EU Law and Economics

Steinbach, Armin Book EU Law and Economics Provided in Cooperation with: Oxford University Press (OUP) Suggested Citation: Steinbach, Armin (2025) : EU Law and Economics, ISBN 978-0-19-892091-5, Oxford University Press, Oxford, https://doi.org/10.1093/9780198920915.001.0001 This Version is available at: https://hdl.handle.net/10419/321951 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/4.0/ EU Law and Economics EU Law and Economics ARMIN STEINBACH Professor of EU Law and Economics, Department of Law, École des Hautes Études Commerciales (HEC) Paris, France Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Oxford University Press 2025 The moral rights of the author have been asserted This is an open access publication, available online and distributed under the terms of a Creative Commons AttributionNon CommercialNo Derivatives 4.0 International licence (CC BYNCND 4.0), a copy of which is available at https:// crea tive comm ons.org/ licen ses/ byncnd/ 4.0/ . Subject to this license, all rights are reserved. Enquiries concerning reproduction outside the scope of this licence should be sent to the Rights Department, Oxford University Press, at the address above. Co-funded by the European Union. Views and opinions expressed are however those of the author only and do not necessarily reflect those of the European Union or the European Education and Culture Executive Agency (EACEA). Neither the European Union nor EACEA can be held responsible for them. Public sector information reproduced under Open Government Licence v3.0 (https://www.nationalarchives.gov.uk/doc/open-government-licence) Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2024947766 ISBN 9780198920885 DOI: 10.1093/ 9780198920915.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Advance Praise Understanding the European Union requires more than studying the rules laid down in the European Treaties. EU Law and Economics opens an insightful and thought-provoking dialogue between two distinct disciplines. Written with lucid erudition, Steinbach deploys economic analysis to illuminate core structural questions about the European legal order. It is an important new contribution to the theoretical and practical understanding of EU law. —Pascal Lamy, Former Director General of the World Trade Organization (WTO) and Vice-President of the Paris Peace Forum EU Law and Economics provides a masterful analysis of the interaction between law and economics in the European Union, and the different perspectives that lawyers and economists have when approaching the why, how, who, and what of integration. The book is destined to become an essential reference for scholars, practitioners, students, and policymakers in law and economics. —Rosa Lastra, Sir John Lubbock Chair in Banking Law and Chair of the Institute of Banking and Finance Law at the Centre for Commercial Law Studies, Queen Mary University of London Armin Steinbach’s research work combines an acutely insightful grasp of the minutiae of the European Union’s legal system with an exceptionally perceptive command of the economics of European policy integration. This masterpiece book will no doubt prove indispensable read for both legal scholars and economists. —Jean Pisani-Ferry, Professor of Economics at Sciences Po (Paris), Senior Fellow at Bruegel (Brussels) and Peterson Institute for International Economics (Washington) In this insightful volume, Armin Steinbach provides an innovative, coherent, and satisfying approach to understanding the rules, processes, and constitutional structure of European integration. This volume shows, to paraphrase Molière in Le Bourgeois Gentilhomme, that for 70 years the EU has been ‘speaking’ law and economics without knowing it. —Joel P. Trachtman, Henry J. Braker Professor of International Law, The Fletcher School of Law and Diplomacy, Tufts University Preface The ambition of this book is not to demonstrate that economics could remedy the insufficient rigour of the EU legal discipline, nor to question the selfsufficiency of legal science. The productive contribution of economics to EU law aims to offer complementary insights into the understanding of EU rules and principles, to uncover empirical premises that are implicit in (normative) legal provisions, and to highlight (in)consistencies between legal and economic views. EU Member States transfer centralized power for some competences, while not for others, inviting us to ask whether their proper understanding should be interpreted through the legal principles of conferral of powers and subsidiarity (law), or through the economic concepts of preferences, spillovers, and economies of scale (economics). EU law oscillates between rigidity and flexibility, provoking us not only to ask what compliance with EU law means (law) but also why compliance occurs (economics); and the actions of institutions created through the Treaties should not only be understood as exogenous events to be measured against a binary ‘law– unlawful’ standard (law) but as an endogenous function of interest, control, and sanctions (economics). More fundamentally, cooperation through law, established by primary law or created through deliberate policymaker decisions in EU secondary law, does not just exist as a predefined positivist legal benchmark (law) but as the outcome of cooperation motives shaping institutional, procedural, and substantive arrangements (economics). With the why of Treatymaking, competence allocation, and rulecompliance being a largely disregarded analytical category for lawyers, the inquiry into motives for state conduct are core for an economic analysis, channelled through preferences and payoffs. Instead of a binary legal standard, economics examines incentives, welfare, efficiency, and distribution. It is the consequentialist orientation of economics that adds value to a lawonly perspective on EU law. By extending the hypothetical homo economicus to the ‘state economicus’, rational choice is a fruitful tool for the study of EU law. Whether economics is the ‘queen of the social sciences’ or whether it just ‘colonized other disciplines by seeming to point the way toward understanding the rational basis of human behavior’1— it certainly offers multiple approaches to study the legal framework in a way that legal scholarship as a norm science fails (and is not intended) to uncover. 1 Liran Einav and Leeat Yariv, ‘What’s in a Surname? The Effects of Surname Initials on Academic Success’ (2006) 20 Journal of Economic Perspectives 175. xiv DetaileD contents 12. Nonconsensual EU law 125 a) EU Treaty optout 126 b) Differentiating integration 131 13. Legislative choices 139 a) Directives versus regulations 140 b) Mutual recognition 143 c) Harmonization 145 14. Enforcement 149 PART IV. WHO COOPERATES UNDER EU LAW 15. The European Council and Council of Ministers 159 16. The European Parliament 165 17. The European Court of Justice 169 18. The European Central Bank 179 19. The European Commission 183 20. Adjudication 187 a) Why international courts at all? 187 b) Efficient breaches of EU law 190 c) Judicial authority: leeway and constraints 193 PART V. WHAT TO COOPERATE ON IN THE EU 21. European Public Goods 201 a) Dealing with tradeoffs 203 b) Identification of European public goods 206 c) Legal options to deliver European public goods 210 Adjusting the scope of competence transfer from Member States to the EU 211 Calibrating decisionmaking majority in the Council 212 Variable participation in integration: ‘club provision’ of public goods 214 Centralization versus decentralization of funding decisions and delivery of public goods 215 Compensation to mitigate preference heterogeneity 217 d) Financing European public goods 218 22. Internal market: economic integration 223 a) Free movement of goods 225 b) Freedom to provide services 227 c) Free movement of workers and freedom of establishment 229 DetaileD contents xv 23. Economic and monetary union 233 a) Fiscal policy coordination and state organization principles 236 State debt in federal states 237 State debt in union states 239 Incentives for debt consolidation in federal and union states 241 b) Debt limitation in the eurozone 245 c) No bailout and monetary state financing 248 Bibliography 251 Index 271 newgenprepdf PART I BASICS Introduction to Part I The usefulness of studying European Union (EU) law based on the categories of why, how, who, and what is not inherently selfevident. Cooperation between states has been examined from a range of disciplinespecific perspectives. By way of example: scholars of international relations have explored ‘cooperation problems’ and ‘characteristics of the states’, using the former term to account for how the interests and constraints that govern the actions of cooperating parties can explain certain design features of international agreements.1 Lawyers, and especially EU lawyers, structure their analysis with recourse to positive legal texts and what they define in terms of relevant institutions, sources of law, and policy instruments.2 Scholars of law and economics, by contrast, have often focused on public international law and international law compliance.3 Taking positive EU law— that is, primary and secondary law, as well as the associated institutional and procedural frameworks— as a point of analytical departure, inquiry into the why, how, who, and what of cooperation paves the way for the integration of various disciplinary perspectives. Some of the ‘cooperation problems’ described in the international relations literature, such as the enforcement or commitment problem, are related to how states use EU law to cooperate: institutional and legal arrangements are tools that represent ‘how’ the benefits of cooperation are obtained. Enforcement issues are also of interest for lawyers and for scholars of law and economics, but they should be distinguished from the underlying motives as to why cooperative relations are established (or eschewed) in the first place. Scholars of international relations interested in how relations are ‘designed’ have focused on five broad dimensions of international organizations (such as the EU). 1 Barbara Koremenos, The Continent of International Law: Explaining Agreement Design (Cambridge University Press 2016). 2 Craig de Búrca, EU Law: Text, Cases, and Materials (Oxford University Press 2020). 3 Andrew Guzman, How International Law Works (Oxford University Press 2008) 22, 33, 71, 119, 183; Jack L Goldsmith and Eric A Posner, The Limits of International Law (Oxford University Press 2006) 23, 83; Joel P Trachtman, The Economic Structure of International Law (Harvard University Press 2008) 72, 119, 208. 2 armin steinBach Treating states as rational actors that shape an international organization (the independent variable), they identify a range of dependent variables, including membership rules; the range of issues addressed; the centralization of responsibilities; rules for controlling the institution; and the flexibility of arrangements.4 Certain features— namely, flexibility and centralization— are of essential importance from an economic perspective, and of relevance for the EU (how and why), while others may be better understood from an actorcentric perspective that considers incentive structures and associated aspects of game theory, including the view that EU institutions are ‘agents’ acting on behalf of Member States as ‘principals’ (who). With a view to ‘centralization’, we find interdependencies between form and substance, for ‘centralization’ not only refers to the empowerment of supranational authorities to monitor, mediate, and sanction, but also to the distribution of substantive competencies between EUlevel institutions and Member States. Thus, our why captures more fundamentally the core motivational drivers for EU members to enter into binding EU law and subject themselves to a constraining institutional framework in the first place. By contrast, the how inquiries into the specific institutional and procedural design of the EU— that is, the way it centralizes responsibilities, maintains flexibility, and legislates and enforces EU law; the who is actororiented and focuses on the economic underpinning of Union institutions acting as administrators, legislators, and adjudicators of EU law, each of them guided by individual incentive structures through (more or less) legally defined control and sanction regimes. Furthermore, the what inquiries into the specific policy fields and competencies of the EU from an economic perspective. EU lawyers typically struggle with the why. EU public lawyers ask questions regarding underlying motivations only in very limited cases (eg to determine the relevant legal basis for public policies); they exhibit almost no interest in the rationale for creating EU primary law in the first place or for crafting the peculiar EU institutional framework. By contrast, asking why is a key research question posed by scholars of international relations and economics. And it is precisely this perspective that we seek to render fruitful for legal analysis. One prospective benefit of such an approach is to furnish economically informed metrics for assessing how competences between the Union and Member States should be allocated. Such an approach, which clearly goes beyond positivist considerations of how competencies are allocated in the Treaties, promises to appeal in particular to legal scholars, given the importance they attach to systematic definitions and doctrinal consistency. By the same token, legal analysis of the instruments set forth by the Treaties can be expanded by asking why and when these instruments should be used. International lawyers traditionally ask questions of who is bound by law or who can invoke certain rights: While the Westphalian legal order from which modern 4 Koremenos (n 1) 42. Basics 3 international law emerged was tied to the stateonly approach, the emergence of international organizations and individuals in the international law arena has given rise to an array of new legal issues.5 International relations scholars have also sought to account for the shift in focus from states to nonstate actors, recognizing the nonstate actors as active agents that make and change rules in global governance.6 Unlike lawyers, who generally consider states to be the sole source of authority (even going so far as to consider domestic affairs irrelevant to international public law7), international relations and economic scholars endeavour to add granularity to the who in international law by breaking up the ‘billiard ball’ and looking inside the state— for example, by applying a political economy perspective that explains the actions of domestic state actors, a perspective relevant to the multilevel and multiinstitutional framework of the EU. Likewise, with methodological and normative individualism as an operational benchmark, economics speaks to the perennial legal issue of who is the subject of legitimacy in the EU, that is, in whose interest the EU should act. Principal– agent considerations enter the analytical sphere of who acts, which is alien to a legal assessment that determines the who with sole reference to the competences, rights, and obligations enshrined in the Treaties. 5 Roberto Virzo and Ivan Ingravallo, Evolutions in the Law of International Organizations (Brill/ Nijhoff 2015). 6 Deborah D Avant, Martha Finnemore, and Susan K Sell, Who Governs the Globe? (Cambridge Studies in International Relations, Series Number 114) (Cambridge University Press 2010). 7 Article 27 of the Vienna Convention on the Law of Treaties 1969. EU Law and Economics. Armin Steinbach, Oxford University Press. © Oxford University Press 2025. DOI: 10.1093/ 9780198920915.003.0001 1 Meandering between rational choice, realism, constructivism, and institutionalism A few words of explanation regarding the methodological arsenal of this work are necessary, that is where it sits on the spectrum between realism and constructivism, between rational choice and behavioural economic scepticism. Building on mainstream economic theory and International Relations literature, this work harnesses rational choice theory as its dominant analytical approach, given its power to account for the behaviour of various actors, whether individuals, firms, nongovernmental organizations (NGOs), or states. The rational choice model presumes that actors seek to maximize their interests in decision situations marked by constraints. This perspective takes EU Member States as the analytical point of departure and suggests that the design of EU institutions, including core legal principles (eg conferral or subsidiarity), are deliberate architectural choices made by Member States as the ‘masters of the Treaties’. Over time, this perspective has proven excessively narrow, as the EU as an organization has evolved to become an actor itself when acting through its institutions— for instance, when it adopts secondary law, when judges of the ECJ perform their judiciary function, or when Union institutions are involved in Treaty amendments, as is perhaps most visible through the supranational character and supremacy of EU law over national law. This makes the EU different from other international organizations which have, from a rational choice perspective, largely been viewed as a derivative function of Member State interests.1 While Treaty changes were ultimately decided and ratified by the EU members, these changes have contributed to the emergence of the EU as an autonomous actor. It is the supranational character of the EU that has emancipated it from typical analytical perspectives in international law, and with this in mind, we must reconsider our assessment of the EU from the standpoint of law and economics, for the EU cannot be studied solely as the expression of Member State will; rather, the EU has been emancipated from being a ‘dependent variable’ to assume the status 1 Barbara Koremenos, Charles Lipson, and Duncan Snidal, ‘The Rational Design of International Institutions’ (2001) 55 International Organization 761; Eric Posner and Alan O Sykes, Economic Foundations of International Law (Harvard University Press 2013) 84. 6 armin steinBach of an ‘independent variable’. This is a methodological extension that may be disruptive to traditional legal perspectives founded on the KompetenzKompetenz of Member States as ‘masters of the Treaties’. To be sure, institutional development in the EU is not solely driven by the deliberate choices of Member States, but also by the autonomous behaviour of EUlevel institutions. To take this institutional autonomy into account is essential, and it requires us to look beyond the EU as a blind agent of Member State will. One important and groundbreaking example of EU autonomy is the legal innovation introduced by the ECJ’s rulings concerning the direct effect and supremacy of EU law, which have been the driver for legal and institutional intervention of the EU into Member States’ legal orders. We will discuss the various rationales that underpin EU cooperation from a rational choice perspective— a perspective that inherently treats the EU as a ‘dependent variable’— but we will also look at the various EU institutions that act as rational choice actors— for example, by examining interactions between EU institutions involved in policy formulation or adjudication from the perspective of game theory. It may be plausible to view the EU as a ‘dependent variable’ during the initial development of the EU Treaties. This is less the case when Treaty amendments occur, because Member States are the pivotal actors under the Treaty revision procedure (Article 48 Treaty on European Union (TEU)). By contrast, with a view to the adoption of EU secondary law, it is plausible to assume that the outcomes depend on the actions and interests of Union institutions, given their indispensable consent for the lawmaking process. The predominant assumption underpinning a rational choice assessment is that state preferences are exogenous, as domestic policy is mostly not taken into account in the law and economics literature, nor in international relations scholarship.2 However, the notion of evolving preferences that has been posited by the constructivist view represents a deviation from the rational choice postulate. In essence, constructivist approaches focus on the social construction of identity between actors in international relations.3 Accordingly, international norms, including international and EU law, have a constitutive function in the sense that they help to give form to the wishes and preferences of states.4 Yet it is not only 2 Koremenos (n 1) 12. 