously serve as a Trojan Horse, leading to the centralisation of powers. Having recalled the different types of competence, we shall now look at how European commitments have reshaped the distribution of powers enshrined in Italian constitutional law.13 As this chapter attempts to make clear, many instances of an EU law-driven reform may be found in the case law of the Italian Constitutional Court. The engine of this process has been the above-mentioned territorial blindness of the EU. Under EU law, the Member State is the only entity liable for non-compliance with the requirements of EU law. On several occasions, the ECJ has shown itself not to care about the domestic power division system when dealing with cases of non-compliance.14 This, in turn, has forced the Italian actors to modify the original distribution of powers enshrined in the Constitution in order to avoid EU penalties.15 To be sure, the Italian Constitutional Court, as we shall see, has always been lenient with adaptations to the list of competences set forth in Article 117. We might say that the European mandate has only increased that predisposition, giving the Italian Constitutional Court an additional reason for centralising competences in favour of the State.16 When investigating the role of the Regions in the implementation of EU law, it is important to distinguish between administrative and legislative activities. As for the administrative side of the implementation, it has been acknowledged that Regions have the power to implement regulations.17 However, when EU law requires a uniform implementation throughout the national territory, the State may co-ordinate the regional administrative acts, discharging its typical role as administrative co-ordinator vis-à-vis the Regions.18 As for the legisla-

revisione costituzionale del titolo V tra nuovo regionalismo e federalismo. Problemi applicativi e linee evolutive (Padua, Cedam, 2003) 69–124. 13 T Groppi, ‘L’incidenza del diritto comunitario sui rapporti Stato-regioni in Italia dopo la riforma del Titolo V’, available at ; MP Chiti, ‘Regioni e integrazione europea’ (1994) Regioni e governo locale 547; A D’Atena, ‘Il doppio intreccio federale: le regioni nell’Unione europea’ (1998) Le Regioni 1401. Recently P Zuddas, L’influenza del diritto dell’Unione europea sul riparto di competenze legislative fra Stato e Regioni (Padua, Cedam, 2010). 14 See, eg Case 169/82 Commission v Italy [1984] ECR 1603; Case 168/85, Commission v Italy [1986] ECR 2945. See also Groppi, above n 13; G Guzzetta, Costituzione e regolamenti comunitari (Milan, Giuffrè, 1994) 171 ff; M Cartabia and V Onida, ‘Le regioni e la Comunità europea’ in MP Chiti and G Greco (eds), Trattato di diritto amministrativo (Milan, Giuffrè, 1997) 605 ff. 15 See, eg, P Bilancia, F Palermo and O Porchia, ‘The European Fitness of Italian Regions’ (2010) 2 Perspectives on Federalism E-1, E-122-E-174 and E-167, available at . 16 Groppi, above n 13. 17 Presidential Decree No 616/1977, Art 6. 18 Art 9 of Law No 86/1989 (Legge La Pergola) set up a complex mechanism aimed at ensuring the implementation of EU law, and introduced the requirement of an annual Community law (legge comunitaria). Legge La Pergola has been repealed by Law No 11/2005 (Legge Buttiglione). See generally L Paladin, ‘Sulle funzioni statali di indirizzo e di coordinamento nelle materie di competenza regionale’ (1971) Giurisprudenza Costituzionale 189; F Bassanini, ‘Indirizzo e coordinamento delle attività

Impact of ‘Regional Blindness’ on Italian Regional State 367 tive implementation, the issue is much more complicated. Originally, the risk of liability for non-compliance with EU law prompted the State to centralise competences to the detriment of the Regions.19 Later on, the Regions were granted the power to implement Community directives.20 However, they could not exercise their power before a State law had been passed containing the fundamental principles of the directive’s implementation. We shall come back to this point in the next sub-section, when dealing with the issue of cedevolezza and the more recent evolutions in Italian legislation. How did the Italian Constitutional Court react to the shift of competences that occurred in the name of EU law? First, it seemed to allow for the possibility that EU law requirements might de facto lead to a shift in the original distribution of competences.21 In sentenza No 224/1994,22 the Italian Constitutional Court stated that EU law might sometimes modify the conditions and prerequisites for the exercise of a regional competence. Earlier, in sentenza No 399/87,23 the Court had stated that EU bodies are not required to respect domestic law in full, including provisions on the distribution of competences between the State and the Regions. At the same time, however, the Court stressed the need to comply with the fundamental principles of the Italian constitutional system, the so-called ‘counter-limits’.24 In the case at hand, EC legislation had provided the Regions with functions that partly differed from those included in the constitutional catalogue of competences. The leading case on the matter is sentenza No 126/1996,25 where the Italian Constitutional Court pointed out that the implementation of EU norms must take the Member States’ structure (centralised, decentralised, federal) into account. However, in the face of the principle of State responsibility for the implementation of EC law, the Court distinguished between the Regions’ firstorder competence and the State’s second-order competence. This second-order regionali’ in D Serrani (ed), La via italiana alle regioni (Milan, Giuffè, 1972) 43 ff; L Carlassare, ‘I problemi dell’indirizzo e coordinamento: le soluzioni giurisprudenziali’ (1985) Le Regioni 29; G Falcon, ‘Spunti per una nozione della funzione di indirizzo e coordinamento come vincolo di scopo’ (1989) Le Regioni 1184. 19

Presidential Decree No 4/1972. Presidential Decree No 616/1977, Art 6. 21 See Corte Costituzionale, sentenza No 382/1993, 18 October 1993, and sentenza No 632/1988, 6 June 1988. Groppi, above n 13; P Caretti and G Strozzi, ‘Luci ed ombre nella più recente giurisprudenza costituzionale in materia di adempimento agli obblighi comunitari’ (1988) Le Regioni 196. However, for a prima facie opposite conclusion, see sentenza No 124/1990, 7 March 1990. 22 Corte Costituzionale, sentenza No 224/1994, 15 June 1994. 23 Corte Costituzionale, sentenza No 399/87, 11 November 1987. 24 On the counter-limits doctrine, see M Cartabia, ‘The Italian Constitutional Court and the Relationship between the Italian Legal System and the European Union’ in A Slaughter, A Stone Sweet and JHH Weiler (eds), The European Court and National Courts—Doctrine and Jurisprudence (Oxford, Hart Publishing, 1997) 133–46; P Ruggeri Laderchi, ‘Report on Italy’ in Slaughter et al (eds), ibid 147–70. The formula ‘controlimiti’ was used for the first time by P Barile, ‘Ancora su diritto comunitario e diritto interno’ in Studi per il XX anniversario dell’Assemblea costituente, vol VI (Florence, Vallecchi, 1969) 49 ff. 25 Corte Costituzionale, sentenza No 126/1996, 17 April 1996. 20

368 Giuseppe Martinico competence permits the State to intervene in order to avoid the risk of liability for non-compliance due to omissions or actions by its Regions. However, it does not allow the State to seize the Regions’ constitutional powers.26 This line of thought also explains why the State may challenge the constitutionality of regional laws that do not comply with EU law obligations before the Italian Constitutional Court. This provides an exception to the ‘procedural impermeability’ characteristic of the Court’s case law, an issue to which we shall return in section III. Sentenza No 126/1996 reflects the outcome of a long jurisprudential journey, whereby the possibility of an alteration of the constitutional distribution of competences for the sake of compliance with EU law was recognised once again.27 At the same time, the Constitutional Court stated that such alterations should be conceived of as exceptions and not as the rule, trying to prevent EU law undermining the national constitutional system. In order to balance the constitutional supremacy and the primacy of EU law, the Italian institutions have been forced to invent and adopt co-operative instruments capable of pushing the State and the Regions towards specific agreements in this regard. The Italian Constitutional Court has established the unconstitutionality of certain State laws adopted without respect for these co-operative procedures.28

B. The Substitution Power and Cedevolezza Italy has a long municipal tradition. The origins of the Italian municipalities may be traced back to the Middle Ages. In that era the comuni were emblematic of the urban revolt against the feudal system.29 Further evidence of Italy’s long municipal tradition is provided by the fact that the National Association of Italian Municipalities was set up as early as 1901. To be sure, Italian municipalities passed through different phases in their history. Most notably, their autonomy was largely affected by fascist centralism. Today, however, municipalities represent Italy’s basic administrative sub-unit.30 Why highlight the Italian comuni if this piece is supposed to focus on the Regions? Because some of the instruments the Italian constitutional system provides for the reconciliation of regional autonomy and the principle of the unity of the Italian Republic31 find their origin in the legislation on municipal autonomy. The best example thereof is the State’s substitution or replacement power.32 26

On this, see Groppi, above n 13. See, amongst many other judgments, Corte Costituzionale, sentenza No 382/1993, 18 October 1993; sentenza No 389/1995, 20 July 1995. 28 Corte Costituzionale, sentenza No 203/2003, 3 June 2003; sentenza No 68/2008, 10 March 2008. 29 See, eg, M Ascheri, La città-Stato (Bologna, Il Mulino, 2006). Of course, those comuni cannot, in terms of functions, be compared to the current municipalities. 30 Italian Constitution, Art 118. 31 Italian Constitution, Art 5 (‘The Republic, one and indivisible, recognises and promotes local autonomy’). 32 Sometimes also called ‘subsidiary power’ or potere sostitutivo. 27

Impact of ‘Regional Blindness’ on Italian Regional State 369 i. A Posteriori Substitution Until 2001, there was no express mention of EC/EU law obligations in the Italian Constitution. As is well known, the Constitutional Court used to find a constitutional basis for the primacy of European law in Article 11 of the Constitution, although that provision refers only to international organisations.33 In 2001, however, a reference to EU law obligations was included in Article 117(1) of the Constitution, which now reads as follows: ‘Legislative power belongs to the State and the Regions in accordance with the Constitution and within the limits set by European Union law and international obligations.’ In the next sub-section, we shall see that this provision has triggered interesting innovations in the Italian Constitutional Court’s case law. From the outset, its interpretation has divided legal scholars.34 According to some, Article 117(1) simply codifies the existing situation: it gives consent to European primacy as it was developed by the ECJ and accepted across the EU.35 Other scholars, however, emphasise the importance of the constitutional status given to the primacy of EU law, and assert that Article 117(1) paves the way for the acceptance of the Italian monist thesis.36 Leaving this discussion aside, the wording of the constitutional provision leaves little doubt that the obligations imposed by EU law apply to both the regional and the State legislators when exercising their respective constitutional competences.37 The express constitutional reference to EU obligations helps to explain the rationale for the State’s substitution power, which has been codified in Article 120(2) of the Constitution: The Government can act for bodies of the Regions, metropolitan cities, provinces and municipalities if the latter fail to comply with international rules and treaties or EU legislation, or in the case of grave danger for public safety and security, or whenever such action is necessary to preserve legal or economic unity and in particular to guarantee the basic level of benefits relating to civil and social entitlements, regardless of the geographic borders of local authorities. The law shall lay down the procedures to ensure that subsidiary powers are exercised in compliance with the principles of subsidiarity and loyal co-operation.38 33 Art 11 of the Italian Constitution determines, amongst other things, that ‘Italy agrees to limitations of sovereignty where they are necessary to allow for a legal system of peace and justice between nations, provided the principle of reciprocity is guaranteed,’ and ‘promotes and encourages international organisations furthering such ends’. See also Cartabia, above n 24. 34 For an overview, see R Chieppa, ‘Nuove prospettive per il controllo di compatibilità comunitaria da parte della Corte costituzionale’ (2007) 3 Diritto dell’Unione Europea 493, 499 ff; A Ruggeri, ‘Riforma del titolo V e giudizi di “comunitarietà” delle leggi’, available at . 35 See, eg, C Pinelli, ‘I limiti generali alla potestà legislativa statale e regionale e i rapporti con l’ordinamento comunitario’ (2001) V Foro italiano 194 ff. 36 F Paterniti, ‘La riforma dell’art. 117, 1° co. della Costituzione e le nuove prospettive dei rapporti tra ordinamento giuridico nazionale e Unione Europea’ (2004) Giurisprudenza Costituzionale 2101; A Pajno, ‘Il rispetto dei vincoli derivanti dall’ordinamento comunitario come limite alla potestà legislativa nel nuovo Titolo V della Costituzione’ (2003) 5 Le Istituzioni del federalismo 813. 37 For a brief overview of Italian regionalism in English, see M Keating and A Wilson, Federalism and Decentralisation in Italy (PSA Conference, Edinburgh, March–April 2010), available at . 38 Italian Constitution, Art 120.

370 Giuseppe Martinico On closer examination, another reference to the State’s substitution power may be found in Article 117(5) of the Italian Constitution: The Regions and the autonomous provinces of Trento and Bolzano participate in preparatory decision-making process of EU legislative acts in the areas that fall within their competences. They are also responsible for the implementation of international agreements and EU measures, subject to the rules set out in State law which regulate the exercise of subsidiary powers by the State in the case of non-performance by the Regions and autonomous provinces.

There has been a huge debate in Italy about the relationship between these two constitutional provisions.39 Though scholars agree that both provisions bestow the State with substitutive powers, it is not clear how these provisions should be co-ordinated. According to some authors, Article 120 of the Constitution provides the genus of substitution powers, whereas Article 117 refers to one species of the genus, being a mere specification of Article 120. According to other scholars, however, Article 120 introduces a form of administrative substitution (which explains why the provision speaks of the ‘Government’), while Article 117—dealing with the distribution of legislative powers—concerns legislative substitution (which explains why it mentions ‘the State’ rather than the Government). Although these provisions are the first constitutionalisation of the State’s substitution power, earlier traces may be found in the legislative regime regarding the municipalities.40 This confirms that the power has its origin in the instruments for controlling municipal action.41 Another codification of the substitution power—with regard to the Italian Regions—may be found in the fundamental charters (Statuti) of certain Special Regions42 and their implementing measures.43 To be clear, these provisions provide forms of ‘administrative’ substitution powers, ie forms of substitution between (State and regional) administrations. The 39 For an overview of the debate, see C Mainardis, Poteri sostitutivi statali e autonomia amministrativa regionale (Milan, Giuffrè, 2007); C Mainardis, ‘I poteri sostitutivi statali: una riforma costituzionale con (poche) luci e (molte) ombre’ (2001) Le Regioni 1369; EC Raffiotta, Gli interventi sostitutivi nei confronti degli enti territoriali, PhD thesis (Bologna, University of Bologna, 2008) available at , 66 ff; L Buffoni, La metamorfosi della funzione di controllo nella Repubblica delle Autonomie. Saggio critico sull’art. 120, comma II, della Costituzione (Torino, Giappichelli, 2007). 40 Testo Unico 1898, Art 193; Testo Unico 1889, Arts 169 and 174; Testo Unico 1908, Art 210; Testo Unico 1915, Art 216. See generally Raffiotta, above n 39, 13 ff. 41 On the relationship between the activity of control and the substitutive power, see FG Scoca, ‘Potere sostitutivo e attività amministrativa di controllo’ in AAVV, Aspetti e problemi dell’esercizio del potere di sostituzione nei confronti dell’amministrazione locale (Milan, Giuffrè, 1983) 17 ff; A de Michele, Il potere sostitutivo dello Stato e delle Regioni nel Titolo V, Parte II, della Costituzione, nella legislazione attuativa e nella giurisprudenza, PhD thesis (Bologna, University of Bologna, 2007) available at . 42 In Italy there are two ‘kinds’ of Regions: 15 ‘ordinary’ Regions and 5 ‘Special’ Regions. The latter category includes Friuli-Venezia Giulia, Sardinia, Sicily, Trentino-Alto Adige/Südtirol, and the Aosta Valley (see Art 116(1) of the Italian Constitution). There is an ongoing debate in Italy about whether the 5 latter Regions are still ‘Special’ now that the ordinary Regions have also acquired exclusive legislative powers in the 2001 constitutional reform. 43 See Statuto of Sardinia, Art 50; Presidential Decree No 480/1975, Art 33 concerning Sardinia; Law No 196/1978, Art 39; Presidential Decree No 182/1982, Art 2.

Impact of ‘Regional Blindness’ on Italian Regional State 371 substitution power laid down in Articles 117 and 120 of the Italian Constitution, by contrast, seems to cover ‘legislative’ forms as well.44 The conditions for the exercise of the State’s substitution power, as well as the procedures to be followed, have been laid down in specific legislation.45 The relevant laws basically provide that the President of the Italian Council of Ministers or the Minister for Community Policy gives the Region a time limit within which to comply, and may also ask for the submission of the question to the permanent Conference on relations between the State, the Regions and the Autonomous Provinces of Trento and Bolzano/Bozen in order to arrange for the necessary initiatives to be taken. In the event of inertia by the relevant territorial actors, the President of the Council of Ministers or the Minister for the Community Policies may propose to the Council of Ministers the initiatives to be taken for the exercise of the substitution power.46 ii. A Priori Substitution From the theoretical point of view, the idea of substitution seems to imply State intervention subsequent to regional inactivity. However, on closer examination, the Italian legal system has also been characterised by a form of ‘preventive’ substitution. Thus, Article 9 of Law no 86/89 established that a State act is necessary for the implementation of EU directives. Such a national act consists of two kinds of provisions. Whereas a first group comprises the fundamental principles that cannot be waived by the regional legislature, the second group contains more specific provisions from which Regions may deviate (disposizioni di dettaglio). According to this scheme, even in the case of regional inactivity, the requirements of EU law would be met thanks to the detailed provisions included in the national law. However, if a Region does pass specific legislation, the power to issue detailed (ie, non-fundamental) implementation measures goes back to the regional legislature. This mechanism is called cedevolezza.47 It has been a recurrent theme of Italian regionalism since 1970, the year in which the ordinary Regions came into existence. Cedevolezza applies to the field of shared competences. It may be seen as a technique devised by the Italian State legislator (and, over the years, supported by the Italian Constitutional Court) in order to avoid the danger of legislative gaps. In order to remove any possible horror vacui, the State legislator has given itself the power to adopt ‘temporary’ detailed provisions that intrude upon the Regions’ legislative competences. These provisions have been sanctioned by the 44 However, the Italian Constitutional Court has recently complicated the picture. See sentenza No 361/2010, 13 December 2010. 45 Law No 131/2003 (legge La Loggia); Law No 11/2005 (legge Buttiglione). See generally G Scaccia, ‘Il potere di sostituzione in via normativa nella legge n 131 del 2003. Prime note’, available at . 46 Law No 131/2003, Art 8; Law No 11/2005, Art 10(3). 47 From the Italian verb cedere, to cede. On cedevolezza, see M Motroni, Le norme cedevoli nel rapporto tra fonti statali e fonti regionali, PhD thesis (Ferrara, University of Ferrara, 2010) available at .

372 Giuseppe Martinico Italian Constitutional Court48 because of their ‘temporary’ and ‘supplementary’ nature, that is, since they are supposed to disappear once the Regions have exercised their legislative competence. One might connect the emergence of cedevolezza to the necessity of guaranteeing legislative continuity. Since the birth of the Italian (ordinary) Regions implied the transfer of competences from the central to the sub-State level, matters which were formerly regulated by State laws suddenly came within the ambit of the regional powers. Considering these old State acts as automatically invalid would have entailed legislative gaps during the passing of new regional legislation. According to this logic, cedevolezza is justified by the emergence of a new constitutional subject (ie, the Regions) and the need to maintain stability in a period of transition.49 The Constitutional Court followed the same logic soon after the constitutional reform of 2001 which, as we know, changed the distribution of competences between the ordinary Regions and the State.50 The technique of cedevolezza is in line with a conception of the constitutional principle of conferred powers as an indication of legislative ‘preference’ rather than a legislative ‘reserved domain’.51 According to this conception, the Italian Constitution, when granting a portion of shared competence to the regional legislator, expressed a preference for this actor while at the same time allowing for temporary and derogable State action in order to ensure legal continuity. In reality, however, cedevolezza has permitted the centralisation of powers and, thus, a de facto alteration of the constitutional distribution of competences. This is especially so since the technique has frequently been resorted to in a context where ‘unitary principles’ were being invoked.52 Even after the 2001 constitutional reform, and despite what the Italian Constitutional Court seems to have suggested in some decisions,53 cedevolezza appears to be still alive. One may wonder whether the new Article 117 of the Constitution, which expressly codifies the Regions’ competence to implement EU law requirements, may represent the end of the technique of cedevolezza in this specific regard. Despite the many doubts raised by scholars,54 this particular kind of cedevolezza has been consolidated by the passing of various ‘Community acts’ or leggi comunitarie after the constitutional reform.55 According to this trend, the State legislature is allowed to adopt detailed provisions in case of regional inertia. 48 Corte Costituzionale, sentenza No 214/1985, 11 July 1985. Cp, sentenza No 40/1972, 24 February 1972. See also sentenze No 192/1987, 21 May 1987 and No 729/1988, 20 June 1988. For a complete overview in Italian, see Motroni, above n 47, 63–94. 49 Motroni, above n 47, 63–94. 50 See, eg, Corte Costituzionale, sentenza No 282/2002, 19 June 2002; sentenza No 176/2010, 10 May 2010; sentenza No 234/2010, 23 June 2010. 51 L Carlassare, ‘La “preferenza” come regola dei rapporti tra fonti statali e regionali nella potestà legislativa ripartita’ (1986) Le Regioni 236. 52 See, amongst others, A D’Atena, Le Regioni dopo il big-bang. Il viaggio continua (Milano, Giuffrè, 2005) 76 ff. 53 See, eg, Corte Costituzionale, sentenza No 282/2002, 19 June 2002. 54 Motroni, above n 47, 95 ff. 55 See, eg, Laws No 39/2002, No 14/2003, and No 306/2003.

Impact of ‘Regional Blindness’ on Italian Regional State 373 The national legislation comes into force only if (and as soon as) the deadline for the implementation of the directive has passed.56 Something similar may be found in Law no 11/2005, which also allows the State to exercise its substitution power through an administrative act (regolamento), even in matters covered by regional legislation,57 including areas of exclusive regional competence. In conclusion, despite the change to the letter of the Constitution, the Italian Constitutional Court and the Parliament too have confirmed many of the principles and practices characteristic of old-style Italian regionalism. Before moving on to the next sub-section, it is interesting to point out that the Italian Constitutional Court recently acknowledged cedevolezza’s bilateral nature by confirming the validity of a law of the Friuli-Venezia Giulia Region.58 The regional legislation included some (derogable and supplementary) fundamental principles, which were adopted in order to prevent the absence of a relevant State act resulting in the incorrect implementation of the directive. Thus, the Region of Friuli-Venezia Giulia implemented Directive 2001/42/EC without waiting for the national legislature to lay down the fundamental principles of the matter. The fundamental principles adopted by the Region were considered to be derogable once the State legislature had passed a framework law.59 The Italian Constitutional Court approved of the regional legislation.60 To conclude this part on the mechanisms used by the Italian State in order to deal with the risk of non-compliance due to regional inertia, it should be recalled that the State may now ask the Regions for the reimbursement of any costs connected with their non-compliance with the requirements of EU law.61

C. The Principle of Subsidiarity Subsidiarity might be regarded as an example of a principle that was originally alien to the Italian Constitution and was introduced under pressure of EU law. As defined in the Oxford English Dictionary, ‘subsidiarity’ signifies ‘that a central authority should have a subsidiary function, performing only those tasks which cannot be performed effectively at a more immediate or local level’.62 This implies

56 See GU Rescigno, ‘Attuazione regionale delle direttive e potere sostitutivo dello Stato’ (2002) Le Regioni 729. 57 Law No 11/2005, Art 11(8). According to some authors, it is possible to find a basis for this practice in the case law of the Constitutional Court, more precisely in sentenza No 425/99, 27 October 1999. See Motroni, above n 47, 132. 58 Regional Law No 11/2005. 59 Regional Law No 11/2005, Art 12. 60 Corte Costituzionale, sentenza No 398/2006, 20 November 2006. See also S Tripodi, ‘L’attuazione regionale delle direttive comunitarie e le clausole di cedevolezza. Nota a Corte cost, sent n 398/2006’, available at ; Motroni, above n 47, 134–37. 61 Law no 11/2005, Art 16bis. See also Bilancia et al, above n 15, E-167. 62 Oxford English Dictionary (Oxford, Oxford University Press), entries ‘subsidiary’ and ‘subsidiarity’.

374 Giuseppe Martinico both a constitutional preference for the exercise of a competence by the periphery and a ‘constitutional restraint on the exercise of those competences granted to the centre’.63 The subsidiarity principle is enshrined in the Treaties64 and in the Protocol on the Application of the Principles of Subsidiarity and Proportionality.65 Its application requires a system of divided competences, while presupposing at the same time an ‘integrated’ system such as a co-operative federation. As a matter of fact, the principle refers to a relationship between two institutional actors (a lower actor or the periphery, and a higher actor or the centre) sharing the same power. At first instance, that power is preferably exercised by the actor which is closer to the citizens. Scholars usually label this first phase as the ‘negative’ side of subsidiarity, since it imposes a duty of non-intervention on the centre. At the same time, however, the subsidiarity principle allows for the possibility to substitute the relevant local actor if the power in question may be exercised in a better or more efficient way by the higher level. This way, subsidiarity works as an ‘elevator’ with regard to powers that could, in theory, be exercised by two institutional subjects. If the preferred lower actor cannot adequately perform the requested action, it will be substituted by the higher actor.66 Having recalled the theoretical notion of subsidiarity, it is interesting to see that the principle was also slipped into the Italian Constitution on the occasion of the 2001 reform. The principle of subsidiarity is enshrined in Article 118 of the Constitution: (1) Administrative functions are attributed to the Municipalities, unless they are attributed to the provinces, metropolitan cities and regions or to the State, pursuant to the principles of subsidiarity, differentiation and proportionality, to ensure their uniform implementation. (2) Municipalities, provinces and metropolitan cities carry out administrative functions of their own as well as the functions assigned to them by State or by regional legislation, according to their respective competences. (3) State legislation shall provide for co-ordinated action between the State and the Regions in the subject matters as per Article 117, paragraph two, letters b) and h), and also provide for agreements and co-ordinated action in the field of cultural heritage preservation. (4) The State, regions, metropolitan cities, provinces and municipalities shall promote the autonomous initiatives of citizens, both as individuals and as members of associations, relating to activities of general interest, on the basis of the principle of subsidiarity.

63 R Schütze, From Dual to Cooperative Federalism. The Changing Structure of European Law (Oxford, Oxford University Press, 2009) 246. 64 Arts 5 and 12 TEU; Arts 69 and 352 TFEU. 65 Protocol no 2 on the Application of the Principles of Subsidiarity and Proportionality [2010] OJ C83/206. 66 R Bin and G Pitruzzella Diritto costituzionale (Torino, Giappichelli, 2001).

Impact of ‘Regional Blindness’ on Italian Regional State 375 In sentenza No 303/2003, the Italian Constitutional Court used the subsidiarity principle as a Trojan Horse to reshape the distribution of competences codified in the Constitution. The Court ruled that the national legislature is allowed to exercise administrative functions in matters falling within the regional (concurring) legislative powers if uniform exercise of these administrative functions is required. It is clear that the Court has expanded the original scope of application of the subsidiarity principle by inventing a mechanism analogous to the idea of ‘national interest’,67 which was present in the wording of the old Title V of the Constitution but was discarded in the 2001 reform. According to many authors, the ‘national interest’ doctrine still functions as an unwritten limit to the Regions’ powers.68 To conclude this section, it is useful to recall the Italian Constitutional Court’s view of the relationship between subsidiarity and co-operation. According to the Court, the subsidiarity principle requires the ‘loyal co-operation’ (leale collaborazione) of the territorial actors, concerted practices and bodies, and a system of agreements between the institutional actors. In sentenza No 303/2003, the Italian Constitutional Court established a set of conditions under which derogation from the constitutional allocation of competences is permissible. More precisely, the derogation must be: (a) proportionate to the public interest that justifies the exercise of regional functions by the State; (b) not unreasonable in the face of strict constitutional scrutiny; and (c) the result of an agreement with the region.69 It has been claimed that the Court adopted in this judgment ‘a procedural and consensual approach to the principles of subsidiarity and adequacy,’ and that the Court ‘denied that such principles can operate as mere verbal formulas capable of modifying, to the advantage of national law, the allocation of legislative powers established by the Constitution’.70 The Italian subsidiarity episode described here tells us at least two things. First, it confirms the centrality of the Constitutional Court in the destiny of Italian regionalism. Secondly, it makes it more difficult to trace the Italian constitutional principle of subsidiarity back to the EU (or German) legal system. Whereas the ECJ traditionally has avoided engaging in debates concerning subsidiarity, the Italian Constitutional Court has actively made use of subsidiarity to reshape

67 On the national interest (interesse nazionale), see A Barbera, Regioni e interesse nazionale (Milan, Giuffrè, 1973). 68 For an account of this debate, see R Bin, ‘L’interesse nazionale dopo la riforma: continuità dei problemi, discontinuità della giurisprudenza costituzionale’, available at ; A Barbera, ‘Chi è il custode dell’interesse nazionale?’ (2001) Quaderni Costituzionali 345; T Groppi and N Scattone, ‘Italy: The Subsidiarity Principle’ (2006) 4 International Journal of Constitutional Law 131, 134. 69 Corte Costituzionale, Sentenza No 303/2003, 25 September 2003. 70 Groppi and Scattone, above n 67, 135. See also A Morrone, ‘La Corte costituzionale riscrive il Titolo V?’, available at .

376 Giuseppe Martinico Italian regionalism. Admittedly, despite this difference in approach, the two courts have reached analogous results: subsidiarity has worked as a factor of centralisation in the EU as well as the Italian legal order. This is striking, given that the original meaning and goal of subsidiarity was to favour the decentralised exercise of shared powers.

III. PROCEDURAL NORMS AND TECHNIQUES CONCERNING THE ITALIAN CONSTITUTIONAL COURT

This section deals with the second group of norms and techniques mentioned at the beginning of the chapter. These legal instruments relate to the proceedings before the Italian Constitutional Court, which serves as the final arbiter of conflicts between the Regions and the State. Though constitutional proceedings can be very technical, I do not intend to go into procedural details here. On the contrary, my aim is to frame these proceedings in a general perspective, building on the idea of ‘procedural impermeability’, that is, the reluctance of the Italian Constitutional Court to get involved in matters of EU law.

A. Convergence and Divergence between the Italian Constitutional Court and the ECJ The procedural impermeability of the Italian Constitutional Court’s case law finds its origins in the historical rivalry between the Italian Constitutional Court and the ECJ. As is well-known, the ECJ and the Italian Constitutional Court started from opposing positions of monism (ECJ) and dualism (Italian Constitutional Court). Later on, however, the Constitutional Court began to talk about two ‘autonomous and separated, although co-ordinated’ systems.71 At the same time, the ECJ showed its appreciation of the national actors’ efforts by occasionally taking a benign and tolerant attitude. Some scholars have defined this situation of partial convergence as a ‘(limited) flexibilisation of supremacies’.72 Characteristic of the above-sketched jurisdictional convergence is the ‘mutability’ of the Italian Constitutional Court’s position. It is possible, in fact, to observe a strong evolution in the Italian Court’s case law. Initially, the Court interpreted the relationship between national and Community acts in the light of chronological criteria, taking as a basic premise that the national act enabling the ratification 71

Corte Costituzionale, sentenza No 170/1984, Granital, 8 June 1984. V Ferreres Comella, ‘La Constitución española ante la clausola de primacia del Derecho de la Unión europea. Un comentario a la Declaración 1/2004 del Tribunal Constitucional 1/2004’ in A Lopez Castillo, A Saiz Arnaiz and V Ferreres Comella (eds), Constitución española y constitución europea (Madrid, CEPC, 2005) 80–89. However, despite this convergence, the tension between the two actors has not disappeared, partly due to the progressive expansion of the ECJ’s activity in areas of national autonomy. 72

Impact of ‘Regional Blindness’ on Italian Regional State 377 of the Treaties was an ordinary law.73 Later on, in sentenza No 183/1973, the Court changed its position, stating that the constitutional basis of the primacy of Community law was to be found in Article 11 of the Italian Constitution.74 Interestingly, this provision envisioned Italy’s participation in the United Nations and other international peacekeeping organisations, but not in the EU. The latter organisation in fact imposes limitations on national sovereignty to achieve ends going beyond ‘peace and justice between nations’. Consequently, the Italian Constitutional Court had to ‘manipulate’ the original meaning of Article 11 in order to allow for such limitations.75 In the beginning, the Italian Court entrusted itself with the task of monitoring respect for the primacy of Community law by regarding a national law in breach of Community law as a violation of Article 11 of the Constitution.76 However, due to the ECJ’s Simmenthal judgment, the Constitutional Court had to change its view. In its judgment in Granital, it decided to pass the task to the ordinary domestic courts.77 Yet even after the Granital judgment, the contrasts between the Courts did not disappear. Though several factors may account for this, one reason is that the ECJ has obtained the confidence of the ordinary judges. As a consequence, the relationship between ordinary judges and the Constitutional Court began to deteriorate. When ordinary courts seek clarification of the relationship between EU law and the Italian Constitution, they do not refer to the Constitutional Court but, rather, to the ECJ. This is the result of the self-exile of the Constitutional Court, which has rarely been willing to decide questions concerning the relationship between legal orders. In 2002, for instance, the Constitutional Court ruled on just 10 (out of 500) cases relating to the Community legal order.78 One might therefore describe the Constitutional Court’s case law as ‘impermeable’ to EU law.