3 In line with critical theorists for a detailed discussion, see Alexander Wendt, ‘Constructing International Politics’ (1995) 20 International Security 71, 71 f.: ‘Critical IR “theory”, however, is not a single theory. It is a family of theories that includes postmodernists (Ashley, Walker), constructivists (Adler, Kratochwil, Ruggie, and now Katzenstein), neoMarxists (Cox, Gill), feminists (Peterson, Sylvester), and others. What unites them is a concern with how world politics is “socially constructed”, which involves two basic claims: that the fundamental structures of international politics are social rather than strictly material (a claim that opposes materialism), and that these structures shape actors’ identities and interests, rather than just their behaviour (a claim that opposes rationalism).’ 4 cf also Martha Finnemore, National Interests in International Society (Cornell University Press 1996) 128: ‘The fact that we live in an international society means that what we want and, in some ways, who we are shaped by the social norms, rules, understandings, and relationships we have with others. These social realities are as influential as material realities in determining behaviour. Indeed, they are meanDering Between rational choice 7 constructivists who point to the changeable preferences of states. Public choice offers a similar understanding by considering interactions with domestic constituencies.5 Notably, state preferences may change as a result of changes in domestic politics. In the case of the EU, the situation is further complicated by the multilevel governance structure that involves several actors interacting with each other in decisionmaking, both horizontally between Union institutions, but also vertically between the EU and Member States, adding complexity and additional strategic aspects to the analysis. Under these circumstances, we must abandon the assumption that states are animated by stable preferences and instead pursue positive analysis of the context in question, rather than employing general models.6 The more public choice considerations add complexity to our understanding and prediction of state interests— that is, the more we construct state preferences as a function of domestic policy issues— the less we can draw general predictions about the conduct of EU states. This is problematic for an economic analysis that seeks to achieve generalizable results. At the same time, assuming that states are unitary actors would preclude such considerations altogether. It thus would appear plausible to adopt both perspectives— that is, to view states as unitary actors animated by concern for national welfare while relaxing this assumption when warranted by contextual factors. Constructivists acknowledge that ideas and beliefs— and not just material interests— determine the behaviour of states, and are therefore relevant in promoting compliance with law.7 In their seminal discussion of constructivist theory, Finnemore and Sikkink assert that ‘norm cascades’ can function as causal chains that induce compliance with international law norms.8 In their view, norms arise when ‘norm entrepreneurs’ try to convince states to accept new norms. In a second phase, a critical mass of states accepts the norms, and they ‘cascade’ down to the remaining states. In the final phase, the norms are accepted and internalized, thus giving rise to a preference for norm compliance. However, rationalists categorically reject this description of the preference formation process. One can use the constructivist logic to explain the sequential spread of EU law into Member State domestic law. Once the ECJ had ruled that EU law had direct effect or supremacy over national law (a judicial innovation not directly derivable from the Treaties), it what endow material realities with meaning and purpose. In political terms, it is these social realities that provide us with ends to which power and wealth can be used.’ 5 Andrew Guzman, How International Law Works (Oxford University Press 2008) 128. 6 Steven P Croley, ‘Theories of Regulation: Incorporating the Administrative Process’ (1998) 98 Columbia Law Review 1. 7 Alexander Wendt, ‘Constructing International Politics’ (1995) 20 International Security 71; Jeffrey T Checkel, ‘Why Comply? Social Learning and European Identity Change’ (2001) 55 International Organization 565. 8 Martha Finnemore and Kathryn Sikkink, ‘International Norm Dynamics and Political Change’ (1998) 52 International Organization 887. 14 armin steinBach maximize the ability of citizens to influence actual decisionmaking.5 In this vein, fiscal federalism theory generally favours local governance as a default position for enhancing democratic participation and accounting for local preferences. Closely related to public goods analysis is the concept of transaction costs, which captures various dimensions of costs associated with cooperation.6 Transaction costs analysis may offer a plausible explanation as to why cooperation within the EU takes place or why it does not.7 Transaction costs can be multidimensional. By way of example, sovereignty costs are incurred whenever states cannot choose their national prerogatives and surrender competencies. In light of the transfer of competences to Brussels, some constitutional courts have argued that specific aspects of state sovereignty should remain under national control. Similarly, negotiation costs occur if EU states must engage in lengthy debate to find unanimity in the Council; contracting costs occur if the complexity of an issue is high; implementation costs occur when there is a burden of complying with EU directives and regulations; and monitoring costs occur when an institutional structure is established to administer an agreement or review compliance with EU law (eg the European Stability Mechanism, or ESM, established during the sovereign debt crisis). The use of transaction costs analysis paves the way for costbenefit analysis that studies the behaviour of EU members seeking various forms of cooperation (eg differentiated rather than full integration). Transaction costs vary depending on a range of factors, including the number of EU states involved, the substance of the issue at hand, the degree of formality, and the institutional environment of the agreement. The EU is based on international treaties, which are the main source of EU law, making contract theory a fertile analytical tool, as it is commonly applied to the assessment of international law.8 The EU is an international organization built on many contracts in the sense that the EU Treaties cover a wide scope of substantive law, procedural rules, and institutional arrangements. Accession to the EU is predicated on the assumption that taking on the rights and obligations associated with membership will generate a positive net payoff. EU Member States delegate 5 Albert Weale and Michael Nentwich, Political Theory and the European Union: Legitimacy, Constitutional Choice and Citizenship (Routledge 1998); Mareike Kleine, Javier Arregui, and Robert Thomson, ‘The Impact of National Democratic Representation on DecisionMaking in the European Union’ (2022) 29 Journal of European Public Policy 1. 6 Building on seminal work, Ronald H Coase, ‘The Problem of Social Cost’ (1960) 3 Journal of Law and Economics 1; Ronald H Coase, ‘The Institutional Structure of Production’ (1992) University of Chicago Law Occasional Paper; Oliver E Williamson, ‘The Economics of Organization: The Transaction Cost Approach’ (1981) 87 American Journal of Sociology 548. 7 Armin Steinbach, ‘The Trend towards NonConsensualism in Public International Law: A (Behavioural) Law and Economics Perspective’ (2016) 27 European Journal of International Law 643, 646. 8 Robert E Scott and Paul B Stephan, The Limits of Leviathan (Cambridge University Press 2006); Joel P Trachtman, The Future of International Law: Global Government (Cambridge University Press 2013) 318. a cursory review of economic methoDs 15 authority over a particular area of concern to the EU in exchange for cooperative gains (such as reciprocal commitments by the other Member States) or for the coordination gains achieved through centralization (such as harmonizing standards). The multiplicity of contracts to which Member States must accede is a distinguishing feature of the EU compared to other international organizations because it widens the number of agreeable solutions through crosspolicy compensation, with asymmetric concessions on one contract (eg in agricultural policy) being compensated by asymmetry in other context (eg market freedoms). While subcontracts produce divergent benefits and costs for Member States, the complementarity of policy fields allows greater positive net payoffs than in narrowly designed international organizations. The questions surrounding optimal Treaty design is a highly salient issue for contracting parties. A contract should be optimal from an exante perspective, that is, it should incentivize the parties to invest in the contractual relationship in order to maximize anticipated mutual benefits. At the same time, however, the parties want to craft a contract that is also optimal ex post, that is, one that maximizes the benefits for parties even after possible uncertainties have arisen. These partially conflicting goals create tension, because each party wants to secure the other party’s commitment, but expost inflexible commitments can compromise the goal of maximizing shared benefits. Specifically, unforeseen circumstances may cause the compliance cost of one party to exceed the benefits that it expected to generate from the contract.9 Treaty specification and expost adjudication of the European Court of Justice (ECJ) are thus key to maintaining contractual gains under conditions of uncertainty. Similar considerations apply to expected compliance with EU Treaties. EU members enter into a contract behind a ‘veil of ignorance’, that is, they do not know who might break the contract. In this situation, they are obliged to make a distinction between, on the one hand, ‘intracontractual flexibility’, which applies to certain behaviour that is due to unforeseen external events and that is not tantamount to a breach of contract; and, on the other hand, ‘extracontractual flexibility’, which refers to a breach of contract due to opportunistic behaviour, and which should trigger corresponding legal consequences. An optimally designed contract permits flexibility in the first case and provides for ‘hard’ law in the second case. While ‘hard’ law punishes those who violate the contract, extracontractual flexibility punishes the party who abides by the contract. Parties can therefore express their contracts in rather vague provisions and delegate the interpretation to third parties (ie EU Commission administration, and ECJ adjudication) in order to ensure the necessary flexibility ex post, at the price of raising the uncertainty regarding the net payoff from cooperation ex ante. This interpretative framework allows us to 9 Scott and Stephan (n 8) 61. 16 armin steinBach study the optimal degree of specificity that contracts should contain, and the functionality of ‘rules’ and ‘standards’ under conditions of uncertainty10— parameters which, in the EU context, are fulfilled by ‘regulations’ and ‘directives’. This has a direct connection to economic arguments for ‘efficient breaches’ of contract, a term that refers to voluntary breaches of contract when adherence would lead to even greater economic losses.11 While such a perspective is incompatible with the binding pacta sunt servanda rule, breaches of EU obligations may be the efficient response when EU Treaties do not offer sufficient flexibility or when other contractual partners prevent an efficient adaption of the EU Treaties as a selfish holdout strategy.12 The interactions between EU states can also be likened to the decisions made by players in a game. Indeed, economic analysis has made extensive use of the techniques furnished by noncooperative game theory. One limitation of game theory is that it precludes consideration of exogenously binding contracts.13 Hence, for international law to be effective, the contractual arrangements must be selfenforcing, because, unlike national law, international law lacks exogenous enforcement tools. In this way, the rules themselves must offer sufficient endogenous incentives for compliance. Compliance is assured because adherence is in the individual interests of the parties to the agreement. With a view to the EU, the cooperation literature shows that cooperation is possible in repeated games.14 The interdependencies that exist between European states, including the Unionwide scope of many policies, entail repeated interaction of a cooperative nature. Incentive compatibility does not require positive net payoffs for all states and compliance on all occasions and under all contingencies, provided the long run benefits for adhering to a specific institutional setup are positive. Sanction mechanisms help to sustain defined institutional arrangements and also promote compliance, but this requires the actual imposition of sanctions when they are warranted. A key problem in this regard is that within international organizations, sanctions are endogenous, thus creating a secondorder public good dilemma. As a consequence, sanctions are underutilized in the international law arena (see below Chapter 14). The EU, which has a supranational character, represents a hybrid case. Compliance with EU law rests on the generation of payoffs in a repeated game as well as retaliation in the event of 10 Louis Kaplow, ‘General Characteristics of Rules’ in Encyclopedia of Law and Economics (National Bureau of Economic Research 1997); Cass R Sunstein, ‘Problems with Rules’ (1995) 83 California Law Review 953. 11 Wenqing Liao, The Application of the Theory of Efficient Breach in Contract Law: A Comparative Law and Economics Perspective (Intersentia Ltd 2015); Melvin A Eisenberg, Foundational Principles of Contract Law (Oxford University Press 2018). 12 Eric Posner and Alan O Sykes, Economic Foundations of International Law (Harvard University Press 2013) 25. 13 Trachtman (n 8) 516. 14 James W Friedman, ‘A NonCooperative Equilibrium for Supergames’ (1971) 38 Review of Economic Studies 1; Drew Fudenberg and Eric Maskin, ‘The Folk Theorem in Repeated Games with Discounting or with Incomplete Information’ (1986) 54 Econometrica 533. a cursory review of economic methoDs 17 misconduct by a third party (EU Commission and ECJ), rather than by the parties to the agreement, which can resolve the secondorder public goods dilemma. The EU has increasingly built up an arsenal of sanction mechanisms, which are largely under the control of the Commission as a third party to Member States (alongside qualified majority voting on the Council). Noncompliance with EU law thus leads to considerable exogenous enforcement which resolves the secondorder collective action dilemma, representing a point of contrast to public international law. As the EU addresses many policy areas and encompasses multiple sectorspecific commitments (subcontracts), one could reasonably presume that various games are underway in parallel.15 From a game theory perspective, one could say that players are bound to each other in multiple ways, including through the links between games.16 Suppose, for example, that Member States interact with each other and with the Commission in a game of fiscal supervision in which Member States have to decide whether to comply or defect. This game may be influenced by behavioural spillovers from a parallel game taking place on a more general level as part of interstate relationships (eg within the European Council, in the interactions between heads of state). In this way, the state of play in an interstate relationship or the likelihood of Member State coalitions (against the Commission) might impact the fiscalscrutiny game and determine its payoff structure. More generally, EU Member States and EU institutions interact in a variety of contexts, with each Member State taking on different roles in each circumstance. In a ‘ruleoflaw game’ a country might take a harsh stance visàvis potential defectors to the EU rule of law while, at the same time, a ‘fiscalsupervision game’ might be underway in which that same EU member is much less concerned with the rigid application of EU rules, and these games could be linked. This potential for interaction between games is at odds with an established legal doctrine: international agreements are considered to be selfcontained regimes that are sufficient unto themselves for settling disagreements, and they also claim primacy over more general law. Accordingly, an issue occurring under the EU fiscal surveillance regime should be dealt with exclusively under the applicable EU fiscal rules— however, in a legal and political union, in which more than one game is going on at the same time, the fiscal dispute is unlikely to be resolved solely based on the application of fiscal rules. The additional discretion that the Commission and Council enjoy is likely to be determined by other games. What is legally designed as a selfcontained regime is influenced in practice by other factors— which is good news for the compliance record under EU law, as the interrelationships between games may induce compliance in a way that makes the larger relationship between Member States selfenforcing.17 15 Trachtman (n 8). 16 Timothy N Cason, Anya C Savikhin, and Roman M Sheremeta, ‘Behavioral Spillovers in Coordination Games’ (2012) 56 European Economic Review 233. 17 Trachtman (n 8) 516. 18 armin steinBach Likewise, game theory may furnish an explanation for the institutional design of the EU. Consider for instance the judicial independence granted to ECJ judges. We can think of judges as participating in a power competition game with other officials. In contrast to the assumptions of positivist lawyers, for whom the Treaty’s stipulation of independence is unquestioned, judges typically have their own preferences regarding the political order.18 The court has a range of possible interpretations that can be brought to a legal text, and judges can satisfy their political preferences by exercising their power to interpret statutes. However, they must fear legislators overruling a court’s judgment and, in the case of the EU, that national courts may not follow ECJ decisions, as domestic courts are obliged to apply EU law as interpreted by the ECJ.19 From this, it stands to reason that that the ECJ will more aggressively apply its judicial discretion when the risk of legislative repeal or conflict with domestic courts is small (see below Chapter 20 c). The EU is based on Treaties that are quasiconstitutional in nature. In this way, constitutional economics offers a fruitful lens for examining EU law. This school of thought, initiated by Nobel prize winner James Buchanan, seeks to abandon the prevailing focus on welfare economics (with its neoclassical ‘maximization paradigm’) as well as the rational choice theory that undergirds realist approaches to international relations.20 By contrast, constitutional economics hinges on the premise of ‘constitutional contract/ exchange paradigms’.21 Individual and collective gains are enabled by constitutional cooperation that improves the ‘laws and institutions’ of the economic and political order and protects the democratic expression of informed preferences. Constitutional economics redirects the focus of economic analysis away from individual utility maximization towards the design of markets and political arenas such that ‘consumer sovereignty’ in markets and ‘citizen sovereignty’ in political domains form the analytical and normative benchmark.22 Normative constitutional economics calls for legal safeguards to ensure that the ‘competitive order’ (based on ‘performance competition’ and price mechanisms) remains embedded in a mutually coherent monetary order (that protects price stability and ensures fiscal discipline); democratic constitutionalism (that holds ‘European network governance’ accountable through multilevel competition, monetary, and other regulatory agencies); and social order (that 18 Jeffrey J Rachlinski and Andrew J Wistrich, ‘Judging the Judiciary by the Numbers: Empirical Research on Judges’ (2017) 13 Annual Review of Law and Social Science 203; Cass R Sunstein, Are Judges Political? An Empirical Analysis of the Federal Judiciary (Brookings Institution Press 2006). 19 Stefan Voigt, ‘Iudex Calculat: The ECJ’s Quest for Power’ (2003) 22 Jahrbuch für Neue Politische Ökonomie. 20 Geoffrey Brennan and James M Buchanan, The Reason of Rules: Constitutional Political Economy (Cambridge University Press 1985); James M Buchanan, ‘The Domain of Constitutional Economics’ (1990) 1 Constitutional Political Economy 1. 21 Armin Steinbach, ‘Constitutional Economics and Transnational Governance Failures’, Constitutionalism and Transnational Governance Failures (Brill/ Nijhoff 2024) 77. 22 Viktor J Vanberg, ‘Market and State: The Perspective of Constitutional Political Economy’ (2005) 1 Journal of Institutional Economics 23, 27. a cursory review of economic methoDs 19 protects labour markets, welfare states, social justice, and judicial remedies).23 Normative constitutional economics criticizes various aspects of neoclassical economics, including its focus on utility maximization and rational rentseeking and the assumption of ‘perfect market competition’ without transaction costs. Gross domestic product as a welfare metric is also attacked given its failure to account for other important measures of human wellbeing, including the universal satisfaction of the populace’s basic needs, the enhancement of ‘human capacities’ (AK Sen), and the protection of constitutional rights. The EU Charter of Fundamental Rights (EUCFR) guarantees civil, political, economic, social, and ‘European citizenship rights’, and thus does not only protect ‘negative freedoms’ (eg constraining abuses of public and private power). Articles 2 and 8 of the European Convention on Human Rights (ECHR) have prompted evergreater number of courts to rule that environmental pollution and climate change infringe on human rights, thus furnishing a legal basis for addressing market failures. Combatting climate change, promoting sustainable development in cooperation with third states, and principles of ‘environmental constitutionalism’ (including the principles of precaution, prevention, and rectifying pollution at its source, as well as the ‘polluter pays’ principle) are included in the EU Treaty provisions on EU environmental policies (eg Treaty on the Functioning of the European Union (TFEU) Arts 11, 191– 193).24 In the arena of international law, Trachtman posits that the allocation of jurisdictional power constitutes an assignment of property rights.25 Accordingly, a wide range of activities by Member States can be conceived as an assignment of property rights, including the agreement on specific rules governing the legality of state conduct, the determination of state obligations and individual rights, and the delineation of competences between states (Union versus Member States) and between institutions (Council versus Commission). States seek to define property rights in order to facilitate coordination and to stabilize expectations. A primary function of property rights according to the literature is to guide incentives for achieving a greater internalization of external costs. Accordingly, a primary function of jurisdictional rules is that of shaping incentives to achieve a greater internalization of external costs among political units. At the same time, Member States, as the initial ‘owner’ of jurisdictional power, engage in property allocation based on whether this allows them to achieve gains. By viewing jurisdictional power as property, we can thus inquire into the optimal allocation of prescriptive jurisdiction in the EU— that is, how to allocate competences to authorize the use of public policies. An interesting conundrum arises in connection with the harmonization of rules within 23 ErnstUlrich Petersmann and Armin Steinbach, ‘NeoLiberalism, StateCapitalism and OrdoLiberalism: “Institutional Economics” and “Constitutional Choices” in Multilevel Trade Regulation’ (2021) 22 Journal of World Investment & Trade 1, 20. 24 ErnstUlrich Petersmann, ‘Constitutional Pluralism, Regulatory Competition and Transnational Governance Failures’ in Constitutionalism and Transnational Governance Failures (Brill/ Nijhoff 2024). 25 Trachtman (n 8) 10, 26. 20 armin steinBach the EU. Specifically, the assignment of property rights to the EU by Member States could be viewed as a means of collusion, as an anticompetitive use of property rights. Indeed, this assignment can be interpreted as an inappropriate restraint on interjurisdictional competition— for example, when the EU sets educational standards for lawyers or doctors, or establishes effectiveness standards for pharmaceuticals. Accordingly, the anticompetitive impacts of collusion in the domain of jurisdiction assignment must thus be taken into account; this is a concern that also arises in fiscal federalism theory (see Chapter 10 a). As a multilevel governance order, the EU is also amenable to principal– agent analysis, a perspective that promises to shed light on the consequences of authority delegation at various levels of government. Here, a collective principal (the Member States) acting on behalf of another principal (national citizens) delegates authority to an agent (the EU) and institutional subagents (EU institutions) in order to accomplish particular purposes. Special attention can be devoted to each element in this multistage principal– agent chain. In this connection, controversy has been triggered by the delegation of authority to nonmajoritarian institutions. Such delegation is typically motivated by one of two rationales: the first is the need for technical expertise (principals delegate certain functions to agents who possess required competencies, but also impose control mechanisms);26 the second rationale is informed by the quest for credible commitments (in this connection, principals deliberately provide a considerable freedom of action to the agent so that this agent can adopt policy to which the principals themselves could not credibly commit).27 Nonmajoritarian institutions such as the Commission, the European Central Bank (ECB), and the ECJ have the following features: they delegate authority to govern specific domains of activity; they are not directly elected; and they enjoy considerable independence.28 For economists, independence in this context means leeway for agents to act autonomously from pressures native to the political economy, in conjunction with the inherent risk that agents will pursue objectives that do not align with the preferences of the principal; associated shirking or selfdealing generates ‘agency costs’.29 Political scientists and lawyers have raised legitimacy as a core concern associated with the nonmajoritarian character of EU institutions.30 Measures designed to reduce legitimacy concerns 26 Giandomenico Majone, ‘Nonmajoritarian Institutions and the Limits of Democratic Governance: A Political TransactionCost Approach’ (2001) 157 Journal of Institutional and Theoretical Economics 57; Giandomenico Majone, ‘Two Logics of Delegation’ (2001) 2 European Union Politics 103. 27 Mark A Pollack, The Engines of European Integration (Oxford University Press 2003). 28 Mark Thatcher, Alec Stone Sweet, and Bernardo Rangoni, ‘Reversing Delegation? Politicization, De‐delegation, and Non‐majoritarian Institutions’ (2023) 36 Governance 5. 29 Eric W Orts, ‘Shirking and Sharking: A Legal Theory of the Firm’ (1998) 16 Yale Law & Policy Review 265; Susan P Shapiro, ‘Agency Theory’ (2005) 31 Annual Review of Sociology 263. 30 Ben Crum and Deirdre Curtin, ‘The Challenge of Making European Union Executive Power Accountable’ in Simona Piattoni (ed) The European Union (Oxford University Press 2015); Deirdre Curtin, ‘Challenging Executive Dominance in European Democracy’ (2014) 77 Modern Law Review 1. a cursory review of economic methoDs 21 by limiting shirking and selfdealing generally establish supervisory arrangements. In this connection, it is necessary to strike an appropriate balance between exante and expost accountability.31 Yet another recommended means of shoring up legitimacy is to ensure a sufficiently specified mandate. In this way, some political scientists have reconceived the principal– agent problem in terms of ‘input legitimacy’ and ‘output legitimacy’.32 The bold normative claims advanced by economists are at least partially to blame for the resistance that has arisen to the proliferation of economic thinking in neighbouring disciplines.33 Accordingly, adequately distinguishing between positivist and normative metrics in economic analysis is far from inconsequential. The positivist conception starts from the premise that economics is primarily an epistemology, not a methodology. The economic toolbox hosts multiple methods, and all of them may be deployed for rational social scientific analysis.34 Properly applied, economic methodologies simply generate relational descriptions. Good economics helps to identify the relationships between law, institutions, and policies and the outcomes that they generate. US President Harry Truman is widely credited with saying ‘Give me a onehanded economist! All my economists say “on the one hand . . . [and then] on the other” ’. Here, Truman was expressing his desire for straightforward economic counsel, rather than an ambivalent discussion of tradeoffs and opportunity costs. In this tradition, economics is a consequentialist and selfconscious tool for analysing social life and the prospective effects of implementing policy measures, one that accepts that normative policy goals should be determined by values outside of the economics discipline.35 Economics is most helpful to law when it offers complementary insight and tools— by studying empirical effects; by revealing the incentives that rules create for states and individuals; by highlighting welfare differences between two states of society; and by disclosing the stability of the legal equilibrium that binds two or more states. By contrast, the normative mode of the economic approach to law seeks to assess the desirability of certain objectives, typically efficiency. Yet it should be relegated to constitutions and lawmakers to determine what normative goals society should pursue. Efficiency is not enshrined in constitutional documents, and legislators rarely pursue efficiency as a primary objective of policy. Employing efficiency as normative benchmark should thus been seen as a scientific exercise, one that 31 Rosa M Lastra and Christina P Skinner, ‘Sustainable Central Banking’ (2023) 63 Virginia Journal of International Law 397, 436. 32 Fritz W Scharpf, ‘Economic Integration, Democracy and the Welfare State’ (1997) 4 Journal of European Public Policy 18; Vivien A Schmidt, Europe’s Crisis of Legitimacy: Governing by Rules and Ruling by Numbers in the Eurozone (Oxford University Press 2020). 33 Liran Einav and Leeat Yariv, ‘What’s in a Surname? The Effects of Surname Initials on Academic Success’ (2006) 20 Journal of Economic Perspectives 175. 34 Trachtman (n 8) 1. 35 ibid 2. 22 armin steinBach allows us to identify firstbest choices in terms of efficiency while also acknowledging secondbest choices, given the primacy of other objectives. Descriptive and normative forms of economic analysis intersect when EU law is investigated using interrogative pronouns (eg who, what, why, how). Over the course of this book, various perspectives are adopted. When we discuss why EU members cooperate through EU law (and when they should), efficiency helps to identify rationales for addressing market failures; for tackling externalities; for unlocking economies of scale; and for reducing (transaction) costs. The how of cooperation is amenable to descriptive economic analysis— it can reveal the economic rationale for flexibility clauses in the EU Treaties; explain when Treaties are incomplete or overcomplete in specifying legal provisions; and explore the incentives arising from EU law enforcement. By contrast, the how may also entail economic normativity— for example, how the subsidiarity test should be applied from an economic perspective; when ‘differentiated integration’ should be the preferred course of integration; and how legislative choices between directives and regulations should be guided. Our discussion of the who of EU cooperation draws on descriptive analysis of public choice theory and principal– agent analysis to discuss the interactions among EU institutions. However, the analysis may be founded on the normativity of efficiency when exploring whether breaches of EU law motivated by efficiency considerations should be permitted in a departure from the pacta sunt servanda principle. Economics is agnostic with regard to the state of the law— it informs the lex ferenda: economics advises how EU Treaties should be amended, how secondary law should be crafted, and how to redesign EU competences in the provision of European public goods.36 De lege ferenda economic advice must be assessed in view of the legal barriers posed by the requirements for Treaty amendment. But economics is also helpful de lege lata when it informs how law should be interpreted— that is, the economic meaning that can be given to legal terms (eg subsidiarity). De lege lata, economic insight must be aligned with the legal canon of interpretation modes, in order to be transferable to the law. The hurdle to introducing economic analysis to the law is the lowest when the law explicitly incorporates economic concepts and benchmarks, such as in EU competition law (eg defining markets) or trade rules (eg determining ‘like’ products). Finally, the what of EU law cooperation addresses policy fields and legal areas that have been neglected in the law and economics literature. Inquiring into European public goods implies the development of a conceptual framework as to how the underprovision of public goods should be addressed— that is, what their economic case is, and what governance options the EU law menu offers to avoid a governance architecture à la ‘onesizefitsall’ for European public goods. With the economic freedoms at the core of the 36 Claeys and Steinbach (n 3). a cursory review of economic methoDs 23 EU integration project, the relationship between individual freedom and public policy intervention is one where efficiency concerns militate against preference heterogeneity. Finally, the Economic and Monetary Union is the field of law in which lawyers and economists have spilled most ink over the past decade, for it raises the fundamental question of how state organization principles interact with debt dynamics and instruments to contain debt. EU Law and Economics. Armin Steinbach, Oxford University Press. © Oxford University Press 2025. DOI: 10.1093/ 9780198920915.003.0003 3 What states and EU institutions care about Exploring why EU members cooperate through legal selfcommitment requires us to clarify what states care about. Whose interests do they pursue when seeking cooperation? A plausible default assumption in line with rational choice is that EU Member States have (wellbehaved) preferences over various goals; the mainstream view in the economic analysis of international law posits that nations tend to pursue their national economic interests while neglecting the interests of foreign states and actors.1 This is a plausible default assumption, one that extends microeconomic modelling of individuals as utility maximizers to the level of collective decisionmakers.2 It is also a pragmatic view in the sense that, irrespective of the obvious fact that even if law is not a direct manifestation of private preferences, individual preferences are also manifested through the mechanism of the state, as state preferences.3 Side streams of economic research have refined the selfish state preference, by adding transnational public goods concerns (beyond a merely national perspective) into the motivation to create a more global welfare perspective;4 or by adopting a perspective that emphasizes powermongering and institutional bias, one that questions benevolent purposes in seeking EU integration.5 Indeed, state leaders may push European integration in pursuit of elitists goals, defying its presumed unbiased connection with citizens’ preferences. In this vein, it is possible that the EU decisionmaking structure enhances the ability of certain political elites or interest groups to achieve their preferred policies at the expense of the broader populace.6 Specifically, the public choice literature on international organizations suggests that domestic policy pursues action through the EU when it 1 Andrew Moravcsik, ‘Taking Preferences Seriously: A Liberal Theory of International Politics’ (1997) 51 International Organization 513, 481; Kenneth W Abbott, ‘Trust but Verify: The Production of Information in Arms Control Treaties and Other International Agreements’ (1993) 26 Cornell International Law Journal 1. 2 Andrew Guzman, How International Law Works (Oxford University Press 2008) 17. 3 Joel P Trachtman, The Economic Structure of International Law (Harvard University Press 2008) 37. 4 ErnstUlrich Petersmann and Armin Steinbach, ‘NeoLiberalism, StateCapitalism and OrdoLiberalism: “Institutional Economics” and “Constitutional Choices” in Multilevel Trade Regulation’ (2021) 22 Journal of World Investment & Trade 1. 5 Roland Vaubel, ‘The Public Choice Analysis of European Integration: A Survey’ (1994) 10 European Journal of Political Economy 227, 232. 6 Joel P Trachtman, ‘Economics of International Organizations’ in Francesco Parisi (ed), The Oxford Handbook of Law and Economics: Volume 3: Public Law and Legal Institutions (Oxford University Press 2017) 506. 32 armin steinBach believes it will achieve its domestic policy goals more effectively, or does so in order to blame the EU for an unpopular domestic policy. Far from being the ‘benevolent dictator’ that some economic models believe the state to be,7 delegation of competences from the state to the EU might allow governments to hide their actions or otherwise avoid accountability for their actions.8 This action may allow states to escape desirable interstate competition in the supply of public goods. A harmonized minimum wage, for example, may serve to maintain competitiveness for some EU members by preventing undercutting from lower wages elsewhere. Thus, while a realist view on state preferences would emphasize its inclination to power and regulatory capture in defiance of citizens’ preferences, the idealist view of state preferences being a pure extension of unbiased aggregated individual preferences does not square well with the rationale of public policy— welfare economics justifies public intervention where markets do not produce efficient results, hence in cases where the outcome of private market decisions are suboptimal. The very justification of public intervention is that private preferences (of those harmed by the inefficient market outcome) are insufficiently accounted for.9 As an analytical starting point, we acknowledge that the peculiar design of EU law is above all the result of deliberate choices of Member States. In focusing on how selfinterested states could cooperate, it is logical to ask what role legal commitments could play. Legal commitments, irrespective whether as substantive obligations, procedural frames, or institutionalized bodies, could be reconceptualized and theorized as arrangements that make cooperation more feasible and durable. Our argument is that the EU institutional design, the choice of its core legal principles (such as the principles of conferral or subsidiarity), the competence allocation for its institutions, or the choice of sources of law reflect deliberate choices by those having an institutional and legal choice in the EU— which are, at least in the historically early stage of the EU, foremost the founding Member States. The evolution of the EU involved many Treaty amendments, with each amendment building on the preestablished institutional setup rather than overhauling the EU’s architecture. With this pathdependent degree of continuity in the EU’s evolution, there remains an institutional learning curve that led through Treaty amendments to deliberate shifts in closer integration. Later and thanks to its supranational guise, the EU itself, through its various institutions and endowed with exclusive competences in some fields, became an autonomous actor, for instance when acting through its institutions by adopting secondary law or through the ECJ as its jurisprudential organ. The EU hence evolved from a ‘dependent’ variable in which it 7 Randall G Holcombe, ‘Make Economics Policy Relevant Depose the Omniscient Benevolent Dictator’ (2012) 17 The Independent Review 165. 8 Trachtman (n 3) 506. 