B. Centralised Review of the Compatibility of Regional Legislation with EU Law: Exceptions to ‘Procedural Impermeability’ Although the Italian Constitutional Court’s case law is, as a general rule, impermeable to EU law, there are some exceptions to be highlighted, most notably as regards principaliter proceedings, the arena of the legislative conflicts between the State and the Regions. In Italy, the constitutional review of legislation may be

73

Corte Costituzionale, sentenza No 14/1964, 24 February 1964. Corte Costituzionale, sentenza No 183/1973, 18 December 1973. 75 Since the 2001 constitutional reform, an explicit reference to the European legal order is contained in Art 117 of the Constitution. 76 Note that the Constitutional Court found that it could not control the compatibility of Community law with the Italian Constitution, because it could only decide on the validity of (ordinary) Italian laws. 77 See n 6 above. 78 M Cartabia and A Celotto, ‘La giustizia costituzionale in Italia dopo la Carta di Nizza’ (2002) Giurisprudenza Costituzionale 4477. 74

378 Giuseppe Martinico triggered and pursued in two different ways, that is, either indirectly (incidenter) during a trial before an ordinary judge, or directly (principaliter) by the State or a Region. Whereas the State Government may appeal directly against a regional law, a Region may do so against a national law or a law enacted by another Region.79 In such cases, constitutional proceedings are conceived to resolve disputes between the State and the Regions concerning the limits of their respective powers. Importantly, a huge difference exists between the’grounds on which proceedings may be brought by the Regions and the State’ available to the Regions and the State. Whilst the former may raise the constitutional question only when their sphere of competence is infringed upon, the State may challenge a regional law on the basis of any kind of constitutional ‘defect’. Principaliter proceedings represent an exception to the diffuse review of the compatibility of domestic law with Community law. The Italian Constitutional Court has agreed to assess the consistency of domestic and Community law within this type of proceeding. In sentenze No 384/1994 and No 94/1995,80 the Court made clear that it is willing to do so if the question is raised in direct proceedings (whether it be by the Regions or by the State). The rationale behind this decision may be found in the particular dynamics of direct proceedings. In that type of case, the ordinary judge, who normally monitors respect for EU law, has no role to play. Were the Constitutional Court to refuse to rule on EU lawrelated questions in principaliter proceedings, a dangerous gap would occur in the protection of EU law. It may be said, therefore, that the residual role of the Constitutional Court in direct proceedings is justified only because the ordinary judge, who is the natural guardian of the primacy of EU law at the domestic level, is missing from the scene. This change in the law governing Constitutional Court proceedings is a significant example of the ‘adaptation’, under the influence of EU law, of a mechanism that was originally devised to guarantee the equilibrium between the regional and national legislatures. In order to ensure respect for EU law, the Italian Constitutional Court agreed to soften its impermeability doctrine. Another exception to the impermeability of the Constitutional Court’s case law occurs when Italian law conflicts with norms of EU law which are not directly effective. Since the Simmenthal mandate of the ordinary courts does not apply in that event,81 the Italian Constitutional Court is the only guardian of the primacy of the supranational system. Lastly, it is interesting to note that the Italian Constitutional Court employs a particular mode of adjudication in cases where the compatibility of Italian law with EU law is at issue. More precisely, the Court regards provisions of EU law as ‘interposed norms’ or ‘interposed standards’ (norme interposte). The idea of the 79 80

Italian Constitution, Art 127. Corte Costituzionale, sentenza No 384/1994, 7 November 1994; sentenza No 94/1995, 20 March

1995. 81

See n 7 above.

Impact of ‘Regional Blindness’ on Italian Regional State 379 interposed norm was first formulated in the 1950s by Carlo Lavagna, an Italian scholar.82 Normally, the standards of constitutional review are drawn from the Constitution itself, constitutional legislation or sources considered equivalent, such as the generally recognised principles of international law. In some cases, however, the Court may decide to broaden this normative spectrum and use norms that have, strictly speaking, a lower status as standards of review. This happens, for instance, when a constitutional provision refers to such norms. If that is the case, ‘a violation of an “interposed norm” is tantamount to a violation of the authorising norm itself ’.83 The most famous example of such interposed norms is the legislation that enables delegated legislation under Article 76 of the Italian Constitution.84 Over the years, and especially after the constitutional reform of 2001, the Italian Constitutional Court has extended the concept of norme interposte so as to be able to treat a violation of EU law as an indirect violation of Article 117 of the Constitution.85 In effect, EU norms fill the blank formula of Article 117(1) of the Constitution with a ‘substantive instruction’. Consequently, if a regional legislative act is found to be in violation of EU law, the act will be declared unconstitutional.86

IV. FINAL REMARKS

Some authors have described EU integration as one of the most important factors leading to the Italian constitutional reform of 2001.87 This is true and, in itself, answers the research question formulated at the beginning of this chapter. All Italian institutional actors, the judiciary too, have adjusted existing norms and mechanisms, including the Constitution itself, to guarantee the effectiveness of EU law. As we have seen, the Italian Constitutional Court has played a fundamental role, both in the history of Italian regionalism and in the European path of the Regions. The Court has used its authority sometimes to support the centralisation of competences, on other occasions to preserve regional autonomy. Currently, the importance of co-operative instruments and principles of subsidiarity and loyal co-operation seems to have been rediscovered. At least 82 C Lavagna, Problemi di giustizia costituzionale sotto il profilo della non manifesta infondatezza (Milan, Giuffrè 1957) 28 ff. 83 A Pizzorusso, V Vigoriti and GL Certomà, ‘The Constitutional Review of Legislation in Italy’ (1983) 56 Temple Law Quarterly 503, 593. 84 Art 76 determines: ‘The exercise of the legislative function may not be delegated to the Government unless principles and criteria have been established by Parliament, and only for a limited period and for specified purposes.’ 85 The European Convention on Human Rights is now treated the same way as EU law. See F Biondi Dal Monte and F Fontanelli, ‘The Decisions No 348 and 349/2007 of the Italian Constitutional Court: The Efficacy of the European Convention in the Italian Legal System’ (2008) 9 German Law Journal 889. 86 See, eg, Corte Costituzionale, sentenza No 406/2005, 24 October 2005; sentenza No 129/2006, 23 March 2006. 87 Bilancia et al, above n 15, E-144.

380 Giuseppe Martinico in theory, these instruments and principles should function as a sort of shield against the alteration of the constitutional lists of competences. In practice, however, the Italian Constitutional Court has proved willing to interpret the constitutional principles of conferral and subsidiarity in a broad and flexible way in the face of EU law. It is therefore my impression that external (ie, international and European) factors do have a decisive impact on the never-ending balancing act between the competing interests of the Italian State and the Regions.

16 The Spanish State Structure and EU Law: The View of the Spanish Constitutional Court MAITE ZELAIA GARAGARZA*

I. INTRODUCTION

T

HE REGIONALISED STRUCTURE of Spain, known as the Estado de las Autonomías or ‘State of the Autonomies’, was established by the Spanish Constitution of 1978 and further developed by the Statutes of Autonomy. The Spanish Constitution defines the basic principles of the Estado de las Autonomías: unity and self-government or autonomy.1 On the one hand, the principle of unity pertains to the unity of the structural elements constitutive of the State: one territory, one Spanish people, one legal system, one judiciary and one Constitutional Court. Moreover, the principle of unity refers to the unity of the State itself: a single State in international law, a single Spanish nationality, a single external representation of the State and a unitary military organisation.

* Professor of Constitutional Law, University of the Basque Country. Research for this paper was carried out in the framework of the research project DER 2011-25795 supported by the Spanish Ministry of Science and Innovation. The author is also a member of the Consolidated Research Group on Fundamental Rights in the EU (GIC07/86-IT-448-07) and the Education and Research Unit UFI 11/05. 1 Spanish Constitution, Art 2: ‘The Constitution is based on the indissoluble unity of the Spanish nation, the common and indivisible homeland of all Spaniards; it recognises and guarantees the right to self-government of the nationalities and regions of which it is composed and the solidarity among them all.’ These two principles (unity and autonomy) are at the same level in the hierarchy of norms as the principle of solidarity. At a lower level, one finds constitutional principles such as the constitutional autonomy of the Autonomous Communities, economic unity, co-operation, and the equality of rights and obligations across the State. Furthermore, Art 1(2) of the Spanish Constitution refers to sovereignty: ‘National sovereignty belongs to the Spanish people, from whom all state powers emanate.’ For an overview of the relevant literature, see S Muñoz Machado, Derecho Público de las Comunidades Autónomas (Madrid, Iustel, 2007).

382 Maite Zelaia Garagarza The principle of self-government or autonomy, on the other hand, implies the existence of distinct governance levels. More precisely, the Spanish Constitution mentions three levels: (a) the State, to which the Spanish people have transferred sovereignty; (b) the Autonomous Communities, vested with political autonomy (ie, autonomous legislative and executive powers in certain constitutional areas); and (c) the provinces and municipalities, which have acquired certain local, basically administrative, powers.2 How do the constitutional principles of unity and autonomy relate to each other? Are they compatible? In a series of judgments issued in the 1980s and 1990s,3 the Spanish Constitutional Court confirmed these principles’ compatibility, while at the same time insisting that autonomy must not destroy unity. Hence, in order to be constitutional, the autonomous powers must be appropriated and exercised with respect for the principle of unity and the constitutional division of competences between the central and the autonomous institutions.4 What is more, the principles of unity and autonomy have no meaning independently of one another: there is no unity without autonomy, and vice versa.5 That being said, the process of European integration entailed a transfer of Member State powers to the EU institutions. This transfer sometimes took place directly, through the conferral of powers in the Treaties. In other cases, however, the transfer of powers was more indirect, especially in those instances where the Treaties did not provide for the requisite powers while EU action was necessary to obtain the objectives set out in the Treaties.6 In the case of Spain, being a ‘composed’ State, the transfer of powers to the EU involves both central7 and regional competences.8 Thus, the process of European integration resulted in the Europeanisation of numerous State as well as regional 2 They do not have a proper political autonomy, with the exception of the Basque provinces, which do possess legislative powers, especially in the field of taxation. See generally, TR Fernández Rodríguez, Los derechos históricos de los territorios forales (Madrid, Centro de Estudios Constitucionales, 1985); M Pulido Quecedo, ‘Normas forales vascas’ (2010) 798 Actualidad Jurídica Aranzadi. 3 See especially, Spanish Constitutional Court, No 32/1981, 28 July 1981; No 5/1982, 8 February 1982; No 158/1986, 11 December 1986; No 135/1992, 5 October 1992. 4 Spanish Constitutional Court, No 4/1981, 2 February 1981. 5 See especially, Spanish Constitutional Court, No 35/1982, 14 June 1982, and more recently No 247/2007, 12 December 2007. 6 Art 352 TFEU provides that the Council may adopt the appropriate measures, acting unanimously on a proposal from the Commission and after obtaining the consent of the European Parliament. Further on Art 352 TFEU (ex Art 308 TEC), see D Chalmers, G Davies and G Monti, European Union Law, 2nd edn (Cambridge, Cambridge University Press, 2010) 214–19; G de Búrca and B De Witte, ‘The Delimitation of Powers between the EU and its Member States’ in A Arnull and D Wincott (eds), Accountability and Legitimacy in the European Union (Oxford, Oxford University Press, 2001) 217. 7 Eg, nationality, immigration and asylum (Spanish Constitution, Art 149(1)2); the judiciary (Spanish Constitution, Art 149(1)5); external trade policy (Spanish Constitution, Art 149(1)10); monetary policy (Spanish Constitution, Art 149(1)11). 8 Eg, the organisation of territory and urbanism (Spanish Constitution, Art 148(1)3); agriculture (Spanish Constitution, Art 148(1)7); welfare (Spanish Constitution, Art 148(1)20).

The Spanish State Structure and EU Law 383 powers. Additionally, as regards the Autonomous Communities’ powers, another centralising tendency was generated by the fact that the institution habitually participating in the Council of the EU is the State government.9 As a result, the constitutional system of division of competences and the balance on which it is based have become unsettled. This chapter explores two possible ways to counterbalance the loss of regional autonomy due to the operation of EU law. It does so through an analysis of two types of cases that have arisen before the Spanish Constitutional Court involving international relations and the implementation of EU law, respectively.

II. PARTICIPATION OF AUTONOMOUS COMMUNITIES IN EUROPEAN AFFAIRS

This section examines the transformation of the constitutional system of power division in the field of international relations.10 The late 1980s saw an intense debate about the difference between ‘international relations’ and the ‘relations with the EU’. Article 149(1)3 of the Spanish Constitution provides that ‘international relations’ are the exclusive domain of the State. From the beginning, this constitutional provision created tension between the State and the Autonomous Communities, which in turn gave rise to case law from the Spanish Constitutional Court, which was strongly criticised in certain Spanish academic circles.11 In its case law, the Court relied on the literal meaning of the constitutional text, pointing out that Article 149(1) of the Constitution used diverse techniques to delimit the field reserved exclusively to the State. It highlighted, however, that there is a pertinent difference between the exclusivity of all the relevant competences (International Relations, Defence and Armed Forces, Justice Administration, the Treasury and State Debt, etc) and the exclusivity of concrete powers (whether this pertains to basic legislation or all legislation in specific matters). In the former case, the exclusive competence of the State not only prevents an Autonomous Community from assuming competences in a matter reserved to the State, but also stops it from developing a policy concerning the matter within the framework of its self-government. However, it may participate in the determination of the said policy through the specific representation that the different Communities have in the Senate.12 According to the Constitutional Court, that position was supported by Article 97 of the Spanish Constitution, which establishes that foreign

9 For a fuller analysis, see P Pérez Tremps, Constitución Española y Comunidad Europea (Madrid, Civitas, 1993) 47–56. 10 See Spanish Constitutional Court, No 165/1994, 26 May 1994. 11 For an analysis of Art 149(1)3, see especially R Bustos Gisbert, Relaciones internacionales y Comunidades Autónomas (Madrid, CEC, 1996); F Fernández Segado, ‘Estado, Comunidades Autónomas y acción exterior en el ordenamiento constitucional español’ (2004) Foro. Nueva Época 91. 12 Spanish Constitutional Court, No 35/1982, 14 June 1982.

384 Maite Zelaia Garagarza policy must be conducted by the State Government, and which clearly excludes autonomous entities.13 Contrary to this constitutional jurisprudence, a significant part of legal academia argued against treating international relations as a matter for exclusive state competence. It was held that the relevant case law produces restrictive effects for the Autonomous Communities by denying them the opportunity to project their will onto the international plane without the mediation of the central institutions of the State. In fact, any of the matters included in the list of Article 148 of the Spanish Constitution, and therefore amenable to action by the Autonomous Communities, could be projected externally. It was advocated that a ‘hard core’ should be identified at the heart of international relations, over which the State monopoly would be maintained, making it possible for the remainder of those fields of competence to be amenable to autonomous activity.14 The Constitutional Court began to modify its case law in the 1990s, thereby gradually approaching the academic position just outlined. Judgment 165/199415 is regarded as the leading case in that regard. However, as early as judgment 153/198916 the Constitutional Court had endorsed the position that there is no constitutional impediment to the Autonomous Communities developing activities that have an international dimension.17 To be sure, Judgment 165/1994 reserves an exclusive competence to the State in classic areas of international policy, such as the external representation of the State (Articles 56.1 and 63.1 of the Spanish Constitution), the ius ad tractatum (Articles 93 and 94 of the Spanish Constitution), the responsibility of the State, the determination of the Spanish position in external matters (Article 97 of the Spanish Constitution), as well as the other issues listed in Article 149(1) with an external dimension (eg, nationality, immigration, emigration, foreign affairs and the right to asylum, customs and tariff regime, external trade, monetary system, external health measures, air transport, air space control

13 Spanish Constitutional Court, No 44/1982, 8 July 1982; No 154/1985, 12 November 1985; No 137/1989, 20 July 1989. See C Conde Martínez, La acción exterior de las Comunidades Autónomas. La institucionalización de gobiernos territoriales y la integración internacional (Madrid, Tecnos, 2000) 61. 14 A Remiro Brotons, ‘La actividad exterior del Estado y las Comunidades Autónomas’ in M Ramírez (ed), Estudios sobre la Constitución Española de 1978 (Zaragoza, Pórtico, 1979) 353; A Remiro Brotons, La acción exterior del Estado (Madrid, Tecnos, 1994); A Remiro Brotons, El Estado, el Derecho interno y la Comunidad Europea (Madrid, Civitas, 1985); A Mangas Martín, ‘Cuestiones de Derecho Internacional Público en la Constitución Española de 1978’ (1980) 61 Revista de la Facultad de Derecho de la Universidad Complutense de Madrid 143; G Jáuregui Bereciartu, Las Comunidades Autónomas y las relaciones internacionales (Oñati, IVAP, 1986); JM Baño León, Las Comunidades Autónomas en la Comunidad Europea (Valencia, IVAP, 1987); F Morata, Autonomía regional e integración europea (Barcelona, IEA, 1987). 15 Spanish Constitutional Court, No 165/1994, 26 May 1994. 16 Spanish Constitutional Court, No 153/1989, 5 October 1989. 17 See O Casanovas i La Rosa, ‘La acción exterior de las Comunidades Autónomas y su participación en la celebración de los Tratados internacionales’ in M Pérez González (ed), La acción exterior y comunitaria de los Länder, Regiones, Cantones y Comunidades Autónomas (Vitoria-Gasteiz, IVAP, 1994) 56–58.

The Spanish State Structure and EU Law 385 and aircraft registration).18 In short, international relations reserved for the State are relations between international legal persons and governed by international law.19 Regardless of this ‘hard core’,20 though, the Autonomous Communities can carry out, and in fact do so in practice,21 activities that have an international effect if and when these activities are necessary for the execution of their powers, and if they have the corresponding statutory competence.22 The discussion of the external competences of the Autonomous Communities was effectively brought to a close in 1994, when the Constitutional Court cleared the way for a Basque Office in Brussels.23 On the basis of the opening created by the case law of the Constitutional Court,24 the reforms of the Statutes of Autonomy have created a distinction between international relations reserved for the State and autonomous action with a foreign dimension, which is amenable to regulation by the Autonomous Communities. At the root of the latter is the autonomous activity in European affairs and foreign action of a broader nature. This has been one of the most significant aspects in the area of the relations between the EU and the Autonomous Communities.