9 However, this applies to negative impact on preferences due to allocative inefficiency. Public policy decisions leading to distributional effects that are disliked by some citizens (and thus misaligned with their preferences) are by nature not aligned with all individual preferences. what states anD eu institutions care aBout 33 was merely the outcome of interstate deliberate choice, to an ‘independent’ variable by being a statelike actor that exerts autonomous influence on other state actors. This makes the EU different from international organizations which, from a rational choice perspective, are largely viewed only as variables dependent on the interests of Member States.10 The claim that EU law is the product of the deliberate choice of Member States even holds when one acknowledges that path dependency limits and determines the institutional evolution of the EU. A distinct question is whose preferences should guide the creation of EU law. The insight offered by Arrow’s impossibility theorem11 suggests that organizations have no rationality of their own and that the process of collective decisionmaking is normatively ambiguous and unstable. EU law and governance establishes several actors on the EU landscape, each of them bound by different institutional interests and restraints and therefore set to pursue different interests. Clearly, rather than assigning the EU as a homogenous rational actor, account must be given to the various Member State and Union actors. Genuine EU institutions that embody EU common interests, such as the Commission and the Parliament, pursue different objectives from the Council that channels Member State concerns. The EU Member States are referred to as the ‘masters of the Treaties’, propagating a legal sovereignty notion that assigns all power to make agreements and create international organizations, a proposition that considers European countries as the ultimate source of authority. However, the outcome of cooperation between nonidentical preferences of Member States does not simply reflect the preferences of the individual state actors, but rather represents their joint efforts and compromises among their preferred outcomes to improve their equilibrium outcome given the strategic circumstances they face. There are different levels of cooperation at which interaction between states and institutions lead to certain outcomes. At the most basic level, Member States agree in intergovernmental conferences on Treaty commitments; in the Council, state representatives interact and strike compromises when crafting EU secondary law; in the EU Commission, interaction between individual Commissioners and directorates produce outcomes; interinstitutional transactions occur between the Council, Parliament, and the Commission. Each of these actors may have specific objectives, and the pursuit of those goals is led by their beliefs about each other’s preferences and the relative costs and benefits of different outcomes. 10 Barbara Koremenos, Charles Lipson and Duncan Snidal, ‘The Rational Design of International Institutions’ (2001) 55 International Organization 761. 11 Kenneth J Arrow, Kenneth Arrow, Social Choice and Individual Values (Yale University Press 2012). EU Law and Economics. Armin Steinbach, Oxford University Press. © Oxford University Press 2025. DOI: 10.1093/ 9780198920915.003.0004 4 The logic of barter trade: Rational choice and constitutional economics Interaction and transaction are at the centre of economics. Rational choice informs us that social and economic interactions are transactions of exchange through which the parties seek cooperation gains.1 Transposed to state level, a state cooperates to the extent it achieves gains through trading commitments. States do not engage in barter trade on one single market. Rather, several markets are used for different aspects of cooperation. The equivalent of the market is simply the forum where states and EU institutions interact— be it in the Council, at summits of head of states, or in the Trilogue involving the Council, Parliament, and the Commission— to cooperate on particular issues in order to maximize their baskets of preferences. A different way to express this trade is to understand cooperation between EU states as a market for trading jurisdiction, with this trade being equivalent to assigning property rights (see below Chapter 10). Goods or services are not the assets traded on these markets, but rather ‘jurisdiction’, defined as the allocation of authority, or the institutionalized exercise of power. Market outcomes of international states, or European states on a regional market for jurisdiction, are thus jurisdiction to prescribe, to adjudicate, and to enforce.2 The trading of jurisdiction occurs on various markets. First, there is an ‘invisible’ and tacit market for certain conduct between states that does not explicitly appear in the black letters of EU Treaties but remains within the realm of tacit agreements. One example would be the Treaties’ silence of certain basic rules of interstate conduct, which are laid down in the UN Charter, but not explicitly mentioned in the EU Treaties. Take as an example the principles of nonintervention, the absence of violence in interstate relations, which rank salient in the UN Charter, but which are not explicitly stipulated in the EU Treaties.3 The validity of these principles may seem obvious in the EU (and tacitly and reciprocally exchanged), due to the greater homogeneity among its members compared to states at a global level.4 1 Joel P Trachtman, The Economic Structure of International Law (Harvard University Press 2008) 10. 2 ibid. 3 There are vague references to the Charter of the United Nations in the EU Treaties but mainly to commit the EU in relation to third countries. 4 One may argue that EU Treaties do implicitly incorporate these fundamental rules of international relations because the UN Charter is binding for EU states as subjects of international law. In terms of norm hierarchy, however, the UN Charter enjoys in many jurisdictions no primacy over domestic constitutions, while the EU Treaties rank above national constitutions. 36 armin steinBach In serving the needs of states, European law is not different from international law in that it can be understood as an attempt to orchestrate cooperation in the face of potential international externalities.5 For instance, the strive for safety plays out in economic beneficial terms in international cooperation more generally, defined as a diversion of anarchy in the interstate relationships in which each state must fear hostilities by its neighbours, forming certainly a key driver at the time when the European Community was born. Security established through a legal order stabilizes expectations and allows states to use scarce resources for more productive purposes than maintaining costly security infrastructure. A basic legal infrastructure increasing security provides then the basis for more cooperative gains, as security produces positive spillovers to many areas of cooperation that would not be attainable if stable expectations were not ensured. What emerged historically as international law building on the ‘Westphalian system’ in terms of state sovereignty, equality of states, and the prohibition of intervention and aggression has been further extended in the EU through its supranational structure, the supremacy of EU law over national laws, and a legal and institutional architecture which fostered a level of economic intertwining that stabilizes the relationship further. Through supranationality, the development of an autonomous legal order claiming prevalence in its Member States’ legal orders, the EU has stabilized expectations on a higher level and reduced the probability of unforeseen hostilities more than on an international law level. In that perspective, cooperation rationales under EU law do not differ fundamentally from international law. Second, and more relevant for this book, there is an interstate market in which commitments are exchanged through explicit and formal (as well as informal) constitutional rules, which ultimately shape the EU as an international organization. EU states as selfregarding units and many states interacting in pursuit of selfinterest create a market where transactions regarding the exercise or renunciation of authority are traded. As an outcome of this trade, states relinquish autonomy in some areas (and craft corresponding authority by EU institutions) in order to obtain certain expected benefits in return.6 Third, following Treaty ratification, there are subsequent markets or submarkets created by the EU Treaties themselves. EU constitutional rules provide rules about how subsequent and subordinate rules, specifically EU secondary law, will be made. On the markets for EU secondary law, institutions, politicians, parties, and even countries compete for political resources of power, ultimately specifying additional commitments within the boundaries of the predefined Treaties. Political institutions (the Commission, the Council, the Parliament) interact, each of them pursuing different preferences and objectives. Fourth, and on a level not leading to the adoption of EU law, within the EU 5 Eric Posner and Alan O Sykes, Economic Foundations of International Law (Harvard University Press 2013) 20. 6 Trachtman (n 1) 10. the logic of Barter traDe 37 institutions, there are submarkets with individual state representatives interacting with each other to craft agreements that secure individual benefits for the parties concerned. As rational choice actors, states employ costbenefit analysis.7 For instance, the principle of reciprocal and ‘mutual recognition’ under EU law (see below Chapter 13 b) allows products that are lawfully marketed in one Member State to be sold in other Member States regardless of whether they comply with the national technical rules of those Member States.8 Likewise, Member States may reciprocally consent to removing domestic obstacles to the internal market through legislative harmonization (Article 114 TFEU). This kind of institutionalized exchange of concessions demonstrates that reciprocal commitments lead to overall welfare gains.9 In an ideal world in which the state leadership pursues overall welfare interests, cooperation ultimately leads to welfare gains. This does not necessarily entail strict Pareto improvements on the domestic level, as international cooperation typically implies distributional effects leading to disadvantages of some (think for instance of protected industries that suffer from trade liberalizing). Welfare economics lets weak Pareto improvements suffice, allowing for the hypothetical compensation of losers of a transaction by those who gain. However, the utilityincreasing trade of reciprocal state concessions do not entirely capture why EU members commit themselves to a supranational design under which each member may experience situations of inferior bargains. States risk being trumped by majority votes in the Council or foregoing the adoption of unilateral measures in cases where domestic preferences would ask for such measures, for example where a merger between companies would be desirable from a domestic perspective but impossible to attain due to exclusive competence of the EU Commission. Cooperation in the EU does not safeguard gains in each individual transaction— EU policymaking, as well as judicial acts of the ECJ, may conflict with national interests. Losses may even be more frequent than gains. Constitutional economics may make sense of this selfrestraining and occasionally selfharming design of rules. Developed by James Buchanan, constitutional economics claims that a fundamental inconsistency exists between the methodological individualism as the classic trademark of the economic approach to social phenomena and the whole concept of a social welfare function. While the traditional focus of economics lies on voluntary market exchanges, constitutional economics extends the ‘mutual gains from trade’ notion to voluntary cooperation more generally understood, including arrangements for collective action, private and public.10 7 Barbara Koremenos, The Continent of International Law: Explaining Agreement Design (Cambridge University Press 2016) 29. 8 Case C8/ 74 Dassonville (1974) EU:C:1974:82 837. 9 Kyle Bagwell and Robert W Staiger, The Economics of the World Trading System (The MIT Press 2004). 10 James M Buchanan and H Geoffrey Brennan, What Should Economists Do? (Liberty Fund Inc 1979) 27. 38 armin steinBach It focuses on the question of how people may realize mutual gains by their voluntary joint commitment to rules, that is, from jointly accepting suitable constraints on their behavioural choices.11 The difference between welfare economics and constitutional economics perspectives on cooperation is that, under the latter, the benefits that participants can expect from such constitutional commitments are not derived from specific anticipated outcomes, but are the overall benefits that result over time from having the continuing process of interaction and cooperation bound by suitable constraints.12 We can speak of ‘enabling constitutionalization’ in reference to rules, which allocate authority to produce rules (like the EU Treaties), but do not foresee the outcomes. There is a veil of uncertainty as to the distributive outcome, because the distributive consequences are unknown in advance.13 The constitutional economic perspective thus illuminates the peculiar design of the EU’s supranational institutionalization. On the market for the trade of jurisdiction, EU members have renounced the exercise of their sovereignty in a number of fields, and the ensuing legislation produced by Union institutions does not guarantee a favourable outcome for any individual EU member. Power has been transmitted to the Commission acting in the Unions’ interests, which in many cases conflicts with national interests, and even if EU members retain legislative power through the Council, governments have surrendered the paradigm of the Westphalian international legal order that had meant to pursue cooperation strictly in line with national interests. Under the EU supranational design, there is by no means a mechanism for Member States to ensure gains in every case. Institutionally salient is this through majority voting in the Council, with decisions potentially taken against a country’s will. And even where unanimity is the predominant decision mode, EU members depend on the Commission for a legislative act to be initiated, they rely on the Commission and the ECJ to enforce mutually agreed obligations rather than reverting to unilateral enforcement, and they may fear Union institutions interpreting or applying EU law in a fashion that contravenes national interests. We should thus be wary of treating states like individuals as homo oeconomii. In line with Buchanan’s diagnosis, nonmarket collective choices in political arenas cannot be determined from the level of individual human action. Society cannot be treated as if it were a choosing entity the same way as individuals decide on markets— rather, it should be seen with its own value scale, thereby abandoning the individualism of the classic economic paradigm.14 It remains to be determined in whose name constitutional economics requires EU constitutional arrangements 11 James M Buchanan, The Economics and Ethics of Constitutional Order (University of Michigan Press 1991) 81. 12 Viktor J Vanberg, ‘Market and State: The Perspective of Constitutional Political Economy’ (2005) 1 Journal of Institutional Economics 23, 27. 13 Trachtman (n 1) 256. 14 James M Buchanan, Freedom in Constitutional Contract (Texas A&M University Press 1977) 235. the logic of Barter traDe 39 to be made. For social arrangements to be ‘socially beneficial’, Buchanantype constitutional economics perspectives require the arrangements to be mutually beneficial, that is, beneficial to all parties involved. Buchanan insists that what may count as ‘better’ in social matters can ultimately only be judged by the persons involved themselves and that, therefore, the relevant test for what qualifies as a ‘good rule’ must be seen in the voluntary agreement of the parties involved.15 To the extent that constitutional economics emphasizes the acceptability of the process that leads to a decision rather than the outcome, it connects to the social psychology literature that points to procedural fairness as what matters for the individual more than the specific outcome.16 With Buchanan remaining faithful to normative individualism for the determination of what should be considered as ‘mutually beneficial’, it was Vanberg who extended this concept to be tantamount to ‘citizen sovereignty’. It captures that the political process should be institutionally governed by the common interest of citizens. Citizen sovereignty thus entails that ‘producers of politics’, politicians and government bureaucrats, are responsive to citizens’ common interests. This contrasts with the legal Westphalian view that emphasizes state sovereignty and state will as reference points for international law,17 as well as with the literature that equates state action with the corporate preferences of the states.18 For constitutional economics it is not the state but rather the citizens’ and consumers’ preferences that form the benchmark for assessing the efficiency of outcomes. The constitutional economic emphasis on citizens’ preferences easily aligns with the traditional legitimacy proposition of constitutional lawyers that traces public authority neatly back to citizens’ votes, albeit with the important difference that Buchanan’s concept does not tie public decision to national citizens but to European citizens. National constitutional law puts the national citizen at the core of democratic analysis, as it is the national citizen who holds domestic politicians accountable.19 Assessing European integration through the lens of national constitutional law thus tends to emphasize the interests of the rather homogenous group of national citizens, a view that is also at the core of the economic theory of federalism, which emphasizes the alignment of homogenous preferences and the supply of public goods (see below Chapter 10 a).20 Constitutional economics defies both the narrow national perspective as well as the legal homogeneity proposition 15 James M Buchanan, Fiscal Theory and Political Economy: Selected Essays (University of North Carolina Press 1954) 122. 16 Lita Furby and others, ‘Public Perceptions of Electric Power Transmission Lines’ (1988) 8 Journal of Environmental Psychology 19; Thomay Priestley and Gary W Evans, ‘Resident Perceptions of a Nearby Electric TransmissionLine’ (1996) 16 Journal of Environmental Psychology 65. 17 Claire Cutler, ‘Critical Reflections on the Westphalian Assumptions of International Law and Organization: A Crisis of Legitimacy’ (2001) 27 Review of International Studies 133. 18 Trachtman (n 1) 37. 19 Andreas Voßkuhle, Europa, Demokratie, Verfassungsgerichte (Suhrkamp 2021) 50, 58. 20 Posner and Sykes (n 5) 15. 46 armin steinBach comply with it. In a prisoner’s dilemma, states have different preferred outcomes and an incentive to defect. Where the distributional implications of an agreement are small, bargaining costs will also be relatively small. In situations where the distributional implications are large— think of the design of the EU agricultural policy or, from an institutional perspective, the setup of voting weights in the Council— bargaining costs will likely to be high.17 In two contexts, negotiations costs are particularly sizeable. First, strategic behaviour is natural conduct in EU relationships. In economic terms, this means one party has incomplete information regarding other countries’ preferences, as strategic behaviour may induce parties not to reveal their preferences. For example, negotiations among eurozone members on fiscal rules may be driven by different motivations. Country A may ask for lenient rules, but Country B does not know whether Country A wants to engage in unproductive fiscal expenditures or whether it seeks to retain leeway for reasonable investments or policy space for stabilization purposes in times of crisis. Similarly, when EU members agree to introduce a carbonborder adjustment mechanism levying nonEuropean imported goods, EU countries may not reveal whether this tool serves to protect its domestic industry or is for genuine climate policy purposes. Second, uncertainty extends to the state of the world.18 When EU members agree on introducing a financial transaction tax, there is a lack of predictability regarding the effect of the tax on various economic indicators. Limited experience with similar ways of taxation will make it more costly to write complete contracts to deal with every contingency.19 Introducing a financial transaction tax with only few EU members may have undesirable sideeffects due to capital movements between participating and nonparticipating countries. In all these instances, countries may want to invest both resources and time in improving their knowledge pertaining to other parties’ preferences or to reduce the degree of uncertainty about the future state of the world.20 Enforcement problems may occur in various guises and cause enforcement costs. EU law is, like all law produced by international organizations that lack nationlike enforcement structures, under risk of not being complied with by EU members. This introduces different perspectives from enforcement or managerial theories, which draw on different reasons for why states comply with or disregard rules in international organizations. Managerial theories refer to a nation’s capacity or inability to comply.21 Barriers to enforcement may be technical in nature and refer to difficulties owing to the complexity of the substance and challenges in 17 James D Fearon, ‘Bargaining, Enforcement, and International Cooperation’ (1998) 52 International Organization 269. 18 Koremenos (n 3) 39. 19 Joel P Trachtman, The Economic Structure of International Law (Harvard University Press 2008) 174. 20 Koremenos, Lipson, and Snidal (n 3) 782. 21 Abram Chayes and Antonia H Chayes, ‘On Compliance’ (1993) 47 International Organization 175. reDucing transaction costs 47 implementation across legal, technical, or institutional dimensions. Lack of bureaucratic institutions or cumbersome legal requirements are examples of enforcement problems. More competent bureaucracies are more likely to implement EU directives without delay, but the domestic structure of decisionmaking— namely veto players at both the national levels and the subnational levels— are equally important predictors of delay.22 Enforcement theories, in line with rationalist and realist approaches, centre on nations’ material incentives to comply with or shirk from international law implementation. Enforcement costs depend on the strength of individual actors’ incentives to cheat on the agreed rules.23 One EU member may prefer not to adhere to it because it can do better individually by defecting from the agreed rules. The enforcement problem arises when actors find unilateral noncooperation so enticing that they sacrifice longterm cooperation. Take, for example, Hungary’s persistent violations and nonobservance of the rule of law principle, the ensuing persistent backandforth in addressing the EU Commission’s request, and the continuation of these violations despite the financial sanctions imposed by withholding funds. The gains of noncooperation may be multifold. When they occur as political gains in the domestic policy arena helping the incumbent government to improve its reelection chances (while accepting overall welfare losses), it is particularly difficult for the EU to influence such gains because sanctions may have a harmful effect not only on the government but also on citizens (eg by withholding funding for EU regions in need as practiced under the budget protection tool24). Closely related are monitoring costs occurring when compliance with EU law must be observed or when an institutional structure is established to administer an agreement. The EU Commission as ‘guardian of the Treaties’ has a genuine monitoring function to the extent that it scrutinizes compliance with EU law and eventually may bring an infringement procedure before the ECJ. On other occasions, additional monitoring bodies may be implemented outside the existing Union structure, such as the EFSM established during the euro crisis to administer financial assistance. The number of participating countries may influence monitoring costs as well as the different monitoring needs of formal and informal formats of cooperation. However, EU law can also be genuinely designed or interpreted (by the ECJ) to reduce monitoring costs. The mobilization of the individual as an enforcer is a case in point. Endowment by the ECJ to invoke EU law directly before national courts decentralizes the monitoring of EU law adherence— individuals detect and tackle national measures incompatible with EU law, which significantly reduces monitoring costs at the EU Commission level. Likewise, innovations of 22 Katerina Linos, ‘How can International Organizations Shape National Welfare States? Evidence from Compliance with European Union Directives’ (2007) 40 Comparative Political Studies 547. 23 Koremenos, Lipson, and Snidal (n 3) 776. 24 Antonia Baraggia and Matteo Bonelli, ‘Linking Money to Values: The New Rule of Law Conditionality Regulation and Its Constitutional Challenges’ (2022) 23 German Law Journal 131. 48 armin steinBach the ECJ, such as liability of EU Member States for incompliance with EU law (the Francovich doctrine25), further strengthens the decentralized and costreducing monitoring. Modification costs are highly relevant with regard to adapting current primary or secondary law to new (factual, legal, or political) circumstances, and may vary depending on the binding nature of the agreement (eg by transforming intergovernmental EFSM law into Union law through the modification of Article 136 TFEU). Member States cannot write complete treaties due to a lack of knowledge and punitive bargaining costs. Yet, modification of treaties through its cumbersome procedure for Treaty amendments is likewise costly. A fully fledged treaty change requires a convention involving national parliaments (Article 48(2)– (5) TEU), while Article 48(6) TEU offers a simplified modification leaving the decision to the European Council. EU members may therefore want to install a transformative clause below the level of Treaty amendment, which allows the EU to react flexibly to new challenges or when it uncovers any gaps in the EU Treaty. Article 352 TFEU is an example of a rule that mitigates the modification costs by offering alternatives to formal treaty changes. Modification may occur through different procedural and institutional steps that involve different actors, with the general rule that the more fundamental the modifications are the more EU and national actors must give their consent in order to secure legitimacy. 25 Joined cases C6/ 90 and C9/ 90 Francovich, Bonifaci and others v Italian Republic (1991) EU:C:1991:428 l05357. EU Law and Economics. Armin Steinbach, Oxford University Press. © Oxford University Press 2025. DOI: 10.1093/ 9780198920915.003.0006 6 Supplying public goods and addressing external effects One could conceive of a world without the EU, possibly even without international law. In such world of autarky, goods and services would be produced and consumed by people in the same state, and no state would take actions affecting other states’ citizens. Likewise, political decisions would be limited to inwardlooking effects that do not affect other countries. In such a context, European states would have no incentive to engage in cooperation, less so through binding selfcommitment. However, in practice there are multiple relations and interactions between states, companies, and citizens, and the nature of these relationships offers a rationale for cooperation through EU law. In a world of increasing economic, cultural, and social interconnectedness, what were conceived as national public goods have turned into transnational public goods, leaving national states struggling to effectively supply them.1 EU law then serves the function of ensuring sufficient supply of EU public goods, addressing externalities, and promoting a level of conduct and commitment that serves to achieve cooperation gains from an overall European perspective. The increase in European surplus can then be divided among participating states to make them all better off.2 The logic of public goods is well known: with nonrivalry and nonexcludability as characteristics of public goods, collective goods such as a clean environment or security are suffering from collective action problems. Freeriding on the climate commitments of others is the rationale behind binding environmental agreements, solving the problem of underprovision of climate efforts. These efforts culminate in EU law and international law, with EU law being the more specific and more binding reduction commitment than the international law climate obligations under the Paris Agreement. However, for the purpose of understanding cooperation on public goods within the EU, we consider the traditional definition of public goods to be too restrictive, as there are other coordination failures that can lead to the underprovision of goods compared to its optimal level that might require public intervention. We adopt a slightly broader definition: a public good 1 ErnstUlrich Petersmann and Armin Steinbach, ‘NeoLiberalism, StateCapitalism and OrdoLiberalism: “Institutional Economics” and “Constitutional Choices” in Multilevel Trade Regulation’ (2021) 22 Journal of World Investment & Trade 1, 1. 2 Eric Posner and Alan O Sykes, Economic Foundations of International Law (Harvard University Press 2013) 13. 50 armin steinBach is simply defined as a good that is not supplied at an adequate level without public intervention (which could take various forms including direct provision, government expenditure, or regulation) due to coordination problems (not only nonexclusion or nonrivalry, but also, eg, network effects). This broader definition captures further incidences of public good cooperation which do not fall under the narrow definition of public goods: ‘pecuniary’ externalities (ie externalities that run through prices and do not constitute market failures) can offer a rationale for EU cooperation. Public investment in an EU member producing crossborder spillovers are felt in other EU Member States as well and may require coordination to optimize their effects from an overall EU perspective. Also, we consider ‘club goods’ to be of relevance (which do not fit the traditional definition because they are not nonexclusionary). Consider the European Economic and Monetary Union (EMU) as a ‘club’ public good. Price stability in the eurozone and fiscal stability are public goods to which all countries should contribute, while, however, leaving incentives to freeride on the other countries’ efforts of fiscal prudence— a dilemma that the EU seeks to resolve by fiscal rules and credible nobailout clauses. Hence, even though club public goods would allow for the exclusion of certain members from the club good (recall the debate on ‘Grexit’), this may, however, not be a viable economic and political option. Efficient club goods thus require incentivizing club members to contribute to the supply of public goods, with the privatization of the benefits of the club accruing to club members offering a strong incentive to join the club. Again, the euro is an example. Clearly, sharing a common currency on the internal market offers economic benefits. Limiting access to these benefits by requiring that prospective members converge their economies towards viable economic levels reveals a logic of incentivization that ensures the effective supply of public goods. Supranational cooperation, such as that of the EU, may suffer from being insufficiently cooperative where a ‘weakestlink public good’ is concerned. Weakestlink cooperation can be traced to Oliver Williamson’s transaction theory.3 He sees vertical integration as a governance response to high asset specificity. In our context, high asset specificity occurs when a Member State makes investments in a good whose value depends on the cooperation of other states. Hence, for a weakestlink public good to generate benefits, all members must invest in the supply of the good— if this does not happen, any state’s investment is wasted. In these cases, there is no compensation possible in the sense of a country making additional efforts to compensate for another state’s reduced efforts. Take policy areas as diverse as the fight against terrorism in the EU, the fiscal conduct of states in the euro currency zone, or the border control of European states fighting illicit trade. In these cases, the Treaty effort level and the number of EU members are complementary in the 3 Oliver E Williamson, The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting (The Free Press 1985). suPPlying PuBlic gooDs anD aDDressing external effects 51 sense that both commitment at the Treaty level and commitment to enforcement positively increase the value of the good concerned. Regarding the fight against terrorism, for instance, in an interconnected EU (without internal borders), the lack of compliance of only one EU member may undermine the value of security for all other members. Likewise, in fiscal policy, the economic misconduct of only one country may destabilize the entire euro area irrespective of the performance of other countries; inadequate measures in the area of banking regulation may undermine financial stability through contagious effects on financial markets; and if EU border control is lax in only one Member State, illicit trade may occur to the detriment of other EU states. In all these cases the supply of the public good suffers, for other EU countries cannot compensate for the underperforming state, hence creating a ‘weakestlink’ public good. With weakestlink public goods being serious concerns, the value of the early ECJ jurisprudence on primacy of EU law over Member States law becomes evident— it ensures consistent application of EU law in all Member States, and it precludes Member States from applying laws that are potentially harmful to the European public good provision. There is thus an ambivalence from an economic perspective: on one hand, EU law primacy ensures consistent protection of EU public goods preventing Member States from defecting from their commitments. On the other hand, uniform application of EU law overrides divergent Member States’ rules, which can be criticized for ignoring diverse preferences by setting rules indistinctively at a central level, hence causing a mismatch between rules and preferences (see Chapter 10 a)). Unlike ‘weakestlink public goods’, ‘substitute public goods’ allow states to compensate for the insufficient efforts of other European states. If European country A lowers its climate ambition and emits more CO2 than it is permitted to under CO2reduction paths of EU climate law, country B could increase its efforts to compensate for it, as it does not matter from the perspective of climate protection who reduces the CO2 emission. Yet, the core challenge also remains with substitute goods, where a country may have an incentive to invest too little and thus free ride on the efforts of other countries. In addition, the equal commitment of countries A and B may be inefficient if the costs of avoiding CO2 emission is cheaper in one country than in the other (hence favouring a costoriented approach to CO2 reduction as implemented through the Emissions Trading Scheme). Closely related to the provision of public goods, externalities offer a straightforward case for entering into binding EU law commitments. In an interconnected and globalized Europe, the actions in or by one state have implications for the wellbeing of citizens in other states. In such settings, an international externality occurs. In economic parlance, many of these externalities are nonpecuniary, in that they do not travel through the price system. The initial motivation for establishing the European Community is probably the most obvious example of negative nonpecuniary externalities motivating the forming of the Treaties of Rome. Fears of 52 armin steinBach territorial expansion, (trade) wars, political threats, and unchecked political power motivated the EU founding members to submit to selfcommitment through the EU Treaties. Other externalities are pecuniary, meaning that the effects on others are felt through changes in price. In that regard, the EMU is a field of multiple pecuniary externalities. Soon after the inception of the European Communities, members identified decentralized monetary policies as a source of instability for other EU members, channelled as pecuniary effects through the price system, which led to the establishment of the European Central Bank (ECB): the effects of Germany (or other large EU members) pursuing expansionary fiscal impulses (positive pecuniary externality) felt in neighbouring countries; domestic state aid distorting competition on the internal market and harming competitors from other countries; national banking regulation affecting domestic banks’ operations, which in turn feeds through EU capital markets into other countries; and low wage development undercutting the competitiveness of other European countries (negative pecuniary externality), perceived as a ‘beggarthyneighbor’ policy. More recently, the EU debt crisis revealed another form of ‘beggarthyneighbour’ policy that resulted from Member States inflicting financial instability on other euro members through fiscally unsound actions. The EU has adopted various legal instruments to deal with externalities. Not only has externality been mitigated through treaties and secondary law by way of implementing directives and regulations— for example, prohibition on inflicting harm on other countries; state aid rules precluding negative effects on foreign competitors; regulations curtailing freeriding on other members’ efforts in the pursuit of climate goals; and fiscal rules disincentivizing EU members from freeriding on the fiscal solidity of other states. At the level of jurisprudence, the ECJ also engaged in the removal of externalities by ruling that not only are discriminatory measures forbidden (which obviously follows from EU law) but also any measures that could potentially hinder the free trade of goods and services (the socalled Dassonville formula4). By eradicating even nondiscriminatory barriers to interstate commerce, the Court abolishes (pecuniary) externalities that regulatory measures impose on other EU members. In other words, the idea of an internal market without barriers incorporates the elimination of externalities of domestic regulation on other states (eg health). From the perspective of European integration, externalities have been addressed in rather a nonuniform manner. While mitigating the negative external effects from company market power or the distortive effects of state aid has been brought rather early under the EU legal umbrella in the integration process, addressing externalities and freeriding in fiscal affairs became imminent when 4 Case C8/ 74 Dassonville (1974) EU:C:1974:82 837. suPPlying PuBlic gooDs anD aDDressing external effects 53 creating the EMU. Other externalityprone policy fields such as security and defence policies remain in Member States’ control and thus are sources of potential externalities (even though cooperation and loyalty requirements apply in these domains). With Article 42(7) TEU obliging its members to act in solidarity when military threats from external actors occur, domestic military expenditure creates positive externalities. A member investing in defence and military capacities increases the overall security for all EU members. In social policies, yet another area of remaining Member State competences, externalities may result from different policy designs, which may create incentives for workers to migrate between EU Member States. Just as with wage policies, social policies can be designed in ways that are perceived as ‘beggarthyneighbour’ policies by other states. With positive externalities just as relevant for EU regulation as negative externalities, the EU internal market offers an example of network externalities at work. Adopting new technological and regulatory standards, more cooperation and participation in the standards increase the value of the standards for all participating countries. By definition, network externalities occur when benefits to all members increase when new members join. International organizations are built on the concept of network externalities. This may be, for example, the case for trade or dispute settlement.5 The more members that join the EU customs union and the EU internal market, the more efficiency gains are possible. Not only do incumbent EU members benefit from the spread of the network, but network externalities also create incentives for the participation of nonEU states, thereby increasing the potential benefits for both new (and existing) EU members joining the standard (and it ultimately lowers new members’ costs to incorporate the growing acquis communautaire as a precondition for accession to the EU). The larger the market and the greater the participation in the EU’s standards, the higher the EU’s global weight in standard setting. Hence, Member States have an incentive to expand standard setting exercises in the EU. Enforcement costs for goods characterized by network externalities are fairly low, as the risk of defection of Member States or third states by deviating from the agreed standard is unlikely. This is because we are in a coordination game (not in a prisoner dilemma situation), offering an endogenous incentive to comply with standards. The network externality effect may not only promote trade between EU insiders, it may also increase the EU’s weight geopolitically, hence strengthening the EU in international organizations and in the international arena more generally. Clearly, the advantages of acceding must be weighed against the costs. Transferring trade policy competence to the EU creates positive network effects for members, but there are sovereignty costs associated with Member States 5 David Epstein and Sharyn O’Halloran, ‘Sovereignty and Delegation in International Organizations’ (2008) 71 Law and Contemporary Problems 89; Joel P Trachtman, The Future of International Law: Global Government (Cambridge University Press 2013) 28. 