III. IMPLEMENTATION OF EU LAW BY AUTONOMOUS COMMUNITIES

The starting point of our enquiry is that EU law, at first sight, is neutral towards the regionalised structure of the Member States. The EU law principles of institutional autonomy, equality between the Member States, and respect for national identities and local and regional autonomy25 support this view.26 The position of the European Court of Justice in this regard is clear: ‘Each Member State is free to delegate powers to its domestic authorities as it considers fit and to implement the European norm by means of measures adopted by regional or local

18 Art 149(1) paras 2, 10, 11, 16 and 20. See J Martín y Pérez de Nanclares, ‘El derecho de información de las Comunidades Autónomas en los procesos de celebración de Tratados internacionales: evolución a la luz de la práctica y de las recientes reformas de los Estatutos de Autonomía’ in E Aja (ed), Informe de las Comunidades Autónomas 2007 (Barcelona, Instituto de Derecho Público, 2008) 89. 19 Fernández Segado, above n 11, 115. 20 This concept was used by M Pérez González, ‘La acción exterior del Estado y las Autonomías: Desarrollos en la práctica estatutaria’ in Cursos de Derecho Internacional de Vitoria-Gasteiz (Vitoria, Universidad del País Vasco, 1986) 265. 21 C Fernández de Casadevante Romaní, La acción exterior de las Comunidades Autónomas-Balance de una práctica consolidada (Madrid, Dilex-Universidad Rey Juan Carlos, 2001) 251. 22 M Sobrino Heredia, La acción exterior de las Comunidades Autónomas. Especial referencia a Galicia (Santiago de Compostela, Fundación Galicia-Europa, 2001) 29. 23 Spanish Constitutional Court, No 165/1994, 26 May 1994. 24 P Pérez Tremps, ‘La acción exterior y la participación europea ante una posible reforma del Estatuto de Cataluña’ in Estudios sobre la reforma del Estatuto (Barcelona, IEA, 2004) 355–66. 25 Art 4(2) TEU. 26 For further analysis of these principles, see A Heritier, D Kerwer, C Knill, D Lehmkuhl, M Teutsch and AC Douillet, Differential Europe. The European Union Impact on National Policymaking (Lanham, Rowman & Littlefield Publishers, 2001).

386 Maite Zelaia Garagarza authorities.’27 At the same time, however, the neutrality of EU law implies that ‘a Member State may not plead provisions, practices or circumstances existing in its internal legal system in order to justify a failure to comply with its obligations under Community law’.28 From the outset, the Spanish Constitutional Court has adopted a similar stance: EU law cannot modify the internal territorial structure of the Spanish State. The Constitutional Court has held that the EU lacks the constituent power necessary for the modification of the constitutional territorial structure.29 More particularly, the Constitutional Court has reiterated that EU law does not make any prejudgment about the level of government through which it is to be implemented.30 The implementation of EU law by the regions is a common practice in Spain. There is no doubt that EU law requires legislative and executive intervention not only from the central State institutions, but also from Autonomous Communities.31 This depends on the division of competences resulting from the Constitution and the Statutes of Autonomy. Specifically with respect to power-sharing, the newlyminted Statutes of Autonomy introduce principles that specify the division of competences as regards the implementation of European norms, in contrast to the previous version, which preceded the accession of the Spanish State to the European Communities and hence did not mention the issue. However, not all statutory texts of the ‘new generation’ follow the same approach. Let us examine the most relevant aspects. The Statute of Autonomy of Catalonia has designed a formula for the implementation and application of EU law, which determines that this task appertains to the Government of Catalonia (Generalitat) if and when dealing with its own competences.32 Likewise, it establishes that the existence of a rule of EU law does not modify the internal distribution of powers that the Constitution and the current Statute establish.33 If the EU should adopt legislation that replaces the basic State 27 See, eg Case 97/81 Commission v Netherlands [1982] ECR 1819, para 12; Case C-157/89 Commission v Italy [1991] ECR I-57, paras 10–11 and 16–17. 28 See, eg, Joined Cases 227/85 to 230/85 Commission v Belgium [1988] ECR 1, para 10; C-33/90 Commission v Italy [1991] ECR I-5987, para 24; Case C-419/01 Commission v Spain [2003] ECR I-4947, para 22; Case C-392/01 Commission v Spain [2002] ECR I-11111, para 9. 29 Spanish Constitutional Court, No 252/1988, 20 December 1988, para 2; No 76/1991, 11 April 1991; No 115/1991, 23 May 1991; 236/1991, 12 December 1991; No 79/1992, 28 May 1992; No 117/1992, 16 September 1992; No 80/1993, 8 March 1993. See also the Declarations by the Constitutional Court No 1/1991, 1 July 1992; No 1/2004, 13 December 2004, analysed in M Aragón Reyes, ‘La Constitución española y el Tratado de la Unión: la reforma de la Constitución’ (1994) Revista Española de Derecho Constitucional 9; A López Castillo, A Sáiz Arnaiz and V Ferreres Comella, Constitución española y Constitución europea: análisis de la Declaración del Tribunal Constitucional (DTC 1/2004, de 13 de diciembre) (Madrid, Centro de Estudios Políticos y Constitucionales, 2005). 30 Spanish Constitutional Court, No 13/1998, 22 January 1998, para 3. 31 For an analysis of the general participation of the Autonomous Communities in the European institutions, see J Bengoetxea, ‘The Participation of Infra-State Entities in European Union Affairs in Spain: the Basque Case’ in S Weatherill and U Bernitz (eds), The Role of Regions and Subnational Actors in Europe (Oxford, Hart Publishing, 2005) 47–66. 32 Statute of Autonomy of Catalonia, Art 113. 33 Statute of Autonomy of Catalonia, Art 189(1).

The Spanish State Structure and EU Law 387 norms, the Statute of Autonomy confirms the competence of the Generalitat to adopt the necessary implementing legislation for the norms of EU law.34 An appeal was brought before the Constitutional Court contesting the constitutionality of these statutory provisions. The Court upheld their constitutionality, with the result that the Autonomous Community of Catalonia can implement EU law within its own sphere of competence. In the same vein, the Constitutional Court considered there to be no existing impediments to the autonomous authorities adopting implementing legislation on the basis of a norm of EU law that substitutes for the basic State norm in a specific subject-matter. Nevertheless, at no time will the State constitutional competence for adopting basic State legislation be displaced or eliminated by EU law. Thus, the ruling asserts that nothing will impede the State authorities from approving, in the future, basic norms in the exercise of their constitutionally reserved competences.35 In the same sense, the Andalusian Statute of Autonomy reserves for the autonomous authority the power to implement and execute EU law whenever it is dealing with matters concerning its competence, and in accordance with what the Andalusian Parliament law establishes. Similarly, in line with the Statute of Autonomy of Catalonia, if the EU should adopt legislation that replaces the basic State norms, the autonomous authority adopts the implementing legislation based on the norm of EU law without any intervention by the State.36 The Statutes of Autonomy of Valencia, Aragon, Castilla y León and Navarre adopt the same approach by pointing out that these Communities have competences at their disposal, without prejudice to State legislation, and without altering the constitutional division of competences and the constitutional set-up.37 The new Statutes of Autonomy have taken into account the fact that, in many cases, the implementation of norms of EU law requires the adoption of measures that go beyond the sphere of competence of the Autonomous Community. This is so as regards the Statute of Autonomy of Catalonia, according to which it is to be the State that takes the appropriate measures, after prior consultation with the Generalitat. Additionally, the Generalitat, in so far as it is permitted, should participate in the decision-making bodies in which the relevant decisions are adopted; and if participation is not feasible, it should at least express its opinion on the adoption of the State decision.38 The constitutionality of these provisions of the Catalan Statute of Autonomy was also the object of an appeal, but it was upheld by the Constitutional Court. The Court accepted that the Statute acknowledges State powers in the matter of 34

Statute of Autonomy of Catalonia, Art 189(3). Spanish Constitutional Court, No 31/2010, 28 June 2010, para 123. Previously, the same position was defended in the judgment of the Spanish Constitutional Court, No 148/1998, 2 July 1998, para 4. 36 Andalusian Statute of Autonomy, Art 235(1)–(2). 37 Statute of Autonomy of Valencia, Art 61(3) d); Statute of Autonomy of Aragon, Art 93(2); Statute of Autonomy of Castilla y León, Art 62(3); Statute of Autonomy of Navarre, Art 68(4). 38 Statute of Autonomy of Catalonia, Art 189(2). 35

388 Maite Zelaia Garagarza adoption of internal mechanisms for the implementation of EU law in cases which require action beyond the Catalan territory. The only limitation imposed by the Constitutional Court pertains to the participation of the Autonomous Community in State bodies, which is limited to those that do not possess a decision-making competence.39 The Statute of Autonomy of the Balearic Islands follows the same approach. On the one hand, it assumes autonomous powers of implementation and execution of EU law. On the other hand, State authorities are required to consult the Autonomous Authorities before implementing a State norm.40 In short, the new Statutes of Autonomy reaffirm the Autonomous Communities’ competence to implement EU law. Some statutory texts contain more detail than others, but all of them confirm the basis of autonomous competences and state that the exercise of those competences does not alter the internal distribution of competences established between the State and the Autonomous Communities. Once the competence to implement a norm of EU law is attributed to the Autonomous Community, the latter may exercise that competence through a legislative or an executive act. When the norm of EU law calls for application to a specific case, and in other words requires more than purely legislative action, the public administration has to adopt specific measures and decisions directed at the parties concerned. In essence, Autonomous Communities implement EU law legislatively through autonomous law.41 When an Autonomous Community implements a norm of EU law, it is at a regulatory crossroads where it needs to determine, for lack of a better expression, what norm has to ‘give way’ to the other. That determination frequently also pertains to State law, since the implementation of EU law by the Autonomous Communities often also involves State ordinances. The outcome depends, as we have just shown, on the territorial entity to which the competence to implement the norm of EU law applies. Having determined the authorities with the competence to implement EU law, we move on to the elements that the legislative autonomous intervention should take into account in the course of its implementation. In this sense, the most relevant conditioning element is the type of obligation to be fulfilled by the Autonomous Community. In other words, the mode of autonomous intervention will be determined by the type of obligation that EU law imposes. The key distinction, which will serve as a starting point, is the distinction between general EU legal instruments (with a normative content)42 and individual EU legal instruments (or decisions).43

39

See above n 35, 2010 decision, paras 111 and 123. Statute of Autonomy of the Balearic Islands, Art 109. 41 See generally, R Jiménez Asensio, La ley autonómica en el sistema constitucional de fuentes del Derecho (Madrid, IEA-IVAP-Marcial Pons, 2001). 42 Arts 288–291 TFEU. 43 Arts 288(3) and 291(2) and (4) TFEU. 40

The Spanish State Structure and EU Law 389 The Autonomous Communities use normative instruments to implement EU norms of a general scope. Such an autonomous normative instrument should incorporate fundamental aspects of the normative content of the EU legal instrument in question. Implementation through a series of individual acts, thus fragmenting the EU normative programme and rendering knowledge, understanding and implementation of the specific EU norms more difficult, would put at risk the integrity of the EU law obligation within the territory of the Autonomous Community in question.44 In principle, it would equally be valid for the Autonomous Community to decide to implement the normative programme through diverse autonomous normative acts. In any case, the normative content of the EU act, whatever its nature, should be implemented in full. For example, the Autonomous Community might decide to implement the central aspects of a directive by an autonomous act, allowing any further implementation to be realised through administrative acts. Needless to say, the law that enables the adoption of further administrative acts should incorporate all the elements of the directive.45At any rate, according to settled case law of the Court of Justice, the transposition of a directive into domestic law does not necessarily require the provisions of the directive to be enacted in precisely the same words in a specific, express provision of national law and a general legal context may be sufficient if it actually ensures the full application of the directive in a sufficiently clear and precise manner.46

That said, according to equally settled case law, the provisions of a directive must be implemented with unquestionable binding force and with the specificity, precision and clarity required in order to satisfy the need for legal certainty, which requires that, in the case of a directive intended to confer rights on individuals, the persons concerned must be enabled to ascertain the full extent of their rights.47

Another aspect to be taken into account when dealing with European normative implementation is the ‘legislative reserve’. In principle, this standard requires that the matter, object of the implementation and, at the same time, object of the legislative reserve, must be regulated by a legislatively approved norm. Until now, no conflict has ever arisen. However, some doubts exist. First, on the assumption that the norm of EU law to be implemented does not determine the type of legal instrument necessary for its transposition, the principle of institutional and procedural autonomy would be activated, so that the Autonomous Community might choose the appropriate normative formula for implementation. Secondly, 44 See, amongst many other judgments, Case 69/77 Commission v Italy [1978] ECR 1749; Case 102/79 Commission v Belgium [1980] ECR 1473; Case 29/84 Commission v Germany [1985] ECR 1661; Case 116/86 Commission v Italy [1988] ECR 1323; Case 360/87 Commission v Italy [1991] ECR I-791; Case 131/88 Commission v Germany [1991] ECR I-825. 45 For a further analysis, see S Muñoz Machado, A Boix, D Sarmiento, M González and M Zelaia, Comunidades Autónomas y Unión Europea (Madrid, Iustel, forthcoming). 46 Case C-50/09 Commission v Ireland (ECJ, 3 March 2011), para 46, and the case law cited there. 47 Ibid.

390 Maite Zelaia Garagarza if a norm of domestic law is found to be incompatible with the norm of EU law that is to be implemented, the Court of Justice has held that the incompatibility of national legislation with provisions of the Treaty, even provisions which are directly applicable, can be finally remedied only by means of national provisions of a binding nature which have the same legal force as those which must be amended.48

It might be asked whether the norm of EU law would indeed have the capacity to replace the internal legislator. Perhaps it would be sufficient for the European norm to be implemented by a binding norm that removes any possibility of legal uncertainty, thus allowing administrative authorities (State or autonomous) to implement the content of the norm of EU law. Lastly, this section turns to the issue of autonomous implementation through decree-laws. Ever since the new Statutes of Autonomy were approved, the Autonomous Community governments have been allowed to adopt decree-laws (provisions with legal status) in exceptional situations.49 Is this type of provision appropriate for the implementation of EU law? In principle, no impediment appears to exist in that regard. Besides, in reply to arguments of the Spanish Government seeking to justify its non-compliance with its obligation to transpose directives on the basis of lack of time, the Court of Justice has in no case prohibited the use of extraordinary instruments such as decree-laws.50 The Union legal order is not opposed to Member States and/or Autonomous Communities using this instrument. However, the national Government has not been in favour of using decree-laws to implement norms of EU law. The reasons for this reticence are not clear, but they might be based on the apprehension of regulating in this manner matters that Article 86 of the Spanish Constitution prevents from being regulated by decree-law. Leaving aside whether the precautions are reasonable or not, there appears to be in fact no real obstacle to using decree-laws. It remains to be seen whether Autonomous Regions whose Statutes provide for that category of legal instruments go on to use it to implement norms of EU law. As previously shown, an Autonomous Community may implement a norm of EU law through its legislative or executive competences. Executive intervention is basically reserved for the implementation of European acts of an individual nature. Specific measures aimed at citizens, the principal beneficiaries or affected by such measures, have to be adopted by the autonomous administrative authority. It is therefore appropriate to examine how these measures affect citizens; that is, whether they are capable of having direct effect and primacy.

48

Case 168/85 Commission v Italy [1986] ECR 2945, para 13. See, eg Statute of Autonomy of Catalonia, Art 64(1); Statute of Autonomy of Valencia, Art 44(4); Statute of Autonomy of Aragón, Art 44(1); Statute of Autonomy of Andalusia, Art 110(1); Statute of Autonomy of Castilla y León, Art 25(4); Statute of Autonomy of the Balearic Islands, Art 49(1). 50 Case C-31/89 Commission v Spain [1990] ECR I-2139. 49

The Spanish State Structure and EU Law 391 IV. CONCLUSION

This chapter has offered a brief analysis of the Estado de las Autonomías, and an exploration of the latest normative and jurisprudential developments in the relationship between the Autonomous Communities and the European Union. The premise of our enquiry is the constitutional limitation of the political autonomy of the Autonomous Communities. The principle of unity is the basic principle of the Estado de las Autonomías. The constituent power instituted a single sovereignty that is attributed to the Spanish people, and the indissoluble unity of the nation. Therefore, the principle of autonomy should be interpreted as the attribution of competences to the autonomous institutions for the management of their respective interests. The exercise of this autonomy will always be limited by the principle of unity, as the autonomy is a power within the State, and its capacity for independent action is different from the notion of sovereignty. The initial case law of the Constitutional Court was along those lines and its current position is very similar, as evidenced by Judgment No 31/2010.51 This ruling is the answer of the Constitutional Court to the contents of the new Statute of Autonomy of Catalonia. This Statute is the archetype of the new generation of Statutes that have tried to extend the Autonomous Communities’ powers to some new areas. The Constitutional Court declared unconstitutional several articles regarding, for instance, the financial system, the use of language and the judiciary. Conversely, the Court upheld the constitutionality of certain tools of regional participation in EU institutions provided by the Statute. In this sense, the position of the Constitutional Court is now closer to the idea that the obligations of the State and of the Autonomies must be interpreted in the light of EU law. Moreover, that idea provides a framework for interpreting the amendments to the division of competences between State and sub-state entities, promoting a correct implementation of EU law by regional governments and the State, and creating procedures of co-operation between them.

51

See n 35 above.