54 armin steinBach losing their autonomous representation at the World Trade Organization (WTO). Dilution of sovereign representation in international organizations (when the EU represents its members) or loss of domestic regulation or standardsetting autonomy add to sovereignty costs. Also, while overall efficiency gains exist with new members joining public goods with network effects, distribution of benefits may vary and affect certain groups of society more than others, which gives rise to internal domestic resistance. More importantly, the EU comprises cooperation across many policy sectors and goods. Some may produce network externalities (trade or financial regulation), while others exhibit weakestlink public goods (eg security, migration). This inconsistency across goods may give rise to diverse patterns of cooperation and to differentiated integration— the Schengen Agreement and euro membership are examples. Treaty optouts and optins as well as enhanced cooperation (Article 20 TEU) are governance options under EU law to account for different public goods enshrined in the EU (see below Chapter 12). EU Law and Economics. Armin Steinbach, Oxford University Press. © Oxford University Press 2025. DOI: 10.1093/ 9780198920915.003.0007 7 Leveraging economies of scale With the EU Member States shifting the level of public power from the national to the supranational level, economics examines the benefits of supranationalization. Economies of scale offer a normative benchmark— tested and widely applied in economic scholarship to federal states1— to assess the benefits of allocating policy responsibilities at the centralized level. These benefits must be held against the costs of harmonization that accrue mainly where onesizefitsall regulation is adopted despite heterogeneity of preferences across the regions.2 Balancing the benefits from economies of scale with the varying preferences of the citizenry, the optimal degree of centralization should follow one logic: all functions where economies of scale offer compelling benefits should be held at the central level, whereas all functions with high heterogeneity of preferences should be kept at the local level.3 Economies of scale are the paradigmatic economic tool offered to assess the subsidiarity principle— the core Treaty principle laid down in Article 5(3) TEU governing the competence allocation between Union and Member States where shared competences are concerned (unlike exclusive EU competence, where the subsidiarity principle does not apply). The legal subsidiarity principle has been notoriously criticized by legal scholarship for not being based on consistent logic, as well as for the perception that Union institutions— the Commission and the ECJ— have interpreted subsidiarity in a fashion that favours centralization.4 This criticism in legal scholarship has been echoed in the political domain, positing that the alleged lack of subsidiarity is tantamount to an intrusive EU encroaching on Member States’ competences. 1 Gordon Tullock, ‘Federalism: Problems of Scale’ (1969) 6 Public Choice 19; Jacques LeBoeuf, ‘The Economics of Federalism and the Proper Scope of the Federal Commerce Power’ (1994) 31 San Diego Law Review 555, 565. 2 Stephen Weatherill, ‘Why Harmonise?’ in Takis Tridimas and Paolo Nebbia (eds), European Union Law for the TwentyFirst Century: Rethinking the New Legal Order (Hart Publishing 2003) 11, 13. 3 Francesco Parisi and Vincy Fon, The Economics of Lawmaking (Oxford University Press 2008) 52. 4 Catherine Barnard, ‘What the Keck? Balancing the Needs of the Single Market with State Regulatory Autonomy’ (2012) European Journal of Consumer Law 201; Stephen 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; Dieter Grimm, Constitutionalism: Past— Present— Future (Oxford University Press 2006) 303; Gareth Davies, ‘Subsidiarity: The Wrong Idea, in the Wrong Place, at the Wrong Time’ (2006) 43 Common Market Law Review 63, 64; Nicholas W Barber, ‘Subsidiarity in the Draft Constitution’ (2005) 11 European Public Law 197; Peter M Huber, ‘Das Kooperationsverhältnis zwischen BVerfG und EuGH in Grundrechtsfragen’ (1997) Europäische Zeitschrift für Wirtschaftsrecht 517. 62 armin steinBach Heterogeneity in preferences, benefits, and costs make EU accession rounds increasingly challenging, because each accession of new EU members make Pareto improvements more difficult to achieve. On one hand, what entices incumbent (as well as acceding) members are the benefits from an enlarged area of common rules, internal market, and geopolitical weight. In addition, some legal instruments are in place that make new EU accessions sufficiently attractive for countries that extend the club advantages of EU membership to other states. Most relevant, by shifting the burden of adaptation onto the shoulders of the acceding states who must adopt the entire acquis communautaire, incumbent members are less likely to lose on the existing level of integration. On the other hand, the dilutionary effect of EU expansion and the heterogeneity effect count as costs for incumbent members. They must accept a dilution of voting rights in the Council as the number of members increases. More generally, with heterogeneity increasing through diverging preferences, it becomes harder to agree between states. As long as there were five founding members of the European Community with more or less similar, or at least complementary interests, members were able to agree in substance in the limited fields in which Member States initially dealt with cooperatively. Yet, as the number of members increased, the heterogeneity within the group also rose. New members are qualitatively different for various reasons, be it culturally, economically, or socially. While heterogeneity limits cooperation, issue linkages may generate new opportunities for mutually beneficial arrangements. The EU expanded both in its number of members and in scope and depth of integration, which facilitated agreements because sidedeals and compensation in unrelated fields became possible. There is, however, a limit to simply expanding the scope of issue areas, as increased scope comes at higher (bargaining) costs. Strategic behaviour may lead some actors to ‘hold up’ the agreement to gain additional benefits. The risk of unravelling, whereby failure in one issue may lead to failure in all linked issues, is then greater.10 Accessions are thus a mixed endeavour: an EU growing in membership increases network and coordination benefits across policy fields for all participating states (incumbent and acceding states), but it also exacerbates diversity and hence makes the landing zone in terms of Treaty rules that satisfy all states’ preferences smaller. High enforcement costs can be an obstacle to cooperation where compliance with EU law is weak. In principle, one of the features distinguishing the EU from international organizations is its effective enforcement structure with infringement procedures and the role of the Commission and the ECJ as enforcers (see below Chapter 14). However, in other areas, enforcement mechanisms are weak and thus invite defection and rule breaches. In the EMU, this has been a perennial issue.11 The EU debt crisis was (at least partially) caused and exacerbated by a 10 Koremenos, Lipson, and Snidal (n 1) 786. 11 Armin Steinbach, Economic Policy Coordination in the Euro Area (Taylor & Francis Ltd 2014). why cooPeration fails 63 lack of compliance with preagreed fiscal rules and led many external observers to question the character of the EU as a supranational role model for other regions.12 The climax of the sovereign debt crisis led the EU to introduce conditionality as a prerequisite for crisis support in order to improve enforcement. Hence, in cases of fiscal transfer and with a view to mitigating moral hazard concerns, the EU has reverted to conditional financial assistance on the delivery of structural and fiscal reforms as an instrument to safeguard the longterm benefits of EMU membership, which would be at risk if individual EMU members could destabilize the common good by opportunistic behaviour. The ensuing discussion on the breakup of the euro area and ‘Grexit’ illustrates how cooperation may collapse over insufficient compliance with what is required by law. Put differently, by giving the EU weak enforcement institutions in EMU affairs, EU members decided to accept high enforcement costs in order to minimize sovereignty costs (through strict fiscal surveillance). More than once throughout the history of the EMU, political considerations trumped legal (and economic) rules— the initial entry requirements for states to join the euro were interpreted leniently, and the Stability and Growth Pact was not rigorously enforced. On a conceptual level, as each EU member determines the payoff for noncooperation as a function of the credibility of sanctions attached to noncompliance, there is a public good dilemma where enforcement costs may be determined by the strength of individual actors’ incentives to cheat on agreed rules.13 Managerial theories of compliance look at it differently.14 Take the Greek flagrant noncompliance with EU fiscal rules when statistical miscalculations were hiding that actual budget deficits were much higher than publicly reported.15 Managerial theories would liken data misreporting and institutional shortcomings to the lack of bureaucratic capacity. In any case, as members of the EMU, all countries benefit from a stable currency as club public goods and factual risksharing smoothing bonds spread, but for this positive spillover to materialize requires investment in markets’ trust through fiscal prudence. In the past, rules failed to constrain disincentives for contributing to the common good of a stable currency.16 Despite a restraining set of EU rules, Member States continue to have an incentive to freeride on the policy efforts of other states. If a Member State is certain that a desired collective action will happen anyway, even without their particular contribution (eg financial stability in the euro area, security efforts in the EU), they have an incentive to leave it to others to deliver the common good. Besides the 12 Charles Wyplosz, ‘Europe’s Quest for Fiscal Discipline’ (2013) European Economy— Economic Papers 2008– 2015 No 498. 13 Koremenos, Lipson, and Snidal (n 10) 776. 14 Katerina Linos, ‘How Can International Organizations Shape National Welfare States? Evidence from Compliance with European Union Directives’ (2007) 40 Comparative Political Studies 547. 15 Benny Andersen, ‘The Crisis in Greece: Missteps and Miscalculations, European Stability Mechanism’ (2020) Discussion Paper Series No 9. 16 Steinbach (n 11). 64 armin steinBach EMU, foreign policy offers another example. Small Member States may be particularly prone to freeriding. While they participate in decisionmaking on international developments, they assume barely any responsibility for concrete action, with their societies unprepared to bear the costs and risks of operational engagement. This conduct has been observed in relation to establishing a common financial base for EU military operations. In case of security challenges, for instance in North Africa, it has been left to bigger countries with regional ties to the region to deal with security challenges, while many other states have tended to freeride on these efforts, contributing little towards setting up an effective EU funding mechanism.17 While the unanimity requirement has been a perennial problem for the adoption of EU secondary law, the extended introduction of qualified majority voting has allowed the Union to move from Pareto improvements (unanimity voting) to weaker Pareto and KaldorHicks solutions (qualified majority voting). Depending on the individual case, those countries that are overruled in the Council may face significant costs outweighing the benefits accruing to other EU members, but the level of thresholds for qualified majority voting (55 per cent of Member States vote in favour and support by Member States representing at least 65 per cent of the total EU population) reduces the instances in which an EU act fails to offer any kind of welfare improvement. The two voting thresholds tend to secure that, from an EUwide perspective, there is at least a possibility that the gains resulting from the enacted secondary law exceed the costs. However, the EU integration process is characterized by making majority voting more frequent, which on one hand is desirable because it makes the impact of national regulatory capture or biased decisionmaking less far reaching compared to unanimity. On the other hand, however, extending qualified voting reduces the willingness of incumbent EU members to accept new members. With each new member, majority thresholds are harder to achieve and blocking opportunities are reduced, thus creating problems that would not exist under the unanimity requirement. A complicating factor is that when negotiating Treaty amendments or adopting EU secondary law, each EU country is facing different restraints on the EU level than they are on the domestic level. We recall Putnam’s twolevel games to make cooperation work.18 At the EU level, national governments seek to maximize their own ability to satisfy domestic pressures, while minimizing the adverse consequences of foreign developments. When negotiating EU Treaties between 17 Stefan Lehne, ‘Is There Hope for EU Foreign Policy?’ (Carnegie Endowment for International Peace 2017). 18 Robert D Putnam, ‘Diplomacy and Domestic Politics: The Logic of TwoLevel Games’ (1988) 42 International Organization 427, 434; Joel P Trachtman, The Future of International Law: Global Government (Cambridge University Press 2013) 43. why cooPeration fails 65 state representatives on the interstate level, bargaining costs matter for finding an agreement that is agreeable to all state leaders. At the national level, domestic groups weigh their interests by pressuring the government to adopt favourable policies, and politicians seek power by crafting alliances with those groups, while complying with domestic legal requirements. Sovereignty costs and implementation costs are a function of the domestic policy constraint.19 Member States can concede to EU partners only what is domestically feasible, that is, the political costs of implementing the EU deal at home must not be too high. The negotiators of EU law thus need to address the concerns of domestic interest groups and, at the same time, reach an agreement that is acceptable for other EU members. In certain circumstances, greater domestic policy constraints can lead to more bargaining advantages in international negotiations in Brussels. A purely welfarist perspective emphasizes an equilibrium where welfare increasing policies for national states overlap. In a world without public choice, states find compromises and accept secondbest solutions (from national perspectives) which may lead to a cooperative equilibrium (even if this is inferior to the optimum as long as it is superior to noncooperation). However, domestic politics constraints determine possible equilibria.20 In that sense, greater domestic policy constraints strengthen the position in negotiations on European cooperation because an EU member cannot be expected to violate the domestic policy constraint.21 Cooperation is thus prone to failure when domestic political welfare (based on the public choice approach) deviates from actual welfare. The state is a dynamic aggregator with parties, social classes, interest groups, and public opinion as determinants of domestic policy constraints.22 To appraise the relevance of the domestic policy restraints one should acknowledge that EU Member States may, as posited by rational choice, pursue a welfarist perspective (maximizing national welfare), but at the same time pursue domestic policy welfare. The latter builds on the liberal theory of international relations, which prioritizes the demands of individuals and societal groups over politics.23 A state’s preferences are thus not necessarily rational in their pursuit of maximizing objective welfare, but simply the aggregation of individual preferences by the state’s political mechanisms.24 With domestic policy constraints in mind, cooperation through EU law is generated (by 19 Katerina Linos, ‘A Theory of Diffusion Through Democratic Mechanisms’, in Katerina Linos (ed), The Democratic Foundations of Policy Diffusion: How Health, Family, and Employment Laws Spread Across Countries (Oxford University Press 2013). 20 Anne van Aaken and Joel P Trachtman, ‘Political Economy of International Law: Towards a Holistic Model of State Behavior’ in Alberta Fabricotti (ed), Political Economy of International Law: A European Perspective (Edward Elgar 2016) 9. 21 Trachtman (n 18) 46. 22 Putnam (n 18) 432. 23 Andrew Moravcsik, ‘Taking Preferences Seriously: A Liberal Theory of International Politics’ (1997) 51 International Organization 513, 516. 24 Trachtman (n 18) 42. 66 armin steinBach crafting EU Treaties or making EU secondary law)— in line with Putnam— when there is sufficient overlap between equilibria in Brussels and those in members’ capitals. Moreover, one should consider the interaction between the welfarist (rational choice) rationale for cooperation and the domestic policy rationale. The very function of cooperation on the EU level is to create outcomes that enhance domestic welfare— either by increasing domestic objective welfare or improving domestic policy outcomes. In other words, cooperation on the EU level widens the scope of possible domestic equilibria that are superior to those available in the absence of EU law. While increases in objective welfare (based on the cooperation rationales discussed above) are desirable, public choice informs us that governments may also make EU laws for the sole purpose of generating domestic policy welfare (even when going against a welfarist benchmark). EU law may be produced in order to satisfy domestic political preferences stabilizing the position of the acting government in the domestic context, irrespective of actual welfare. EU law is effective from that perspective if it changes domestic politics (by stabilizing the incumbent position). An important factor countervailing solutions favouring domestic policy welfare over objective welfare is the use of qualified majority voting as the predominant decision mode in the Council. This decision mode reduces the number of countries and occasions in which domestic policy welfare trumps the objective welfare function of EU law. In many cases, this twolevel requirement does not play out in a prohibitive manner: a government may agree to concessions on the EU level when the adoption of an EU directive or regulation is at stake. Despite domestic policy resistance, the government may consider other types of benefits to be sufficiently high in the interest of overall welfare gains. What weighs in is that domestic policy costs often materialize with certain timelags, as EU regulations take effect after a delay, for example when EU directives require Member States to take implementing measures. The effect of such timelags is that governments may blur their accountability for an EU decision to which they agreed in the Council. The timelag between a Member State’s responsible action (eg the adoption of a directive in the Council) and the materialization of political costs (domestic implementation) allows national politicians to externalize some of the costs that would otherwise hinder the agreement on an EU level— this somehow relaxes the constraint under Putnam’s twolevel game. In other cases, however, the domestic policy level may be an insurmountable requirement of the twolevel game. Take EU Treaty amendments under Article 48 TEU requiring not only an accord between EU governments but also ratification, with the negative referenda in France and The Netherlands offering a case in point. Governments were able to come to an accord on the European level (hence on one of Putnam’s two levels), but the solution found at EU level did not overlap with the range of acceptable solutions under the domestic policy constraint. Since the referenda were indispensable for the EU Treaty amendments to enter into force, the domestic policy constraint was prohibitive and ultimately hindered the entering into force of the EU Constitutional Treaty. why cooPeration fails 67 There is evidence to suggest that information provided to ordinary persons that a policy is required by international law shifts opinion in favour of that policy.25 Member States can thus facilitate the implementation of EU directives by adequately communicating to their constituents that compliance with EU law is required. A different logic with similar effect unfolds when governments use the delegation of governance tasks to third parties in order to avoid blame once policies become contested.26 Member States’ governments can benefit by attributing unpopular decisions to EU requirements. Legislating unpopular domestic laws that implement binding EU directives can lead to a blame game at the expense of the EU, while proving highly useful for enabling national politicians to keep political costs with EU compliance low. Domestic policy constraint may interact with a welfarist perspective in the particular case of externalities, which as discussed above provide a strong rationale for (European) cooperation. However, the domestic policy situation may be such that voters have an interest in externalizing costs by imposing them on other countries. This may lead to an equilibrium in which the desirability of cooperation is independent of the magnitude of spillovers. The rationale— as modelled by Loeper— is that crossborder cooperation (in the EU) requires the policymakers of one Member State to provide more public goods— and thus its voters to pay higher taxes. Therefore, the higher the externality, the higher the costs of cooperation on the voters of that particular Member State. This puts reelection at risk because voters will prefer lower taxes in combination with higher externalities undermining the cooperative regime.27 Uncertainty may hinder cooperation, not only where benefits or costs are hard to predict but also where EU members are reluctant to disclose information that could make them more vulnerable. Uncertainty is reduced through institutions, notably the EU Commission and the ECJ acting as bodies that disclose information from Member States (see below Chapter 20 a)). Involving neutral brokers representing Community interests (or a neutral Member State’s interest like the President of the Council) can enhance certainty and facilitate informationsharing. There may also be behavioural economic factors at work that undermine scope for cooperation. Psychologists and economists have constantly investigated the systematic heuristics and biases contradicting the rationality assumption, in order to establish a more realistic model of human behaviour.28 While European law has 25 Adam Chilton and Katerina Linos, ‘Preferences and Compliance with International Law’ (2021) 22 Theoretical Inquiries in Law 247. 26 Tim HeinkelmannWild and others, ‘Blame Shifting and Blame Obfuscation: The Blame Avoidance Effects of Delegation in the European Union’ (2023) 62 European Journal of Political Research 221. 27 Antoine Loeper, ‘CrossBorder Externalities and Cooperation among Representative Democracies’ (2017) 91 European Economic Review 180. 28 See, in particular, Daniel Kahneman and Amos Tversky, ‘Prospect Theory: An Analysis of Decision under Risk’ (1979) 47 Econometrica 263; Gerd Gigerenzer and Daniel G Goldstein, ‘Reasoning the Fast and Frugal Way: Models of Bounded Rationality’ (1996) 103 Psychological Review 650. 68 armin steinBach yet to be systematically analysed via behavioural economics,29 some behavioural economic insight may be intuitive, in particular where it may impede cooperation. Loss aversion may play a role where EU cooperation is associated with risks and losses rather than with potential and opportunities.30 Risks can become salient, creating sentiments of losses. If a country’s state leaders are subject to heuristics, this may bias an objective costbenefit analysis, and perceived losses may lead them to reject an agreement that would increase overall welfare.31 As discussed, sovereignty costs are crucial for a rational country to decide whether it agrees to surrender competences to the EU or not. If a country suffers from a strong bias by overemphasizing the loss, this will compound sovereignty costs and hence entail a reduced willingness of the country to enter into the agreement. Loss aversion may well be used as a tool of political manipulation. As ‘framing’ plays an important role in political communication, politicians may try to generate loss aversion in their constituency in the pursuit of their political agenda, a propagandistic technique that populist parties frequently employ. For instance, fostering fear of losing one’s own currency to trigger resistance or opposing the Schengen Agreement in order not to lose control over borders are among the issues that can easily be framed to trigger behavioural economic insight. Framing may also play out in the internal market context, with regulation promoting liberalization of trade within the Community or with external states as an example. Consider the impact of trade concessions, where liberalization can be framed either to impact ‘employees’ or ‘consumers’. While the employee frame would illuminate the risks such as loss of employment due to increasing competition, the consumer perspective highlights the benefits for consumers through lower prices and a greater variety of products.32 Yet, loss aversion may also lead to excessive weight being placed on the loss of transferring competences to the EU. Concerns may be multifold: will resources be squandered in bureaucratic excess by the Commission? Are EU institutions likely to transgress their competences, hence undermining the principle of conferral? States view centralization warily, and for the same reasons states are also concerned with maintaining tight control over institutional arrangements so as to minimize agency costs, which, however, is difficult to ensure. There is no mechanism through which individual members could easily tackle a competence breach by the Commission. While the annulment procedure under Article 263 TFEU allows Member States to review EU legislative acts, this possibility runs void if the ECJ is biased towards integration itself (see below Chapter 17). There is a 29 For international law, see Gregory Shaffer and Tom Ginsburg, ‘The Empirical Turn in International Legal Scholarship’ (2012) 106 American Journal of International Law 1. 30 Surveys show that public opinion does not typically align with a country’s optimum welfare strategy, see ‘Americans on Globalization: A Study of US Public Attitudes’ (2000). 31 Russell B Korobkin and Chris P Guthrie, ‘Heuristics and Biases at the Bargaining Table’ (2004) 87 Marquette Law Review 795. 32 Anne van Aaken, ‘Behavioral International Law and Economics’ (2014) 55 Harvard International Law 421, 457. why cooPeration fails 69 mass of literature elaborating on the restrained scrutiny of the ECJ over the EU Commission.33 From a public choice perspective, both the EU Commission and the ECJ have aligned interests as their importance increases with the scope of EU law. Limited monitoring capacities on the part of Member States and the lack of sanctions for the agent’s selfish actions weigh as costs for current EU members and prospective losses for future EU members. Often, states overvalue the costs of their trade concessions significantly, to the extent that realizing an agreement becomes impossible.34 Meanwhile, the actors often overestimate the value of their own concessions and underestimate the value of their adversary’s commitment during a negotiation. This situation makes bargaining more difficult and can lead to an impasse.35 It is comparatively more difficult to achieve an agreement when bargaining over the allocation of losses compared to bargaining over the distribution of gains. This often manifests in a permanent struggle over EU Treaty reforms, where the transfer of authority to the EU and perceived losses of sovereignty amount to a bargain over losses, which are clearly identifiable and immediate, while the corresponding gains through integration are vague and uncertain in terms of their actual materialization in the future (eg through future secondary law). Hence, from a behavioural perspective a great deal of information depends on the ‘framing’ of an agreement. Policymakers are inclined to be more reluctant when states perceive an agreement rather as loss (eg domestic industries ‘losing’ market shares), despite overall welfare gains (eg through higher consumer rents). Meanwhile, policymakers might tend to agree when perception is dominated by profit (eg when accession to EU Treaties is dominated by a narrative of gaining access to EU institutions and markets). Drawing on behavioural economic insight, EU Treaty reforms are more likely to be achieved when the benefits for consumers and exporters are the dominant narrative in public perception influencing the government’s choice architecture, as they are among the groups that typically benefit from deeper (economic) integration. In turn, governments whose focus is the protection of the working class (rather than the business sector) may be inclined to underscore the impact on workers in the importing or protected industries.36 Further, a status quo bias might also play out. For individual states, consent avoids welfare losses. Consequently, ambitious Treaty reforms— such as the Constitutional Treaty in 2004— which aim to make profound changes to the structure of the EU may not be concluded because of a bias for the status quo.37 The 33 Wolf Sauter, ‘Proportionality in EU Law: A Balancing Act?’ (2013) 15 Cambridge Yearbook of European Legal Studies 439, 450. 34 van Aaken (n 32) 468. 35 Jack S Levy, ‘Prospect Theory and International Relations: Theoretical Applications and Analytical Problems’ (1992) 13 Political Psychology 283, 290. 36 Bruce E Moon, Dilemmas of International Trade (Routledge 2018) 29. 37 van Aaken (n 32) 457. 70 armin steinBach status quo bias may be particularly relevant where potential benefits are uncertain giving rise to ambiguity aversion. While traditional fields of EU Treaty reforms (eg enhancing internal markets and removing barriers to trade) allow exante quantification of costs and benefits, future effects in other areas are far less certain (eg social policy coordination, institutional changes to decisionmaking rule in Council, and lifting the EU Treaties to become a ‘Constitution’). Major Treaty changes implying institutional changes aiming to overhaul the current practice and engender unprecedented substantive obligations and even threaten national cultural roots by introducing an EU flag or anthem are prone to failure due to ambiguity and loss aversion. However, the effects of aversion can be diminished by avoiding salient reference points. Consider the differences between the failed EU Constitutional Treaty and the successful Lisbon Treaty. The latter avoided being considered a constitution and hence removed all constitutionlike features from the reform agenda (eg explicit supremacy of EU over national law; EU symbols such as a flag or anthem; relabelling EU ‘regulations’ as ‘laws’). However, in substance the Lisbon Treaty maintained many of the core ideas of the Constitutional Treaty, in particular the institutional (eg revised decisionmaking in the Council; a permanent European Council president) as well as substantive changes (eg abolishing the EU’s socalled threepillar structure). The EU’s evolution from a failed constitution to the accepted Lisbon Treaty was hence effective in responding to patterns of loss and ambiguity aversion. PART III HOW TO COOPERATE UNDER EU LAW Introduction to Part III There are many legal variations of cooperation across EU members, with a differing scope of participation, variable bindingness of commitments, and customized degree of centralization. The Treaties foresee, for example, enhanced cooperation and multispeed integration; they provide for optouts in Treaty commitments; in some areas, the EU reverts to soft law rather than hard law; some EU competences are allocated to Brussels as exclusive competences, while others are subject to a subsidiarity test to which the economic analysis offers interpretative guidance. With this variability and diversity in form, participation, and scope of commitments under EU law, the focus of this chapter lies on the design of EU law. Having established the core rationales for why EU members seek to commit to binding law, we may further explore how cooperation under EU law may unfold. We may draw from the arsenal of economic methods introduced in the Introduction and from the basic rationales for entering into EU cooperation, as discussed in Chapter 1, to explore different design features of EU law. For example, federalism theory inspires the legal principle of conferral, with citizens’ preferences and economies of scale as analytical benchmarks; public good theory underpins the subsidiarity principle; contract theory identifies where EU law provides for exception and escape clauses; transaction costs analysis informs phenomena of hard law and soft law under EU law; game theory makes predictions about compliance with EU law and informs the enforcement institutions with the roles of the EU Commission and the ECJ. Likewise, we will benefit from the analysis of Chapter 1 on the rationales of cooperation, in order to understand why EU members choose a specific Treaty design the way they do in light of their individual government objectives and payoff function, the spillovers of policy actions. Given the focus of this Part on the general arrangements tailored in EU law, that is, the basic principles stipulating the legal framework of EU Treaties, we should recall that the EU is a selfconscious product of its Member States— illustrated by the fact that EU members are the ‘Masters of the Treaties’, even if the production of EU secondary law is a product of multiple actors combining EU institutions and 78 armin steinBach Fon have shown that when a Treaty is left open (ie when it is predicated on future accessions of new members), the initial treaty content is set optimally on the basis of the incumbent states’ expectation of Treaty enlargement.7 Hence, drafting semiopen Treaty sections would have to anticipate a growing (and potentially more heterogenous) membership. The rational Treaty designer anticipates growing heterogeneity of preferences between participating states and thus designs rules in a way that minimizes negative repercussions. Looking at the euro membership as an example, incumbent euro members would be well advised to specify upfront in the initial Treaty the precise conditions under which accession to the euro for new members is possible. It is straightforward that euro members have a strong economic interest in predetermining these conditions, as the eurozone has weakestlink public good character (see above Chapter 6) and thus bears the risk of freeriding and negative spillovers. Consequently and unsurprisingly, the path to euro adoption is welldefined, with Article 140 TFEU enumerating the ‘high degree of sustainable convergence by reference to the fulfilment of criteria’, which are further specified in Protocol 13 of the TFEU, with even permissible inflation and budget deficit threshold numerically defined.8 The high degree of upfront and Treatybased specification of entry requirements is in line with the above mentioned logic— in a semiopen Treaty, the participating countries should anticipate future accessions by choosing a Treaty design that ensures the alignment of future Treaty participants with the principles and goals of the Treaty founders. By contrast, the closed sections of EU Treaties do not make such high specification necessary, as the veto principle allows incumbent members to decide adhoc whether they consider accession beneficial. Accordingly, accession requirements to the EU are only vaguely referred to under Article 49 TEU, by mentioning the necessary respect of the values referred to in Article 2 TEU.9 (EU) Treaty formation and amendment further depend on the payoff structure for incumbent and acceding members. In general, from the perspective of the incumbent EU members, accession of new members causes a different kind of potential costs which incumbents seek to minimize. Granting conditional access to the EU is in line with rational choice suggesting that restrictive membership is necessary due to uncertainty about the acceding member’s preferences.10 The negotiation process of accession with new members can be understood as a procedure for revealing private information about the applicant’s preferences. The incumbent EU members do not want new members who freeride on public goods provided 7 Parisi and Fon (n 3). 8 The fiscal surveillance procedure as well as economic policy coordination governance, which are explicitly mentioned in Article 139(2)(a) and (b) TFEU will kick in once a new member enters the common currency area. 9 The admission criteria were also absent in earlier Treaty versions and were then laid down at the June 1993 European Council in Copenhagen. 10 Barbara Koremenos, The Continent of International Law: Explaining Agreement Design (Cambridge University Press 2016). memBershiP of eu treaties 79 across the Union (eg legal stability, solidarity, security), but instead they seek to take on members who contribute to the public goods and comply with the acquis communautaire. Adoption of the acquis communautaire is a costly signal ensuring an equilibrium where only those who share certain characteristics are included.11 By forcing acceding EU members to undertake concessions and bear significant adjustment costs, they increase the likelihood that the applicant will be a compliant and cooperative EU member. While this serves to reduce one source of costs for incumbent states associated with new accessions, there are significant institutional costs due to the dilution of voting power and influence on policy outcomes. Each round of accession requires negotiations about the redesigning of the institutional representation in order to balance the influence of current and new EU members. Hence, potential benefits of increased EU Treaty participation and deeper Treaty content must be balanced, from the perspective of the EU incumbent, against the costs of diluted power. With each accession, much depends on the marginal benefits and costs— whether gains from increased participation outweigh dilution of influence. Connected to dilution of voting power, future adoption of secondary law through EU institutions may come at a higher price with extended membership— the risk of being overruled in the Council increases sovereignty costs, and heterogenous preferences raise negotiation costs, all of which must be factored in by a rational incumbent before giving approval to new members.12 Importantly, opportunity costs of countries vary. Variable opportunity costs signify that countries may be able to pursue gains from cooperation outside the EU to different degrees. Variable opportunity costs invite the insight of distributive rationalist approaches, which argue that what matters is that some states have better outside options than others.13 According to this logic, states with a greater ability to achieve their preferred outcomes without cooperation can leverage the threat to act unilaterally to shape multilateral institutional rules; on the international level, the United States has used both its informal and formal power to design and control international (economic) institutions. This would suggest that on the EU level, powerful states skew voting rules in their favour. However, the Lisbon Treaty rules modifying qualified majority voting led to a decrease in the total power of larger countries, while the total power of smaller states increased, which supports the existing phenomenon of smaller countries being overrepresented in EU institutions.14 At least regarding the voting structure of the EU, larger incumbent 11 Koremenos, Lipson, and Snidal (n 2) 784. 12 Mareike Kleine, Informal Governance in the European Union: How Governments Make International Organizations Work (Cornell University Press 2013) 54. 13 Lloyd Gruber, Ruling the World: Power Politics and the Rise of Supranational Institutions (Princeton University Press 2000); Phillip Y Lipscy, ‘Explaining Institutional Change: Policy Areas, Outside Options, and the Bretton Woods Institutions’ (2015) 59 American Journal of Political Science 341. 14 Alessandro Scopelliti, ‘The Political DecisionMaking Process in the Council of the European Union Under the New Definition of a Qualified Majority’ (2008) 73 Il Politico 180. 80 armin steinBach countries have not translated their comparative low opportunity costs into relative advantages over smaller acceding states. EU enlargement happens when benefits for incumbent and acceding members are significant, where the general rationales for cooperation apply (Part II). Enlargement may be particularly plausible where policy areas are characterized by coordination games and network effects, that is, in situations where benefits increase for all members as the number of participants grows (eg joint product standards on the EU internal market).15 Also, admitting a new EU member may create additional valueenhancing opportunities for all states, for example through stimulating economic growth or stabilizing external relationships.16 In turn, with weakestlink public goods the situation is more ambivalent. In principle, cooperation in weakestlink public goods declines with the number of states, as the cost of monitoring increases and the probability of a weakest link rises. If new EU members undertake insufficient efforts to combat illicit trade or to secure border protection, this lowers the payoff for all other members. In these cases, either cooperation fails, or alternatively, there is plausibility for centralization of the weakestlink good in order to ensure minimum enforcement, provided that the EU is sufficiently prepared to ensure a consistent level of protection for these goods. One example is the EU Schengen Agreement: as border protection is a weakestlink public good, several Eastern European countries have not been able to join this element of the EU Treaties. Incumbent countries request a high level of border protection to ensure that the benefits of removing internal border controls are not undermined by some members. Euro membership is another illustrative example: new members must subject themselves to convergence criteria and EU fiscal rules requiring solidity of national budgets. A common currency has features of a weakestlink public good, as a critically unstable country may cause repercussions for all other members. 