17 Economic and Monetary Union: Caught between Brussels and Luxembourg? The Influence of EU Law on Belgian Federalism Case Law STEF FEYEN*

I. INTRODUCTION

E

UROPEAN UNION (EU) law influences Belgian federalism in a multitude of ways. Some strains of influence are quite conspicuous. As in many other federal Member States, a substitution mechanism was established to ensure that violations of EU law by federated entities might be remedied at the central level.1 These and other institutional mechanisms are dealt with in other contributions to this volume. The broad question underlying this chapter, however, is how EU law impacts on the federalism case law of the Belgian Constitutional Court (BCC), formerly the Court of Arbitration. ‘Influence’ may be understood in many different ways. For one thing, one might approach the concept from a ‘political’ perspective, focusing on the fact that the BCC has occasionally evacuated controversial federalism-related issues to the Court of Justice through preliminary references.2 One might also take a more ‘dogmatic’ approach by highlighting, for instance, that EU law sometimes seems

* PhD Fellow of the Research Foundation—Flanders (FWO), Institute for Constitutional Law, KU Leuven. I should like to thank Elke Cloots, Stefan Sottiaux, Valérie Verbist and Willem Verrijdt for helpful comments, suggestions and discussion. The usual disclaimers apply. 1 Belgian Constitution, Art 169; Special Law of 8 August 1980 on institutional reform, Art 16(3). See also R Vuillermoz, La Belgique, l’Espagne et l’Italie face à l’intégration communautaire: quelle adaptation des rapports entre l’état et ses collectivités régionales? (Brussels, Bruylant, 2003) 476–85. Cp K Lenaerts, ‘The Application of Community Law in Belgium’ (1986) 23 CML Rev 253, 259–60. 2 A plausible recent example of this phenomenon is BCC, No 50/2011, 6 April 2011, especially para B.7.2 (suspending proportionality review while waiting for an answer from Luxembourg). Also the care insurance saga, discussed below, might be analysed in this light.

Economic and Monetary Union: Caught between Brussels and Luxembourg? 393 to force the federated entities to co-operate.3 However, these are not the forms of influence central to this chapter. Instead, this contribution concentrates on the ‘economic and monetary union’ (EMU). The BCC introduced the EMU doctrine shortly after its creation. Later on, the ‘principle’ was enshrined in the Special Law of 8 August 1980 on institutional reform (hereafter ‘SLIR’).4 Of course, the EMU is not a Belgian invention. Both the BCC and the Belgian federal Parliament were thoroughly inspired by EU law when seeking to safeguard ‘the institutional framework of an economy built up out of federated entities, which is characterised by an integrated market (economic union) and unity of currency (monetary union)’.5 In fact, a central tenet of the EMU, ie the ‘free movement of goods and production factors amongst the federated entities of the State’,6 is reminiscent of the European fundamental freedoms. More typically Belgian, by contrast, is the principle of territoriality which underlies the allocation of competences. According to the BCC, this principle requires that each federated entity’s jurisdiction be restricted to the entity’s territory.7 This contribution explores whether, and to what extent, the European and Belgian conceptions of the EMU principle converge. In other words, it is examined to what extent EU law has influenced the BCC case law regarding the EMU. However, before starting our inquiry, the basics of Belgian federalism, including the role of the BCC, should be outlined.

II. BELGIAN FEDERALISM: AN OVERVIEW

Since 1994, Article 1 of the Belgian Constitution proclaims that ‘Belgium is a federal state composed of Communities and Regions’, thus confirming the existence of two distinct kinds of federated entities in Belgium. On the one hand, the Communities have generally been empowered to deal with educational, social and linguistic matters in their own (monolingual) territories.8 The Regions, on the other hand, have jurisdiction in certain economic areas, in environmental and energy policy matters, and in several other domains.9 By

3

A recent example of this phenomenon is BCC, No 33/2011, 2 March 2011, para B.11. Art 6(1) VI (3) SLIR. 5 BCC, No 47, 25 February 1988, paras III.6.B.4–III.6.B.5. 6 Ibid. 7 Established case law since BCC, No 9 and No 10, 30 January 1986, paras III.5.B.1 and III.8.B, respectively. For the Regions, the same territorial principle has been upheld in BCC, no 33/2011, 2 March 2011, para B.5. 8 Ie, the Flemish Community in the Dutch-speaking territory, the French Community in the French-speaking territory and the German-speaking Community in the German-speaking territory. In the bilingual Region of Brussels, this type of power is exercised by the two former Communities, by a special Commission or by the federal Government, depending on the subject matter in question. 9 The Flemish Region has jurisdiction in the Dutch-speaking territory, the Walloon Region in the French- and German-speaking territories, and the Brussels Region in the bilingual Brussels territory. 4

394 Stef Feyen and large, conflicts between the legal acts of those distinct sub-State entities, and between these acts and federal legislation, have been prevented by the adoption of a system of exclusive competences. This system implies, first, that each legal situation must be qualified to fit the material scope of powers of the federal State, the Communities or the Regions. A second element of the system of exclusive competences is the aforementioned ‘territoriality principle’. As already noted, this principle entails that the sub-State entities’ powers are restricted to their own respective territories.10 The BCC was established as a necessary corollary of the Belgian federalisation process. The Court’s original task was to ensure observance of the rules governing the allocation of competences.11 Even though the bulk of its work today concerns fundamental rights issues, the Constitutional Court still fulfils its traditional function.12 To that end, it has developed a number of principles, on which it relies when assessing whether a legislature remained within the boundaries of its powers. These principles include, most notably, the principle of conferral,13 the principle of proportionality,14 the principle of federal loyalty,15 (the principle of) implied powers16 and the EMU.17 The Court’s interpretation of the said principles has clearly evolved over time. As André Alen, a judge of the BCC, explains: In the first years of its existence, the Court of Arbitration laid the foundation for the allocation of competences, which resulted [on the one hand] in a broad conception of the powers of the Communities and the Regions … and [on the other hand] in the maintenance of an exclusively territorial system of power division.18

Later on, Alen continues, the BCC gradually expanded the notion of the EMU, and started tolerating certain deviations from the territoriality principle.19 Thus, the BCC has tried to strike a proper balance between the principles the observance of which it must ensure.

10 See n 7 above. The bilingual Brussels area is a special case in this regard. Due to the complexity of this case and its limited relevance to the subject of this chapter, I shall refrain from discussing it. 11 Belgian Constitution, Art 142; Special Law of 6 January 1989 on the Constitutional Court, Arts 1, 1° and 26(1) 1°. Hence the BCC’s original name: ‘Court of Arbitration’. The Court acts as an ‘arbiter’ of conflicts between the federal State and the sub-State entities, and between the latter entities themselves. In 2007, the Court’s name was changed to ‘Constitutional Court’. 12 Moreover, there are some intricate connections between the two areas of jurisdiction. See, eg, BCC, No 119/2003, 24 September 2003, para B.3.1. 13 Belgian Constitution, Art 35. See, eg, BCC, No 25, 26 June 1985, para III.B.1. 14 See, eg, BCC, No 14/91, 28 May 1991, para B.3.3. 15 Belgian Constitution, Art 143(1). See, eg, BCC, No 119/2004, 30 June 2004, paras B.3.2–B.3.3; BCC, No 124/2010, 28 October 2010, para, B.39.2. 16 Art 10 SLIR. 17 Ie, to the extent that the EMU may be conceptualised as a ‘principle’ (see section III. below). 18 A Alen, ‘Twintig jaar grondwettigheidstoezicht op wetgevende normen’ in A Alen (ed), 20 jaar Arbitragehof (Malines, Kluwer, 2005) 6. 19 Ibid 7 ff.

Economic and Monetary Union: Caught between Brussels and Luxembourg? 395 III. THE EMU: CONVERGENCE?

A. Freedom of Trade and the EMU In the Belgian mixed economy, the Government seeks to ensure the effective operation of the free market. Even though this implies that the Government can regulate the economy, governmental power to interfere with market processes is obviously not unlimited. The BCC assesses the constitutionality of such interferences in the light of the freedom to trade, which it interprets as an individual right that must be safeguarded against undue legislative burdens.20 The intensity of the Court’s review of freedom of trade restrictions is generally rather low.21 Besides, as a principle geared towards protecting a free market in a more collective sense, the freedom of trade is imbued with a specific meaning.22 Indeed, the Belgian EMU might be regarded as an ‘institutional’ version of the free trade principle: it ensures that institutions and, more particularly, the federal structure of the Belgian State do not unduly interfere with the free market.23 In order to guarantee respect for the free trade principle within the Belgian economic union, the BCC has adopted various doctrines. To begin with, the Court has held that customs or charges having equivalent effect levied by federated entities are incompatible with the Belgian EMU. For instance, in the Water Export case, an action for annulment was brought against a Walloon decree introducing a tax on the export of water from the Walloon Region. The Court held that such a tax, the object of which was the transit of water, amounted to a customs duty between the federated entities. Therefore, and irrespective of whether the tax might have had a constitutional basis in a regional autonomous fiscal power, the Court invalidated the Walloon decree on EMU grounds.24 Not only such charges, but also measures which allegedly had an adverse effect on the free movement of goods more broadly conceived,25 or on the free

20 Cp, eg, BCC, No 6/96, 18 January 1996, para B.2.1 (merely relying on Art 6(1) VI (3) SLIR) and BCC, No 9/2011, 27 January 2011, para B.26.1 (relying on the freedom of trade as enshrined in the décret d’Allarde, read in conjunction with Art 6(1) VI (3) SLIR and with the general principle of nondiscrimination as enshrined in Arts 10–11 of the Belgian Constitution). 21 The Court finds a violation only ‘if [the freedom] is unnecessarily restricted, in a manner which is clearly disproportionate to the pursued goal’ (eg BCC, No 29/2010, 18 March 2010, para B.33.2; No 99/2010, 16 September 2010, para B.6.2; No 56/2011, 28 April 2011, para B.8.2; No. 166/2011, 10 November 2011, para B.10; No. 180/2011, 24 November 2011, para B.17.2). 22 See, eg, BCC No 55/92, 9 July 1992, para B.8; No 149/2010, 22 December 2010, para B.8; cp No 180/2011, 24 November 2011, para B.17.3 (leaving open the question of Belgian specificity). 23 X Delgrange and JP Keppene, ‘Convergences et divergences dans le contrôle constitutionnel de la non-discrimination et l’Union économique et l’unité monétaire’ in X Delgrange (ed), L’Union économique et l’unité monétaire en Belgique (Bruges, La Charte, 1993) 92. 24 BCC, No 47, 25 February 1988, para III.6.B.4. 25 See, eg, BCC, No 51, 20 May 1998, paras B.3.1–B.3.6; No 119/2004, 30 June 2004, para B.7.1.

396 Stef Feyen movement of persons,26 services27 or capital,28 have been subject to constitutional scrutiny. In reviewing such measures, the Court has sometimes—prima facie— relied on internal Belgian law,29 while referring to EU law on other occasions.30 Consequently, the question arises as to whether the Belgian and the European conceptions of the EMU converge.

B. The Rules Governing the Belgian and the European EMU The rules governing the European EMU may be summarised as follows:31 every restriction32 to one of the four freedoms must be justified in order to be in conformity with EU law, either by relying on one of the enumerated exceptions,33 or by relying on the so-called ‘rule of reason’.34 In addition, tariff barriers per se are prohibited under EU law.35 Although this general scheme is fairly simple, it has not always been easy to apply. How the Court of Justice applies the scheme or ‘structure’ in practice depends on the ‘sub-norms’ to which it resorts. For the purposes of this chapter, ‘sub-norms’ are doctrines consisting of norms that seek to apply the general structure to particular cases. Exploring what particular ‘sub-norms’ the Court of Justice has been taking into account when monitoring free movement restrictions is a strenuous task, and determining the exact content of these norms even more so. The most easily traceable ‘sub-norms’ are those which have been enshrined in the Treaties. For instance, certain grounds of justification have been accepted only in the context of the free movement of goods, not in the context of freedom

26 BCC, No 11/2009, 21 January 2009, paras B.7 ff (Belgian Care Insurance Judgment) (free movement of workers); BCC, No 83/2010, 8 July 2010, para B.9 (Flemish Inheritance Tax Judgment) (freedom of establishment). 27 See, eg, BCC, No 56/96, 15 October 1996, paras B.6 ff; No 119/2004, 30 June 2004, para B.2.4. 28 See, eg, BCC, No 83/2010, 10 July 2010, para B.9. See also No 67/2000, 14 June 2000, paras B.3.1–B.3.5. 29 For examples, see nn 24–27 above. 30 Eg BCC, No 130/2005, 19 July 2005, paras B.15–B.17 (planning to rely on the European conception of free movement of capital); No 139/2008, 22 October 2008, para B.16.4 (relying on the European conception of free establishment); No 128/2011, 14 July 2011, paras B.13.4 ff (relying on the European conception of free movement of services); also see No 166/2011, 10 November 2011, paras B.29.1 ff.; No 180/2011, 24 November 2011, paras B.14.1 ff.; cp No 40/2012, 8 March 2012, paras B.8.1 ff. 31 K Lenaerts and P Van Nuffel, European Union Law (London, Sweet & Maxwell, 2011) 202–03, 246–47, 277–78 and 289–90. 32 See, eg Case 8/74 Dassonville [1974] ECR 837, para 5 (goods); Case C-19/92 Kraus [1993] ECR 1663, para 32 (persons); Case C-76/90 Säger [1991] ECR I-4221, para 12 (services). 33 Arts 36, 45(4), 52, 62 and 65 TFEU (permitting, in certain circumstances, direct discrimination on grounds of nationality). 34 Case 120/78 Rewe-Zentral [1979] ECR 649, para 8 (goods); Case 96/85 Commission v France [1986] ECR 1475, para 11 (self-employed persons); Case C-204/90 Bachmann [1992] ECR I-249, para 28 (employees); Case 205/86 Commission v Germany [1986] ECR 3755, paras 27–29 (services); Case C-367/98 Commission v Portugal [2002] ECR I-4731, para 49 (capital). 35 Art 30 TFEU. See generally Lenaerts and Van Nuffel, above n 31, 207–11.

Economic and Monetary Union: Caught between Brussels and Luxembourg? 397 of movement for persons.36 Other ‘sub-norms’, however, have been established by the Court of Justice. Here, the problem of determining their status and content becomes more apparent. Even though the Court as well as legal commentators have tried to clarify these norms, the debate about their status and content continues. In the light of these difficulties, it is often the outcome of the case that becomes the most obvious criterion of comparison. The BCC has adopted a structurally similar approach. Like the Court of Justice, the BCC investigates whether there is a restriction on the freedom of movement and, if so, whether that restriction can be justified. However, certain authors have criticised the BCC for not sufficiently aligning itself with the ‘sub-norms’ established in the ECJ case law. It has been claimed that the BCC adheres to a different conception of the term ‘restriction’, and does not conduct its proportionality review as stringently as it should.37 More precisely, it has been argued that the BCC has failed, on occasion, adequately to assess the purpose of the restrictive measure,38 and that its proportionality test does not take sufficient account of the Court of Justice’s ‘least restrictive means’ test.39 These authors conclude, therefore, that if such cases had been brought before the Court, the outcome would have been different (mutatis mutandis). In the light of the abovementioned difficulties associated with the establishment of sub-norms, we can only speculate about that. Yet it does seem fair to say that the BCC has been quite lenient in monitoring respect for the Belgian EMU principle.40 It appears that only if the domestic measure constitutes a customs duty, a charge having equal effect or some other manifest trade obstruction will the BCC strictly enforce the principle. Hence, the Court gives Belgium’s various legislatures considerable leeway to balance economic and other interests as they see fit.41

36 See, eg, ‘the protection of industrial and commercial property’ as enshrined in Art 36 TFEU. There is no similar ground of justification in the Treaty provisions on freedom of movement of persons. 37 Delgrange and Keppene, above n 23, 91 ff; C Vanderveeren and J Vuchelen, ‘Een Vlaamse fiscaliteit en de Belgische economische monetaire unie’ in C Vanderveeren and J Vuchelen (eds), Een Vlaamse Fiscaliteit binnen een economische en monetaire unie (Antwerp, Intersentia, 1998) 63; P Van Ypersele, ‘Quelques réflexions “catastrophistes” sur les risques que comportent les divergences de contrôle’ in X Delgrange (ed), L’union économique, above n 23, 126 ff. See also M Melchior and P Vandernoot, ‘Contrôle de constitutionnalité et droit communautaire dérivé’ (1998) Revue belge de droit constitutionnel 3, 24–25. 38 BCC, No 32/91, 14 November 1991, paras 5.B.1.8–5.B.1.9, as discussed in Delgrange and Keppene, above n 23, 101–02. 39 Delgrange and Keppene, above n 23, 114–16. 40 Very few violations of the Belgian EMU have been found. Moreover, most of those violations merely pertained to the Belgian customs union. See, eg, BCC No 29/96, 15 May 1996, para B.8.4; No 55/96, 15 October 1996, para B.4.2.6; No 34/97, 12 June 1997, paras B.7.5–B.7.7; No 128/2001, 10 October 2001, para B.12; Flemish Inheritance Tax Judgment, above n 26 (discussed below); No 123/2010, 28 October 2010, paras B.5.2–B.6. 41 Cp I Verougstraete, ‘Een nieuw soort rechters’ in Liber Amicorum Walter van Gerven (Deurne, Kluwer, 2000) 88 (arguing that the Court of Cassation takes a more ‘social’ approach than does the Court of Justice).

398 Stef Feyen Consequently, we may infer that the general schemes relied on by the BCC and the ECJ, though similar in structure, are applied in divergent ways. Accordingly, a prima facie conclusion might be that structural convergence does not necessarily entail substantive convergence, that is, similarity of outcome. However, in order to establish whether this preliminary conclusion is justified, we need to address another question first, which is logically prior to the premises of our preliminary conclusion: Are the EMU principles emanating from different legal systems comparable anyway?

C. The Question of Convergence in a Broader Perspective The thesis that the Belgian and European EMU conceptions coincide structurally, though not substantially, is based on a number of unarticulated assumptions. First, it seems to presuppose that the Belgian federated entities and the EU Member States are equivalent, such that the Belgian EMU and the EU fundamental freedoms ought to be construed in the same way.42 Secondly, the preliminary conclusion seems to assume that ‘convergence’ occurs either between general rules (structural convergence) or between particular outcomes (substantive convergence). Yet one might wonder whether there are some other relevant parameters of convergence in addition to structure and substance. Indeed, is it not preferable to assess ‘convergence’ in terms of a broader, principled rationale, such as the establishment of an effective internal market? Implicitly answering the latter question in the affirmative, some Belgian authors have defended the view that, structurally, the Belgian EMU should not be interpreted ‘in line’ with the European EMU. The rationale underlying this position is that Belgium, due to its strong, trans-regional interlocking economy, displays a greater degree of market unity than does Europe.43 Put differently, these commentators

42 Nonetheless, in some areas of EU law, this presupposition does not hold. Two important examples are the Azores/Rioja line of cases, where the federated entity is made the relevant point of comparison (for discussion, see K Lenaerts and N Cambien, ‘Regions and the European Court: Giving Shape to the Regional Dimension of the Member States’ (2010) 35 EL Rev 609, 629–33), and the Lancry line of cases, where the ECJ held that the relevant reference framework was not the Member State (see, eg Case C-163/90 Legros and others [1992] ECR I-4625, paras 16–17; Joined Cases C-363/93, C-407/93, C-408/93, C-409/93, C-410/93 and C-411/93 Lancry and Others [1994] ECR I-3957, paras 27–29; Joined Cases C-485-486/93 Simitzi [1995] ECR I-2655, para 17; see also A Tryfonidou, ‘Resolving the reverse discrimination paradox in the area of customs duties: the Lancry saga’ (2011) 22 European Business Law Review 311, 321 ff). 43 See, eg, Vanderveeren and Vuchelen, above n 37, 63; C Vanderveeren et al, De Economische en Monetaire Unie in de Belgische Staatshervorming. Juridische en economische aspecten (Antwerp, Maklu, 1987) 49. See also J Velaers and J Vanpraet, ‘De materiële en territoriale bevoegdheidsverdeling inzake sociale zekerheid en sociale bijstand (II)’ (2009) Tijdschrift voor Bestuurswetenschappen en Publiekrecht 195, 211.

Economic and Monetary Union: Caught between Brussels and Luxembourg? 399 argue that principled convergence may well necessitate structural divergence, ie different rules structuring the implementation of the EMU.44 Lastly, our preliminary conclusion appears to take for granted that the Belgian legal order is distinct from the EU legal order. Indeed, ‘convergence’ does not imply that EU law is simply applied, but presupposes the existence of ‘purely internal situations’ to which EU law cannot be directly applied. These ‘purely internal situations’ are then merely influenced by EU law. It is in terms of such ‘influence’ that the question of convergence should be assessed. Yet drawing the line between both legal orders is not always easy. Although the ECJ still maintains that EU law does not apply to purely internal situations,45 it has found itself able to decide on these situations when they are an extrapolation of EU law.46 In addition, the scope of EU law itself is one of the most controversial issues in EU law, not the least because of the influence of EU citizenship, which is, according to the ECJ, destined to be ‘the fundamental status of nationals of the Member States’.47 We think that any student of the degree of convergence between a domestic EMU principle and the European fundamental freedoms should be aware of the caveats revealed above. No single theory seems to be available in the light of which the question of convergence might be addressed in the abstract. Contentious judgments about controversial issues are inevitable. The scholarly discussions in the aftermath of the Flemish care insurance saga, which will be discussed below, clearly exemplify this. Many Belgian authors asserted that the Belgian and the European EMU conceptions are no longer in line.48 Although this contention may be true from a certain perspective, it is inevitably based on certain unarticulated assumptions. Keeping this in mind, we now turn our attention to the case law of the BCC and the ECJ. More particularly, two series of cases will be analysed: the Flemish care insurance saga, and the Flemish inheritance tax case. That will allow us to partially sidestep these abstract issues and to look at the question of ‘convergence’ from a more practical perspective. 44 See Vanderveeren et al, above n 43, 52–54 and 63. See also G Brouhns, ‘Débats’ in Delgrange (ed), L’union économique, above n 23, 45–50 (on how the general normative framework of the Belgian EMU is enshrined not only in Art 6(1) VI (3) SLIR, but also in distinct but related legal provisions). Cp Van Ypersele, above n 37, 124–25 (seeing a relevant difference in the centrifugal nature of the Belgian federation, as compared to Europe’s centripetal tendencies). 45 Cp Case C-34/09 Ruiz Zambrano (Court of Justice, 8 March 2011), Opinion of AG Sharpston (arguably advocating a reconsideration of the case law). 46 Cp Joined Cases C-297/88 and C-197/89 Dzodzi [1990] ECR I-3763, paras 36–42, especially para 37; Case C-448/98 Criminal Proceedings against Guimont [2000] ECR I-10633; Case C-280/06 Autorita Garante della Concorrenza e del Mercato [2007] ECR I-10893, paras 19–29, especially para 21; Case C-310/10 Agafi¸tei (Court of Justice, 7 July 2011), paras 38–39, but cp paras 46–47. But see Case C-48/07 Les Vergers du Vieux Tauves SA [2008] ECR I-10627, para 27. 47 Case C-184/99 Grzelczyk [2001] ECR I-06193, para 31. See also Case C-135/08 Rottman [2010] ECR I-1449, paras 43 and 56; Ruiz Zambrano, above n 45, para 41; Case C-391/09 Runevicˇ-Vardyn and Wardyn (Court of Justice, 12 May 2011), para 60. 48 See, eg Velaers and Vanpraet, above n 43, 210–11. See also H Verschueren, ‘Social Federalism and EU law on the free movement of persons’ in B Cantillon, P Popelier and N Mussche (eds), Social Federalism: the Creation of a Layered Welfare State. The Belgian Example (Antwerp, Intersentia, 2011) 222.

400 Stef Feyen D. The Flemish Care Insurance Saga The Flemish care insurance saga has been widely discussed, both in Belgian and in European academia,49 not the least because it is provides a quintessential example of the tension between EU law and the Belgian rules governing the allocation of competences.50 In this line of cases, the BCC was asked to review whether the Flemish care insurance scheme passed constitutional muster. The scheme was established for the benefit of persons suffering from a serious disability, on the basis of the Flemish Community’s autonomy in social assistance matters. The personal scope of the obligations and entitlements under the scheme was primarily determined on the basis of a residence criterion. After Flemish law had been modified in response to objections made by the European Commission, the scheme applied: (a) compulsorily to residents of Flanders; and (b) on a voluntary basis to residents of Brussels. However, this general scheme was qualified by two exceptions. First, persons who were, on the basis of the Social Security Regulation,51 subject to the social security scheme of another EU Member State were excluded from the ambit of the Flemish care insurance scheme. Secondly, persons who were, on the basis of that Regulation, subject to Belgian social security law and worked in the areas circumscribed under (a) and (b) above, but did not live in Belgium, were included. In a series of judgments, the BCC ruled that the Flemish Community had not exceeded the material and territorial limits of its powers by adopting the Care Insurance Decree.52 Nonetheless, in 2006, the BCC decided to make a number of references to the Court of Justice for a preliminary ruling on the decree. The most important question was whether the care insurance scheme violated the Treaty provisions on the free movement of persons, as persons who worked in the Flemish or Brussels Regions but lived in the Walloon Region were still excluded from the scope of the Flemish insurance scheme.53

49 The most important cases in this ‘saga’ are Case C-212/06 Government of the French Community and Walloon Government v Flemish Government [2008] ECR I-1683 (hereafter ‘European Flemish Care Insurance Case’) and the Belgian Care Insurance Judgment, above n 26. See also Herwig Verschueren’s contribution in ch 12 of this volume; P Van Elsuwege and S Adam, ‘The Limits of Constitutional Dialogue for the Prevention of Reverse Discrimination’ (2009) 5 European Constitutional Law Review 327, 328–30. 50 See, eg, T Vandamme, ‘Case note accompanying Case C-212/06’ (2009) 46 CML Rev 299. See also E Cloots, ‘Germs of Pluralist Judicial Adjudication’ (2010) 47 CML Rev 645, 667–68; D Pieters, De Europese Unie tegen een Vlaamse sociale zekerheid (Leuven, Acco, 2010) 7–8, 23 ff and 55 ff. 51 Council Regulation (EC) 1408/71 on the application of social security schemes to employed persons and their families moving within the Community [1971] OJ L149/2. 52 BCC, No 33/2001, 13 March 2001, paras B.3.4, B.3.7 and B.3.9 (where the Court held that the Flemish Community acted within the material scope of its powers); BCC, No 51/2006, 19 April 2006, paras B.7–B.9.8 (where the Court decided that the Flemish decree respected the territorial boundaries of the Flemish legislature’s powers). 53 BCC, No 51/2006, 19 April 2006, paras B.12.3, B.12.4 and B.13.4.

Economic and Monetary Union: Caught between Brussels and Luxembourg? 401 The Court of Justice distinguished in its ruling between purely internal situations and situations falling within the scope of the European Treaties.54 More precisely, it held that EU law was applicable to Walloon residents who worked in Flanders if, and only if, they had exercised their freedom to move within the EU. By contrast, Walloon residents who worked in Flanders but had never exercised their European free movement rights remained excluded from the scope of EU law. In a next step, the Court ruled that, as far as situations falling within the scope of EU law were concerned, the residence condition for entitlement to the care insurance scheme violated the free movement requirements laid down in what are now Articles 45 and 49 TFEU.55 In its 2009 judgment, the BCC reproduced the distinction the Court of Justice had drawn between those persons who had made use of their right to freedom of movement within the EU, and those who had not.56 As far as persons falling within the scope of EU law were concerned, the BCC faithfully copied the Court’s considerations.57 However, as regards the purely internal—Belgian—situations, the Court’s position was somewhat ambiguous. Though the Court’s conclusion was straightforward—the residence criterion did not violate the Belgian EMU58— the reasoning behind it was less clear. Moreover, the BCC’s reasoning in the 2009 judgment seems to contrast with the approach taken in 2006, when it was decided to get the Luxembourg Court involved. In the 2006 case, the BCC investigated whether the Belgian EMU principle was illegitimately restricted. The BCC decided it was not, since the Flemish care insurance scheme involved only small amounts of money and had a limited impact.59 In the 2009 case, by contrast, the BCC seemed to have changed its mind. More precisely, the BCC ducked the question, not going beyond a mere ‘suggestion’—inspired by the European case law—that the care insurance scheme might constitute a restriction on the Belgian EMU.60 Nevertheless, the very fact that the BCC reproduced the distinction between purely internal situations and situations falling within the ambit of EU law in its 2009 judgment implies that, at least potentially, the BCC remains willing to construe the Belgian EMU requirements differently from the Court of Justice’s reading of the fundamental freedoms. Moreover, there are several other cases in which the distinction between the Belgian and EU legal orders was central to the BCC’s reasoning.61 In an earlier decision pertaining to the Belgian EMU, the BCC 54

European Flemish Care Insurance Case, above n 49, paras 37–39. Ibid para 38–39. But see para 40. 56 Belgian Care Insurance Judgment, above n 26, paras B.7 and B.8.1. 57 Ibid para B.8.1 ff, especially paras B.9 and B.10.1. See also S Feyen, ‘Zorgen voor morgen’ (2009) Tijdschrift voor Bestuurswetenschappen en Publiekrecht 169, 173–74. 58 Belgian Care Insurance Judgment, above n 26, dictum. In fact, the Court merely confirmed its earlier case law on this point (ibid para B.12.1, quoting BCC, No 51/2006, 19 April 2006, para B.10.3). 59 BCC, No 51/2006, 19 April 2006, para B.10.3. See also Belgian Care Insurance Judgment, above n 26, para B.12.1. 60 Belgian Care Insurance Judgment, above n 26, para B.13.1, referring to European Flemish Care Insurance, above n 49, para 54. 61 See especially BCC, No 56/2009, 19 March 2009, paras B.4.3–B.4 and B.7.3 (cp Case C-433/04 Commission v Belgium [2006] ECR I-10653), confirmed in BCC, No 124/2009, 16 July 2009. 55

402 Stef Feyen even expressly denied that the European fundamental freedoms might be relied on in the context of purely internal situations.62

E. The Flemish Inheritance Tax Case The Flemish inheritance tax case originated in a Flemish decree that introduced an exemption from inheritance taxes relating to family undertakings. The exemption was available only if the undertaking employed, in the three years preceding the date of death of the deceased, at least five workers in the Flemish Region.63 The Court of Justice found that this condition amounted to ‘indirect discrimination’,64 incompatible with the freedom of establishment, as it provided for a difference in treatment between taxpayers on the basis of the place of employment. The Court of Justice was not convinced by the justification put forward by the Belgian Government. A first argument that the Government invoked in its defence related to the survival of small and medium-sized undertakings and the maintenance of employment. The ECJ, however, held that Belgium had ‘not been able to show the need to limit the exemption at issue to “family” undertakings’, and had disregarded the fact that similar arguments were applicable to undertakings from other Member States.65 The second defence argument, based on fiscal efficiency concerns, was equally dismissed. The Court pointed out that the difficulty of obtaining information from abroad could not justify a categorical refusal to grant tax exemptions, as less restrictive alternatives were available. In the Court’s view, it would not be excessively burdensome to ‘request the taxpayers concerned to provide themselves the evidence which the authorities consider necessary to be fully satisfied that those benefits are granted only where the jobs in question fulfil the criteria set out under national law’.66 In 2010, the BCC was called upon to decide whether the decree passed constitutional muster, this time under the Belgian EMU.67 The BCC first reiterated that a mere discrepancy between regional inheritance tax rates, which arguably might have an influence on the behavioural pattern of taxpayers, did not, as such, 62 BCC, No 132/2005 19 July 2005, paras B.6.1–B.6.3; see No 13/2012, 2 February 2012, para B.5.3; see also V Verbist, ‘Het standpunt ten aanzien van omgekeerde discriminatie in de Belgische rechtsorde’ in A Alen and J Theunis (eds), Leuvense Staatsrechtelijke Standpunten III (Bruges, Die Keure, 2012) forthcoming. 63 Flemish Inheritance Tax Judgment, above n 26, paras B.1–B.2. 64 Case C-464/05 Geurts and Vogten [2007] ECR I-9325, paras 20 and 22. The discrimination here was considered ‘covert’ or ‘indirect’, because the criterion of distinction (ie, having employed five or more workers in the Flemish Region) in fact leads to the same result as direct discrimination on the basis of the seat of a company (ibid para 20). 65 Geurts and Vogten, above n 64, paras 26–27. See also Opinion of AG Kokott, para 39 ff. 66 Geurts and Vogten, above n 64, para 28. See also Opinion of AG Kokott, paras 53–54. 67 It should be noted that the decree was modified in 2007. The criterion of employing workers in the Flemish Region was thereby abolished in response to the ECJ judgment. The BCC had to judge upon the constitutionality of the decree as it stood before this amendment (see Flemish Inheritance Tax Judgment, above n 26, para B.1). Nevertheless, the amendment may have influenced the Court’s reasoning (ibid para B.3).

Economic and Monetary Union: Caught between Brussels and Luxembourg? 403 infringe on the Belgian EMU.68 However, so the Court continued, the decree rendered more difficult the investment in corporations which did not at least employ five employees in the Flemish Region, since the consequences of an inheritance tax regulation were taken into account in an investment decision.69 This amounted to a restriction on the freedom of establishment and the free movement of capital. In the Court’s opinion, this restriction could not be justified. The argument based on employment and business survival did not succeed, since the condition in question was not necessary to attain that goal.70 At first sight, the European and Belgian EMU conceptions converged in this case. The Court of Justice and the BCC reached a similar conclusion: the Flemish tax exemption violated the freedom of movement. Moreover, they followed a similar route to arrive at that conclusion. Both courts dismissed the defence arguments based on business continuation and the maintenance of employment. Although neither of the courts denied that such considerations were, under certain conditions, capable of justifying tax benefits,71 the Court of Justice as well as the BCC found the ‘place of employment’ condition to be unnecessary to attain that aim.72 The Court of Justice argued, more precisely, that in relation to the objective of preventing inheritance tax from jeopardising the continuation of family undertakings, and therefore the jobs which they bring, undertakings having their seat in another Member State are in a situation comparable to that of undertakings established in [Flanders].73

The BCC, on the other hand, noted that promoting business continuity and maintaining employment ‘can be sufficiently safeguarded by a condition of minimum employment, irrespective of the Region where that employment occurs’.74 Although the outcome of and the reasoning adopted in both judgments largely coincide, it is important to emphasise that the BCC still regards the Belgian legal order as distinct from the EU legal order. Indeed, it is quite clear that the Belgian inheritance tax case was not directly governed by EU law. Rather, it was subject to national law, the reading of which was notably ‘influenced’ by EU law.

F. Comparative Conclusion The BCC’s judgments in the care insurance case (2009) and the inheritance tax case (2010) are similar in that they both confirm that the Belgian and the European legal orders are distinct from one another. However, there are also important differences between the two cases. A first prima facie disparity concerns 68 69 70 71 72 73 74

Flemish Inheritance Tax Judgment, above n 26, para B.8. Ibid B.9 (referring to Geurts and Vogten, above n 64, para 17). Flemish Inheritance Tax Judgment, above n 26, para B.9. Geurts and Vogten, above n 64, para 26; Flemish Inheritance Tax Judgment, above n 26, para B.9. Geurts and Vogten, above n 64, para 27; Flemish Inheritance Tax Judgment, above n 26, para B.9. Geurts and Vogten, above n 64, para 27. Flemish Inheritance Tax Judgment, above n 26, para B.9 (emphasis added).

404 Stef Feyen the impact of the European EMU principle on its Belgian counterpart. Whereas in the care insurance case, the BCC denied that Belgian constitutional law required convergence between the domestic EMU principle and the fundamental freedoms enshrined in the Treaties, it did base its reading of the former principle on the ECJ case law in the inheritance tax case. This disparity, however, is merely of a prima facie nature, taking into account, as explained above, the unarticulated assumptions associated with it. A second, more essential, difference relates to the reasons proffered by the BCC for ‘deviating’ from the European free movement conception in the one case but not in the other. In the care insurance case, the Court expressly referred to the Belgian system of power division, which apparently trumped countervailing considerations relating to the EMU.75 The Court failed, however, to give reasons for why it aligned itself so easily with the Court of Justice in the inheritance tax case. Indeed, why did the BCC think it appropriate to apply the European EMU conception to a purely internal situation in the latter case? We believe that the BCC could have been clearer on this point. It is important that a coherent doctrine be developed in order for the sub-State legislatures to know what the limits of their constitutional powers are in the light of both Belgian and European free movement requirements. However, that development promises to be complicated, as the federated entities might find themselves in a catch-22 situation. For instance, in a recent debate preceding the adoption of a new decree, the Flemish government emphasised that, under the constitutional system of power division, it is permitted to establish a tax reduction only for the benefit of persons resident in Flanders. However, the government also recognised the need to respect its obligations under EU law, most notably the free movement requirements.76 Accordingly, as the care insurance case exemplified, there is a tension between the free movement requirements and the requirements set forth by the constitutional system of power division, more particularly, the principle of territoriality which requires federated entities to remain within their territorial jurisdiction. Fundamental choices will be far from easy to make, as they confront the BCC with a serious dilemma. If the Belgian EMU principle were to be conceptualised in line with its European counterpart, the Belgian constitutional system of power division would risk being ‘entirely eroded’.77 However, if the principle were to be construed differently in these two legal orders, the possibility of reverse discrimination will inevitably pop up, as the care insurance case powerfully illustrated.78

75

See especially Belgian Care Insurance Judgment, above n 26, para B.14. Flemish Parliament, Travaux Préparatoires 2008–2009, No 2012/1, 108. 77 J Velaers, ‘Social Federalism and the Distribution of Competence in Belgium’ in Cantillon et al (eds), above n 48, 156. 78 If the ECJ had followed AG Sharpston’s suggestion of adjudicating ‘purely internal situations’, the problem would not have been resolved. Indeed, under that hypothesis, compliance with the ECJ judgment would have compelled the BCC to modify the Belgian constitutional system of power division. 76

Index Åland Islands, 210, 292 Aleinikoff, Alexander, 347 Alen, André, 394 Alexy, Robert, 352 Amar, Akhil, 113 Aristotle, 57, 58, 59, 61, 62 Arte Channel, 188 asbestos, 77 Australia, 26 Austria: competence disputes, 91, 93 Constitutional Court, 87, 89, 91, 93, 95 equality principle, 336 federalism, 86, 212 nationality law, 33 regional participation in EU law-making Bundesrat, 215 Committee of the Regions, 223 European Council, 220 information, 213, 216 interregional bodies, 216 social federalism, 265 subsidiarity scrutiny, 227 Baltic euro-region, 237 Barroso initiative, 238 Barry, Brian, 179 Belgium: Co-operation Agreement, 197–8, 204–5, 208, 217, 221 Constitutional Court access to, 95–6 collegiate decisions, 89 competence disputes, 91 composition, 88–9 EMU doctrine, 393, 394–404 jurisdiction, 336–7 legislative annulment actions, 91–2 origins, 87, 394 principles, 394 regionalism and, 392–404 constitutional patriotism, 173–4 Constitutional Treaty and, 204–5 ‘declaration on national parliaments,’ 202, 204 federalism, 1, 86, 212, 245, 287 EU free movement, 14, 396–404 exclusive competences, 394 Flemish Care Insurance see Flemish Care Insurance case

Flemish Inheritance Tax case, 402–4 intra-state free movement, 395–6 jurisdiction, 393 overview, 393–4 place of employment, 269 regional powers, 393 social security, 265, 269, 272–6, 284 taxation powers and, 395 Flanders, 171, 178, 182, 183 language, 292, 307 social housing, 307, 311 inheritance tax, 15–16 languages dominance of French, 299–300 Flanders, 292, 307 official languages, 291 Lisbon Treaty and, 204–5 Maastricht Treaty and, 197 naming law, 32–3 nationality laws, 35–7, 117 nationhood, 171 regional participation in EU law-making Committee of the Regions, 223 European Council, 197–8, 221 information, 213 interregional bodies, 217–18 veto rights, 208–9 subsidiarity and, 204–5, 227–8 Bermann, George, 66–7 Blair, Tony, 250 Boom, Steve, 108 Brennan, William, 113, 122, 123 Burger, Warren, 123 Burma, 159 Bush, George W, 163 Calabresi, Steven, 334 CALRE, 234, 246 Canada: constitutional patriotism, 173–4 EC—Asbestos and, 77 International Fund for Ireland and, 26 multiculturalism, 185 nationhood, 171 public culture, 179 Quebec, 171, 178, 179, 182, 183 Supreme Court, 95 Catalonia, 182 children’s rights, 119, 324

406 Index cities, 233 citizenship: EU see EU citizenship liberal nationalism and, 169 moral standing and, 62 nationalism and, 175 Cold War, 158 collective bargaining, 323, 326–7, 341–2 Committee of the Regions, 198–9, 206, 223–4, 239, 245–6, 250 Common Agricultural Policy, 22, 240–1, 344 constitutional courts see national federal constitutional courts constitutional patriotism: critique of liberal nationalism, 175–80 EU and, 184–5 meaning, 171, 173–4 national boundaries and, 188 political community and, 169, 170 constitutional rights: balancing v categorisation American theory, 338–40 ECJ case law, 340–5 barriers to free movement ECJ approach, 323, 324–5 ECJ differential approach to regional barriers, 324–38 contents, 322–3 courts see national federal constitutional courts ECJ duty to respect, 330–2 EU rights and, 125–9, 324–5 individual rights, ECJ preference for, 333–7 societal values, 323 Constitutional Treaty, 46, 141, 204, 207 COREPER, 247 COSAC, 206, 246 cosmopolitanism, 61–2, 63, 187 Council of Europe, 235, 236 Cronin, Ciaran, 175 culture: multiculturalism, 176–7, 185, 242–3 nation and, 171 national liberalism, 169–70, 171–3, 175–80 national pluralism, 170, 180–8 societal culture, 172 Cyprus, 192 Dashwood, Alan, 241–2 Davies, Gareth, 68–9 de Búrca, Gráinne, 66, 67 Delmas-Marty, Mireille, 13 democracy: beyond nation-state, 169 democratic subsidiarity, 66 equal representation, 244 EU deficit, 55, 244, 364 international law and, 70

language and, 289–90, 291 legitimacy and, 70–1 liberal nationalism and, 169, 170 national parliamentary accountability of EU ministers, 200–2 new visions, 234 Denmark, 224 Derrida, Jacques, 169, 179 De Schutter, Helder, 291–2, 335–6 Dworkin, Ronald, 15 ECOSOC, 247 Eeckhout, Piet, 116 environmental protection: US—Shrimp, 77–9 Estella, Antonio, 67, 69 Estonia, 306 EU: citizenship see EU citizenship co-decision procedure, 26 competences, 24–5 see also subsidiarity conferral, 139–42 decisions, 65 exclusive competences, 144–5 implied competences, 144, 147–9 international relations, 136–52, 139–52 Lisbon Treaty, 238 shared competences, 146–7, 243 Constitution, 85–6 nature, 45–6, 242 sui generis entity, 132 constitutional decisions procedure, 207–8 regional interests and, 207–9 decision-making and regional interests, 191–209 constitutional decision-making, 207–9 democratic deficit, 55, 244, 364 federalism see EU federalism free movement see free movement international relations see EU international relations internationalism, 364 law-making see EU law-making non-discrimination see non-discrimination principle principles, 294 Union of peoples, 60–3 EU citizenship: application to member states, 116–21 constitutional patriotism and, 169 EU federalism and, 31–43 EU legal nexus, 31–3, 38, 41, 275 legal basis, 55 liberal nationalism and, 170 McCarthy case, 39–42, 43, 119–20 national pluralism and, 170 proportionality, 38

Index 407 prospects, 184 reverse discrimination, 35, 39, 280 Rottmann case, 15, 32, 33–5, 36, 38, 41, 42 Ruiz Zambrano case, 35–9, 42–3, 104, 116–18, 120–1 EU federalism: autonomous constitutional regions and, 230–48 concept, 134–5 constitutional structure, 2–4 diagonal dimension, 31–43, 44 early meaning, 13 fundamental rights and, 110–30 horizontal dimension, 23–31, 44 international relations and, 131–3, 135–52 jurisdiction and, 53–63 legal basis, 23–31 nature, 45–6, 84–5 subsidiarity and, 65–82 unity and diversity, 13–15, 52 vertical dimension, 15–23, 43–4, 45 EU international relations: categories, 140 CFSP, 132, 140–4, 146–7 competences, 139–52 conferral, 139–42 ECJ case law, 147–52 exclusive competence, 149–52 existence, 143–4, 147–9 implied competence, 144, 147–9 international cooperation, 146 nature, 144–7, 149–52 shared competence, 144, 146–7 treaty making, 145–6, 149–52 US compared, 136–7, 138 CSDP, 141, 144 federalism and, 131–3, 135–52 intergovernmental approach, 132 Lisbon Treaty, 140–1, 142, 143–4 Neighbourhood Policy, 141 principles, 142 terminology, 136 EU law-making: regional participation: application of subsidiarity, 225–8 Committee of the Regions, 198–9, 206, 223–4, 239, 245–6, 250 decision-making process, 219–28 ECJ, 223 European Council, 219–22 European Parliament, 222–3 information, 212–14 internal process, 212–19 interregional bodies, 216–19 liaison offices, 224–5 mechanisms, 214–19 overview, 210–29 preparatory phase, 212–14 upper chambers, 215–16

European Arrest Warrants, 125 European Charter of Fundamental Rights: application to member states, 103–4, 114–16 children’s rights, 119 constitutional rights and, 116–21 free movement and, 329–30 linguistic diversity, 310 Lisbon Treaty and, 102, 112 objective, 119 private and family life, 119 residence rights, 119 European Commission: Barroso initiative, 238 national parliaments and, 203–4 regions and, 246–7 European Convention on Human Rights, 102 European Council: EU constitutionalism and, 86 national parliamentary accountability, 200–2, 203, 204 regional participation, 196–8, 219–22, 247 European Court of Justice: access to, 105–6, 223, 247–8 annulment actions, 98, 106 compared to national federal constitutional courts composition, 104–5 functions, 98–104 overview, 98–107 competence disputes function, 99–100 Kompetenz-Kompetenz, 107, 108 procedures, 105–7 significance, 107 composition, 104–5 constitutional rights and approach, 324–5 duty to respect, 330–2 federal rights compared, 323–38 preference for individual rights, 333–7 dissenting opinions, 105 final authority, 106–7 functions compared to national federal constitutional courts, 98–104 competence disputes, 99–100 federal constitutional court, 98–100 human rights court, 102–4 international court, 102 Supreme Court, 100–1 uniformity of interpretation, 100, 125 fundamental rights constitutional rights, 325–6 development, 98–9 incorporation, 113–21 preference for individual rights, 333–7 hierarchy of rights, 326 Italian Constitutional Court and, 376–7

408 Index jurisdiction, 53–9 EU international relations, 144 language regulation and, 288 error potential, 294–6 Groener, 295, 296 Lithuanian case, 301–6 names doctrine, 302–5 new legal order, 59–61 political safeguards of federalism and, 15 pragmatism, 363 pro-European activism, 101 proportionality review, 314 regions and access, 23, 223, 247–8 appointments, 248 approach, 325–6 balancing v categorisation, 342–5 bias, 346–8, 361 categorical approach, 345–60 differential approach to constitutions, 323–38 recognition, 239–41 scepticism, 330–1, 348 survey, 322–61 tax powers and, 249–64 subsidiarity and, 69, 99–100, 106, 247–8 European Parliament: annulment actions, 26–30 co-decision procedure, 26 composition, 53 national parliaments and, 203, 204 regional representation, 199, 222–3, 246 European Trade Union Confederation (ETUC), 349–50 federalism: bias against, 346–8, 361 concepts, 1–2, 45–7, 110 EU and US, 134–5 constitutional courts see national federal constitutional courts constitutional federalism, 51–2, 53 definition of federation, 86–7 devolutionary federalism, 52 EU see EU federalism historical federalism, 47–50 international relations and, 131–66 member states see regions pluralist federalism, 52–3 shunning label, 2 Feeley, Malcolm, 333, 334 Ferreres Comella, Victor, 100 Finland, 210, 287, 292 Fischer, Joschka, 47–8 Flemish Care Insurance case: Belgian position, 348 categorical reasoning, 342, 352–3, 361 critique, 276–84

free movement and, 400–2, 403–4 internal situations, 354–5, 356–7, 359, 401 judgment, 272–6 justification for interference, 325–6, 337–8 legal uncertainty, 278–9, 285 place of employment principle, 269, 273, 279–81 pre-Lisbon Treaty case, 330 residence requirements, 400 reverse discrimination, 279–81, 285 social security and devolution, 277–8 treating regions as member states, 280–4 foreign relations see EU international relations; international relations France: expulsion of Roma, 103 overseas territories, 245, 255 regional participation in EU law-making, 213, 218 wine-growing, 192 WTO asbestos dispute, 77 free movement: Belgian federalism, 395–404 Charter of Fundamental Rights and, 329–30 collective bargaining and, 323, 326–7, 341–2 constitutional barriers ECJ approach, 324–5 ECJ differential approach to regional barriers, 324–38 ECJ proportionality review, 314 EU legal nexus, 31, 32 fundamental rights and, 114–15, 324–7 intra-state situations, 342–3, 351, 355–6 language regulation and, 293–319 limitation clauses, 341 regional barriers, 275, 323 ECJ approach, 325–6 ECJ differential approach to constitutions, 323–38 social devolution, 266, 325–6 treating regions as member states, 280–4 restrictions by private undertakings, 315–16 social rights and, 349–50 social security and see also Flemish Care Insurance case coordination, 267–8 EU interventions, 266–72, 325–6, 336–7 states of employment and residence, 268–72 fundamental rights: balancing v categorisation American theory, 338–40 ECJ case law, 340–5 collective bargaining, 323, 326–7, 341–2 constitutional rights see constitutional rights ECJ as human rights court, 102–4 ECJ development, 98–9

Index 409 EU Directorate, 103 EU federalism and, 110–30 EU rationale, 37–8 EU rights v constitutional rights, 125–9 free movement and, 114–15, 324–5 incorporation of EU rights in member states EU citizenship, 116–21 non-discrimination principle, 115–16, 241 scope, 113–21 individual rights, ECJ preference, 333–7 moral rights and, 335 United States see United States Gamper, Anna, 87 Gans, Chaim, 176 GATT: Article XX exceptions, 76–80 Germany: children’s rights, 324 Constitutional Court access to, 95 competence disputes, 90, 97 composition, 88–9 dissenting opinions, 89–90 ECJ dialogue, 128 on EU federalism, 84 fundamental rights and, 101 Lisbon case, 244 constitutional legitimacy, 55–6, 61 federalism, 14, 86, 212, 245 human dignity, 336 Maastricht Treaty and, 197 nationality laws, 33–4 regional participation in EU law-making Bundesrat, 215–16 Committee of the Regions, 223 Conference of the Länder, 216 European Council, 220–1 information, 213 Maastricht Treaty and, 237 regional state aids, 252–3 social federalism, 265 subsidiarity and, 225, 227 Gibraltar, 21, 240, 261–3 globalisation, 48–9, 155–6, 187 GMOs, 72 Green, Leslie, 335 Habermas, Jürgen, 48–50, 169, 174, 175, 179, 184 Halonen, Tarja, 210 Held, David, 187 Holocaust, 160, 161 Homer, 322 Howse, Robert, 69–76 human rights see constitutional rights; fundamental rights Hungary, 103, 178

Iceland, 178 inheritance tax, 402–3, 404 International Court of Justice, 163 International Fund for Ireland, 26–30 international relations: EU and US federalism, 131–66 Spain, 383–5 terminology, 136 Ireland, 26–30, 223, 224, 295 Italy: comuni, 368 Constitutional Court annulment actions, 94 collegiate decisions, 89 compatibility of regional law with EU law, 377–9 competence disputes, 94, 96–7, 366, 367–8, 372, 375–6, 378 composition, 88–9 ECJ and, 376–7 procedural impermeability, 363–4, 376–7 procedural rights, 95 region-state relations and EU law, 364–76, 379 subsidiarity, 375–6 counter-limits doctrine, 367 dogmatic legal system, 363 language regulation, Bickel and Franz, 115, 297, 299, 305 municipal tradition, 368 nationality discrimination, 115–16 peacekeeping, 377 primacy of EU law, 377 region-state relations cedevolezza, 371–3 centralisation of powers, 372, 379 competence disputes, 94, 96–7, 366, 367–8, 372, 375–6, 378 conferred powers, 365–8, 372 EU law and, 366–8 subsidiarity, 373–6 substitution power, 368–73 survey, 364–76 regions, 1, 2, 86, 245, 287 autonomous provinces, 220, 292, 308 compatibility with EU law, 377–9 EU regional blindness and, 362–80 language, 292, 308 participation in EU law-making, 213, 220, 222, 223, 224 regionalised state, 212 state relations, 364–76 taxation powers, Sardegna, 263, 282–3, 359–60 Jackson, Vicki, 347 Joerges, Christian, 70 jus cogens, 241

410 Index justice: national pluralism, 183, 185 nationalism and, 175–6 new visions, 234 social justice, 173 Kant, Immanuel, 62, 63, 74 Kelsen, Hans, 100 Kommers, Donald, 89–90 Koopmans, Tim, 13 Korea—Beef, 79–80 Kosovo, 180 Kymlicka, Will, 172, 176, 180, 187, 290, 291, 299, 313, 335 Lacroix, Justine, 179–80 language regulation: ECJ case law, 288 Groener, 295, 296 Lithuanian case, 301–6 names doctrine, 302–5 EU law and federal states designation of official languages, 293–6 error potential, 294–6 evidence of proficiency, 296–7 language facilities, 297, 299 proportionality review, 300 protectionism, 297 residence conditions, 297–300 survey, 293–301 EU law and unitary states, 301–6 federal member states, 287–321 legitimate aims communication, 309, 319 language promotion, 309–10 linguistic diversity, 310–11 private language, 316 public language, 308–11 neutrality impossible EU law and federal states, 293–301 EU law and unitary states, 301–6 official language policies, 292–3 Polish-speaking minority in Lithuania, 301–6 political theory, 289–92 survey, 289–306 neutrality possible legitimate aims, 308–11 proportionality, 311–15 survey, 306–15 private language, 307 EU law and, 315–19 objectives, 316 political theory, 316–19 scenarios, 315 proportionality EU inquiry, 311–15 intensity, 314–15 narrow sense, 312–13

necessity, 311–12 review, 300 suitability, 311 public language, 287–315 rationale, 192 Lazer, David, 67 legal personality, states, 235–6 Lenaerts, Koen, 45, 47, 52–3, 54, 59 liberal nationalism: EU and, 184–5 exclusion critique, 170, 175–80 meaning, 171–3 national boundaries and, 188 political community and, 169 Lithuania: Polish-speaking minority, 301–6 Lugano Convention, 152 Luxembourg, 224 MacCormick, Neil, 242 McDonald, Michael, 318 Madison, James, 110 Malta, 192 Mayer-Schoenberger, Viktor, 67 member states: application of EU rights to EU citizenship, 116–21 prohibition of nationality discrimination, 115–16 scope, 112–21 constitutional autonomy, 235, 241 autonomy test, 255–60, 262 EU treaties and, 250 new interpretation, 244–5 social welfare, 266–7 Spain and, 385 taxation powers, 255–60, 262 constitutional rights v EU rights, 125–9 equality, 244–5, 250, 385 federalism see regions governing monopoly, 233 harmonisation, 152 identity see national identity legal personality, 235–6 parliamentary accountability EU Council, 200–2 subsidiarity scrutiny, 203–7 ratification of EU treaties, regional approval, 208 sovereignty, 251 veto rights, 26–9 Menon, Anand, 60 Miller, David, 184 Müller, Jan-Werner, 175, 188 multiculturalism, 176–7, 185, 242–3 Myanmar, 159 Nagel, Robert, 333, 347 Napolitano report, 231

Index 411 nation states see also member states culture and, 172–3 democratic deficit, 70 identity see national identity meaning of nation, 171 moral standing, 62 national pluralism, 180–8 nations and, 171 reciprocal duties, 62–3 sovereignty, 251 national federal constitutional courts: access to, 95–6 centralist bias, 87, 372, 379 collegiate decisions, 89 competence disputes commonality, 95 ECJ compared, 105–7 function, 87 procedures, 90–4 prominence, 96–7 composition, 88–90 ECJ compared, 104–5 core commonalities, 95–6 dissenting opinions, 89–90 ECJ compared, 98–107 functions, 87 ECJ compared, 98–104 key features, 86–97 legislative annulment actions, 88, 97 political dimension, 89 preliminary references and, 128 regional representation, 88 Supreme Courts, 95 ECJ status, 100–1 national identity, 44, 108–9, 170, 172, 176–8, 183–8, 193, 197, 238, 242, 244 national pluralism: domestic and European, 180–8 EU citizenship and, 170 nationalism: exclusion critique, 175–80 liberal nationalism, 169–72, 175–80 past ideology, 1 nationality discrimination: language regulation and, 293–306 prohibition, 115–16 social security, 269 neoliberalism, 48 Netherlands, 218, 224 New Zealand, 26 Nicolaïdis, Kalypso, 45–6, 69–76 Nimmer, Melville, 358 non-discrimination principle: devolution and, 241, 344 incorporation of principle, 115–16 indirect discrimination, 402 language regulation, 293–306

social security, 269 state aids, 350–1, 355–6 Northern Ireland, 222, 237, 241, 243 Palermo, Francesco, 87 Patten, Alan, 299 PEACE programme, 27, 237 Pescatore, Pierre, 13 Plato, 57–8, 61, 62 pluralist federalism, 52–3, 54, 61 Pollock, Stewart, 122 Portugal: asymmetrical devolution, 21 Azores, 17–19, 239–40, 243, 254–8, 261, 262, 264, 343, 347, 350, 357 Committee of the Regions and, 223–4 overseas territories, 245, 255 subsidiarity principle, 225 Post, Robert, 110 private and family life, 119, 307, 314 public goods, 316–18 Quebec, 171, 178, 179, 182, 183 Rawls, John, 50, 62 Raz, Joseph, 316–17 Réaume, Denise, 317–18 regions: autonomous constitutional regions autonomy test, 255–8, 259–60, 262 meaning, 230 special status, 231, 239–45 terminology, 242 balancing v categorisation, 338–45, 357–60 blindness, 235–6 Italy and, 362–80 treaties, 364 categorical approach to alternative to balancing, 357–60 assessment, 345–60 bias and, 346–8, 361 boundaries, 353–7, 359–60 devolution structures and, 351–3 imperfect knowledge, 346 merits, 345–53 methodological issues, 355–7 minimising errors, 345–6 pitfalls, 353–60 seriousness of errors, 348–9 treaty rationales, 350–1 cities, 233 Committee of the Regions, 198–9, 206, 223–4, 239, 245–6, 250 Common Agricultural policy and, 22, 240–1, 344 communal identity, 332 Council of Europe and, 236 devolution, 249–50

412 Index diversity and complexity, 232–3, 239, 280–4 ECJ and see European Court of Justice ECJ approach to, 325–6, 385–6 bias, 346–8, 361 categorical approach, 345–60 categorical v balancing, 340–5 differential approach to constitutions, 323–38 error potential, 345 imperfect knowledge, 346 survey, 322–61 ECOSOC and, 247 EU constitution and, 15–17 EU law and, 4–10 EU law-making and application of subsidiarity, 225–8 Committee of the Regions, 223–4 decision-making process, 219–28 ECJ, 223 European Council, 196–8, 219–22, 247 European Parliament, 222–3 information, 212–14 institutional mechanisms, 214–19 internal process, 212–19 interregional bodies, 216–19 liaison offices, 224–5 overview, 210–29 preparatory phase, 212–14 upper chambers, 215–16 EU policy evolution, 235–8 1st phase, 235–6 2nd phase, 236–8 3rd phase, 238 accommodation, 242, 245–8 blindness phase, 235–6 myopia phase, 236–8 recognition, 242, 245–8 silent transformation, 241–2 visibility phase, 238 EU principles, 385–6 European Commission and, 246–7 European Council and, 196–8, 219–22, 247 European integration and, 332 European Parliament and, 199, 246 hybrid forms, 135 influence mechanisms, 231 interests constitutional decision-making and, 207–9 direct and indirect defence, 196 EU decision-making and, 191–209 EU treaties and, 193–5 institutional mechanisms, 196–207 judicial proceedings and, 347 political process, 192, 195 judicial bias against, 346–8, 361 Lamassoure proposal, 231 languages see language regulation

liaison office, 224–5 Lisbon Treaty and, 5, 193–4, 195, 200–7, 211, 238, 323, 328–9, 331 Maastricht Treaty and, 196–9, 236–7, 250 member states’ autonomy, 235, 241, 244–5, 250, 255–60, 262, 266–7, 385 Napolitano report, 231 paradox of EU regionalism, 239 proportionality and, 195 relevance, 233–5 respect policy, 185 social regulation, 272–84 state aids and see state aids structure of devolution provisions, 351–3 subsidiarity and see subsidiarity taxation see taxation REGLEG, 231, 234, 238, 247 reverse discrimination: EU citizenship and, 35, 39, 280 language regulation and, 305 social security devolution, 279–81, 285 Ritter, Cyril, 32 Roma, 103 Romania, 180 Rubin, Edward, 333, 334 Schauer, Frederick, 340, 353, 358–9 Scotland, 22, 179, 222, 241, 243, 292, 347 self-determination, 172–3, 176–7, 317 Siedentop, Larry, 49–50 social democracy, 49 social federalism: European strand, 266–72 free movement, 267–72 impact on devolution, 272–84 overview, 265–86 treating regions as member states, 280–4 social rights, 327, 349–50 social security: circles of solidarity, 271–2, 279, 285–6 devolution and EU federalism, 272–84 EU interventions, 272–6 Flemish care insurance see Flemish care insurance case free movement and, 266–72 categorical reasoning, 342, 352–3 coordination, 267–72, 278, 285 states of employment v residence, 268–72, 353 legal uncertainty, 278–9 member state autonomy, 266–7 place of employment principle, 269, 273, 279–81 reverse discrimination, 279–81, 285 Socrates, 57–8 sovereignty: continuing sovereignty, 56 popular sovereignty, 54–6

Index 413 Spain: Andalusia, 387 Aragon, 387 asymmetrical devolution, 21, 243 autonomous communities Constitutional Court and, 381–91 EU law and, 14 implementation of EU law, 385–90 legal basis, 381 participation in EU affairs, 213, 214, 218, 220, 222, 223–5, 228, 383–5 powers, 382 regionalised state, 212 social welfare, 265 Balearic Islands, 388 Basque Brussels delegation, 225 Basque taxation powers, 240, 253–4 UGT-Rioja, 19–20, 23, 258–62, 350, 361 Castilla and León, 387 Catalonia, 178, 182 implementation of EU law, 387–8 Constitutional Court access to, 95 annulment actions, 92–3, 95 competence disputes, 92–4, 96 implementation of EU in regions, 385–90 procedural rights, 95 regional participation in EU, 383–5 regional powers and, 381–91 constitutional patriotism, 173–4 EU and, 382–3 historic territories, 240 international relations, 383–5 naming law, 33 nationalism, 178 Navarre, 387 regional participation in EU law-making CARCE, 218 Committee of the Regions, 223–4 Constitutional Court and, 383–5 European Council, 220, 222 information, 213, 214 subsidiarity scrutiny, 228 regionalism, 1, 86, 245 Constitutional Court and, 381–91 cultures, 180, 182 governance levels, 382 languages, 292, 294 state-region relations, 386 unity principle, 381 Valencia, 387 Spinelli Group, 1 state aids: meaning, 251–2 non-discrimination principle, 355–6 regional powers and, 17–21, 239–40

Azores, 17–19, 239–40, 243, 254–8, 261, 262, 264, 343, 347, 350, 357 early ECJ cases, 252–4 Gibraltar, 20–1, 240, 261–3 Sardegna, 263, 282–3, 359–60 UGT-Rioja, 19–20, 23, 258–62, 350, 361 treaty rationales, 350–1 states see member states; nation states Stein, Eric, 83 Stilz, Anna, 180 structural funds, 27, 234, 236, 237, 250 sub-state entities see regions subsidiarity: competence decisions and, 65 critiques, 65–6 fundamental critiques, 68–9 Howse and Nicolaïdis, 69–76 non-justiciable principle, 69 overview, 68–76 deliberative subsidiarity, 67–8 democratic subsidiarity, 66 ECJ and, 69, 99–100, 106, 247–8 effectiveness subsidiarity, 67 etymology, 68 EU law and, 81–2 executive subsidiarity, 66–7 Italy, 373–6 Kantian account, 74 meanings, 66–8, 373–4 legal definition, 66, 73 principle, 65–82 new formulation, 211 origins, 237 procedural subsidiarity, 67 proportionality and, 243 regions and, 194 application of principle, 225–8 national parliamentary scrutiny, 203–7 scrutiny by national parliaments, 203–7 shared competences, 243 treaties and, 374 Lisbon Treaty, 238 Maastricht Treaty, 237 Protocol, 67, 69, 194, 201–5, 211, 224, 226–7, 374 vertical and horizontal dimensions, 69 WTO, 69–73, 76–81 EC—Asbestos, 72, 77 Korea—Beef, 79–80 US—Shrimp, 77–9 US—Tuna, 80–1 Sullivan, Kathleen, 339, 345 Sweden, 224 Tamir, Yael, 176 taxation: regional powers: autonomy test, 255–8, 259–60, 262

414 Index Azores, 19–20, 239–40, 243, 254–8, 261, 262, 264, 343, 347, 350, 357 Belgian inheritance tax, 402–4 centrifugal trend, 15–16 change of vision, 239–41 early ECJ judgments, 252–4 ECJ case law, 249–64 Gibraltar, 20–1, 21, 240, 261–3 new interpretation, 241–2 new norms, 243–4 Sardegna, 263, 282–3, 359–60 state aids, 17–21, 252–4, 350–1, 357 treaty rationales, 350–1 UGT-Rioja, 19–20, 23, 258–62, 350, 361 Tocqueville, Alexis de, 50 trade unions, 323 United Kingdom: Anglo-Irish Agreement (1985), 26 deportation, 33 federalism, 1, 2, 245, 287 asymmetrical devolution, 21 devolution, 250 Horvath, 22, 241–2, 344 regionalised state, 212 nationality law, 39, 41 Northern Ireland, 222, 237, 241, 243 public culture, 179 regional participation in EU law-making European Council, 220, 222 information, 214 soft law mechanisms, 219 subsidiarity scrutiny, 228 regional powers CAP, 22, 240–1, 344 Gibraltar, 20–1, 240, 261–3 Scotland see Scotland social federalism, 265 Wales, 22, 179, 222, 241, 243, 292 United Nations Charter: US courts and, 163 United States: abortion, 125–6, 127 Bill of Rights centralising force, 112, 129 floor protection, 122–4, 126–7 incorporation, 112–13, 117–18 origins, 112, 113 Reconstruction amendments, 113 state rights and, 122–4, 125–7 Burma sanctions, 159 competences allocation, 153–7 implied competences, 147 international relations, 153–65

Constitution 1787 Constitution, 112–13 absolute terms, 341 n87 constitutional rights theory, 338–40 federalism, 14 foreign relations competences, 153–7 Supremacy Clause, 122–3, 153, 158, 159 consular relations, 162–4 consumerism, 50 democracy, 50 federalism concept, 134–5 Constitution, 14 constitutional rights and, 101 debate, 110, 111, 112–13, 333–4 foreign relations and, 131–3, 135–9, 153–65 jurisdiction and, 54 model for EU, 111, 129 new judicial federalism, 122, 123 foreign relations competences, 137, 138, 153–65 federalism and, 131–3, 135–9 Supreme Court case law, 154–5, 157–65 terminology, 136 free speech, 126 Hyde Amendment, 125–6 International Fund for Ireland, 26 private language regulation, 319 Supreme Court, 95, 100, 123–4, 333–4 US—Shrimp, 77–9 US—Tuna, 80–1 Van Gerven, Walter, 51, 53, 54 Van Parijs, Philippe, 50, 290–1, 299, 311 Vienna Convention on Consular Relations, 163 Waldron, Jeremy, 58, 60–1 Wales, 22, 179, 222, 241, 243, 292 Walzer, Michael, 62 Watts, Ronald, 54 Wechsler, Herbert, 14 WTO: dispute settlement and civil society, 71–2 EC—Asbestos, 72, 77 GMOs, 72 Korea—Beef, 79–80 national treatment, 79–80 SPS Agreement, 72 subsidiarity, 69–73, 76–81 TBT Agreement, 80–1 US—Shrimp, 77–9 US—Tuna, 80–1