15 Sébastien Dupuch, Hugues Jennequin, and El Mouhoud Mouhoub, ‘EU Enlargement: What Does it Change for the European Economic Geography?’ (2004) 91 Revue de l’OFCE, Presses de SciencesPo 241; Paolo Cecchini, The European Challenge 1992: The Benefits of a Single Market: European Challenge— Benefits of a Single Community (Gower Publishing Ltd 1998). 16 Arjan M Lejour, Vladimir Solanic, and Paul JG Tang, ‘EU Accession and Income Growth: An Empirical Approach’ (2009) 16 Transition Studies Review 127. EU Law and Economics. Armin Steinbach, Oxford University Press. © Oxford University Press 2025. DOI: 10.1093/ 9780198920915.003.0010 10 Centralization The objectives for pursuing cooperation within the EU, laid out in Part I, refer to the motivation of interjurisdictional efficiency.1 Interjurisdictional efficiency seeks to determine the best allocation of resources among different jurisdictions. Interjurisdictional efficiency is achieved when the public policies offered by the government satisfy the collective demands of citizens at the lowest cost.2 Interjurisdictional efficiency addresses the foundational problem and rationale for the EU— it acknowledges the legal Westphalian system of relationships between European sovereign states. There are multiple motives for cooperation as a result of interjurisdictional (in)efficiencies, such as externalities either between individuals in one state and individuals in another or between public policies. These externalities can either be channelled through prices (and amount to market failure) or, if reflected in prices, they can be felt as spillovers in neighbouring states (and amount to ‘beggarthyneighbour’ policies).3 Hence, interjurisdictional inefficiencies arise if Member States produce decentralized spillovers which are likely to ignore the overall economic costs or benefits which will lead them to overor underprovide the activity. In this chapter, we are concerned with intrajurisdictional efficiency. In theory, it concerns the choice of public policy satisfying the collective demand within a given jurisdiction in a way that ensures citizens’ willingness to pay for those activities. Intrajurisdictional efficiency addresses the question whether public activities are pursued either at the EU level or at Member State level. In general terms, the EU should set the public policy if the benefitstocost ratio is more favourable compared to when Member States pursue the policy individually. Intrajurisdictional efficiency strategies may aim at reducing bargaining and transaction costs or establish institutions to reveal and disseminate information that would not be available in a setting of bilateral interaction between EU Member States, such as the EU Commission and the ECJ. Likewise, centralization can facilitate the enforcement 1 Frank H Easterbrook, ‘Federalism and European Business Law’ (1994) 14 International Review of Law and Economics 125, 129; Alan O Sykes, ‘Externalities in Open Economy Antitrust and Their Implications for International Competition Policy Competition, Free Markets, and the LawSymposium on Law and Public Policy1999’ (1999) 23 Harvard Journal of Law and Public Policy 89. 2 Robert P Inman and Daniel L Rubinfeld, ‘Rethinking Federalism’ (1997) 11 Journal of Economic Perspectives 43; Charles M Tiebout, ‘A Pure Theory of Local Expenditures’ (1956) 64 Journal of Political Economy 416. 3 Hendrik Spruyt, ‘Institutional Selection in International Relations: State Anarchy as Order’ (1994) 48 International Organization 527. 82 armin steinBach of agreements: when bilateral or regional cooperation relies on decentralized enforcement, there is a perennial risk of undercompliance because the costs of enforcement deter the enforcing party, unlike with a neutral enforcer (such as the EU Commission and the ECJ) who mutualizes the costs of enforcement, giving rise to a secondorder collective action problem (see below Chapter 14). The intrajurisdictional question of whether and how to centralize within the EU revolves around two core legal provisions in the EU Treaty— the principle of conferral and the principle of subsidiarity. While the principle of conferral governs the limits of EU competences, the use of those competences is governed by the principle of subsidiarity. The principle of conferral embodies the deep Westphalian international law idea that the EU’s competences are voluntarily conferred on it by its Member States and that the EU must remain within these borders— the EU is treated as a ‘dependent variable’, crafted in line with the preferences of Member States as ‘masters of the Treaties’ or as ‘independent variables’. The legal EU jargon of ‘competences’ is a key term which encapsulates the degree to which Union institutions or Member States can exercise authority in a given policy area. With competences being nothing more than authority, we may leverage Trachtman’s property right analogy applied to international law. In this vein, European law may be understood in terms of reciprocal restraints on state autonomy, and therefore as transactions regarding the exercise of jurisdiction. This lends itself to the argument that jurisdiction is analogous to property in a private context. Jurisdiction gives states the power to rule on activities and assets, while property gives individuals the right to control certain assets.4 Both jurisdiction and property rights are constructed by law to encompass certain rights, and both jurisdiction and property rights are transferable. The property rights theory offers insights into the EU’s nature as an international organization to which members have surrendered certain sovereign functions (ie their power to exercise jurisdiction). One may inquire whether and to what extent centralization— a transfer of jurisdiction— should occur. The theory of property rights suggests an initial vesting of rights in favour of those who are likely to value them most, in order to minimize transaction costs.5 Let’s assume two states, one which pursues domestic regulation for domestic policy reasons, but which creates a negative effect on another country. The property rights strategy would ask from a normative perspective which country would value the right to regulate most, and to this country sole authority should be given. Speaking of a jurisdiction having a value is alien to legal jargon, but it connects to an economic effect orientation. Societies positively value the pursuit of public values, but they negatively value being adversely affected (eg through negative policy spillovers.) 4 Joel P Trachtman, The Economic Structure of International Law (Harvard University Press 2008) 72, 31. 5 Richard A Posner, The Economics of Justice (Harvard University Press 1983). centralization 83 Hence, the value a government attaches to jurisdiction is proportionate to the benefits it can obtain through regulation, or through avoidance of the application of another state’s regulation.6 In such situations, the property rights theory would suggest that jurisdiction should be allocated to the country that would be chosen in negotiation between countries in the absence of any transaction costs. An inefficient allocation would be one where all negative or positive effects of a country’s regulation would materialize in another country. In turn, if mixed effects occur (or if the regulating country values the domestic policy positively while the affected country negatively), the dispersion of jurisdiction among all affected jurisdictions would be the optimal solution.7 In this case, the appropriate entity for exercising prescriptive jurisdiction may not be a national government but an international organization— such as the EU. In practice, it is clearly difficult to determine how much a country values the right to regulate or how it perceives foreign policies as costs. This suggests that the authority to allocate property rights should be given to the entity who is able to analyse costs and benefits, and to initiate appropriate reallocative transactions.8 From this perspective, in the context of multiple countries such as in the EU, a neutral entity is positioned to act as broker and agent to collect information and to assess crossborder comparisons of costs and benefits. This logic is indeed enshrined in the EU’s shared competence principle (Article 5(3) TEU) — it paves the way for two alternative allocations of jurisdiction. Either Member States remain competent in cases where the acting state pursues a policy (one to which it attaches positive value). As long as the state manages to achieve the policy objective by acting alone, the jurisdiction remains with the Member State. Or, the single state does not achieve the policy objective effectively— either because it causes harm to another state (negative externality), or because it remains insufficiently implemented. In these cases, the EU monitors the overall effects within the EU and assumes competence to act. a) Principal of conferral: which role for the EU and Member States? Inquiring into the principle of conferral laid down in Article 5(1) TEU from an economic perspective can unfold in two dimensions. First, the principle of conferral can be understood as an allocation device in the sense that it stabilizes a certain competence distribution between the EU and Member States— this distribution 6 Trachtman (n 4) 41. 7 ibid 42. 8 Guido Calabresi, The Costs of Accidents: Legal and Economic Analysis (Yale University Press 1970). 84 armin steinBach can be efficient or inefficient from an intrajurisdictional perspective. The principle of conferral is thus inextricably linked to the question of whether a competence should be dealt with at Member State or EU level. The principle of conferral enshrines the notion of KompetenzKompetenz, positing that it is Member States who decide on the scope of surrendering competences to the EU.9 This positivistlegal approach can be contrasted with an economic view as to how the competence allocation between the centre and the decentralized entities, that is, between the EU and Member States, should be designed (irrespective of its positivist allocation). This question leads us to federalism theory, which offers insight into the relationship between the EU and its Member States.10 In this context, ‘homogeneity’ has been a term widely used by lawyers, political scientists, and economists in order to make inferences on the optimal distribution of competences between the centre and its members.11 Cultural homogeneity has been discussed as a prerequisite for statehood and was seen as a barrier to more EU centralization,12 while homogeneity of preferences captures the economic yardstick to determine the optimal level of regulation.13 Moreover, lawyers take recourse to ‘legitimacy’, and political scientists have distinguished between input and output legitimacy to justify public authority, a distinction that one can capture with the outcome focus of welfare economics versus the process orientation of constitutional economics.14 While the first question concerns the role of the principle of conferral in making the optimal distribution of powers, we can also ask a second question— whether the principle of conferral is effectively applied and respected within the existing allocation of competences in the EU. This second perspective accounts for the judicial innovations crafted by the ECJ, propelling a dynamic ushering of EU law into doctrinal inventions such as ‘implied powers’ or effet utile which can be understood as legal techniques undermining the principle of conferral.15 The first question 9 Erin Delaney, ‘Managing in a Federal System without an “Ultimate Arbiter”: Kompetenz in the EU and the AnteBellum United States’ (2005) 15 Regional & Federal Studies 225. 10 Wallace E Oates, ‘On the Evolution of Fiscal Federalism: Theory and Institutions’ (2008) 61 National Tax Journal 313. 11 Matthias Mahlmann, ‘Constitutional Identity and the Politics of Homogeneity’ (2005) 6 German Law Journal 307; Oliver Mader, ‘Enforcement of EU Values as a Political Endeavour: Constitutional Pluralism and Value Homogeneity in Times of Persistent Challenges to the Rule of Law’ (2019) 11 Hague Journal on the Rule of Law 133. 12 Stephen Weatherill, ‘Why Harmonise?’ in Takis Tridimas and Paolo Nebbia (eds), European Union Law for the TwentyFirst Century: Rethinking the New Legal Order (Hart Publishing 2003) 11, 13; Stelio Mangiameli, ‘The European Union and the Identity of Member States’ (2013) No 369 L’Europe en Formation 151. 13 Wallace E Oates, Fiscal Federalism (Harcourt Brace Jovanovich 1972). 14 Viktor J Vanberg, ‘Market and State: The Perspective of Constitutional Political Economy’ (2005) 1 Journal of Institutional Economics 23, 28. 15 Carl Lebeck, ‘Implied Powers beyond Functional Integration— The Flexibility Clause in the Revised EU Treaties’ (2007) 17 Journal of Transnational Law & Policy 303; Elio Maciariello, ‘EU Agencies and the Issue of Delegation: Conferral, Implied Powers and the State of Exception’ (2019) 4 European Papers 723; Urška Šadl, ‘The Role of Effet Utile in Preserving the Continuity and Authority of European Union Law : Evidence from the Citation Web of the PreAccession Case Law of the Court of Justice of the EU’ (2015) 8 European Journal of Legal Studies 18. centralization 85 can be boiled down to asking whether public policy action is better placed at the EU level or at the decentralized Member States level. The second question asks whether and how the principle of conferral should be adhered to— through a strict or flexible interpretation. Federalism theory: decentralization as the optimal solution? The principle of conferral can be explored by asking whether there should be many individual EU members acting separately instead of one single EU entity. The theory of system competition anchored in institutional economics offers an answer. Charles Tiebout offered a model describing fiscal competition between many independent local governments.16 The Tiebout model is an extension of the purely competitive market model with complete information. Governmental or interjurisdictional competition follows the logic of efficient private market competition and builds on similarly strong assumptions, but the basic insight from theories of federalism is that jurisdiction over a problem should be allocated to the lowest level of government capable of internalizing the relevant externalities.17 At the centre of this is the idea that states compete for citizens as mobile resources. The rationale for decentralized individual or member states is that they can better accommodate the specific preferences of their citizens and thus provide better public service offers (‘race to the best fit’).18 It reasons that the state supplies public goods at specific tax prices and uses this ‘product’ to compete with other states for citizens. With the focus on individuals’ preferences, from the Tiebout perspective, the world of competitive governments will allow for the analogy of a marketplace in which individuals move among local jurisdictions to select the public goods and public services combination that they most desire. By definition, citizens can choose their preferred legal regime in the same way consumers choose among competing products. Under certain assumptions— such as frictionless legal competition, institutional rules such as mutual recognition— there will be a competitive legal equilibrium maximizing aggregate welfare. From a political economic perspective, competition is intended to foster legal diversity in order to be an effective constraint on public policymakers who are not inclined to act in benevolent ways. Just as competition on the supply side reduces the margin and promotes quality, the competitive environment in jurisdictional markets disciplines the lawmaker’s discretion. In order to maximize 16 Tiebout (n 2). 17 Inman and Rubinfeld (n 2). 18 David Christoph Ehmke, Institutional Congruence the Riddle of Leviathan and Hydra, vol 16 (Nomos Verlagsgesellschaft 2018) 38. 86 armin steinBach procompetitive pressure, transaction costs should be minimized in order to facilitate citizens’ choices. The resulting hypothesis is that law gives states an incentive to compete by providing efficient legal rules.19 It follows from this political economic logic that decentralized solutions that leave competences at Member State level, are all the more appropriate the more heterogeneous regional preferences are.20 Since individual needs are different due to culture, ethnicity, history, and the state of the economy, decentralized solutions can better design suitable public goods, that is, tailoring their supply to individual preferences.21 The more diverse the preferences between regions and the more homogeneously they are distributed within regions, the more advantageous the competition between EU members is, and the more disadvantageous Europeanization is.22 However, this theory does not fare well in the practical realities of the EU. The Tiebout framework posits a number of strong assumptions, one of which is mobility of households among jurisdictions at no cost, or that businesses can easily choose the jurisdiction that offers the taxexpenditureregulation package that it most prefers.23 Yet mobility costs remain high in the EU, despite the free movement of persons and employees stimulating mobility.24 The euro crisis triggered some, though limited, south– north migration, yet considerable barriers to the ‘exit’ remain (eg language, recognition of professional qualification). Also, a (too) strong assumption of the Tiebout model is the perfectly elastic supply of public goods in EU Member States, one that assumes each state to be capable of replicating all of the attractive economic features of its competitors. The idea of ‘inefficient politicians’ being driven out of the market through election ignores the realities of political economics. In addition, information presuming households, citizens, and businesses to be fully informed about the fiscal and regulatory policies of each jurisdiction, appears an overly ambitious theorical assumption. Households and businesses are often illinformed— it is difficult for them to ascertain the implications of regulations, tax rules, and public activities, let alone to make a comparison with other EU Member States given regulatory idiosyncrasies in EU Member States. 19 Francesco Parisi and Vincy Fon, The Economics of Lawmaking (Oxford University Press 2008) 52, 7. 20 Stephen Weatherill, ‘Protecting the Internal Market from the Charter’ in Sybe de Vries, Ulf Bernitz, and Stephen Weatherill (eds), The EU Charter of Fundamental Rights as a Binding Instrument: Five Years Old and Growing (Hart Publishing 2016) 151. 21 Robin Boadway and Anwar Shah, Fiscal Federalism: Principles and Practice of Multiorder Governance (Cambridge University Press 2009) 35. 22 Charles B Blankart, Föderalismus in Deutschland und in Europa (Nomos Verlagsgesellschaft 2007) 59. 23 Tiebout (n 2); William W Bratton and Joseph A McCahery, ‘The New Economics of Jurisdictional Competition: Devolutionary the New Economics of Jurisdictional Competition: Devolutionary Federalism in a SecondBest World Federalism in a SecondBest World’ (1997) 86 Georgetown Law Journal 201. 24 In the US, moving between states is mostly job motivated and not due to a more attractive public service environment, Frank H Easterbrook, ‘Antitrust and the Economics of Federalism’ (1983) 26 Journal of Law and Economics 23. centralization 87 Individuals and businesses face high costs in resolving this lack of comparability and undertaking a costbenefit calculus. Adding to this, while system competition may be efficient from an allocation perspective, it has crossregional distributional effects through ‘beggarthyneighbor’ policies. EU countries can try to entice international market shares away from other EU members for their industries by reducing wages and thus increasing their current account surplus, which impairs the relative competitiveness of other countries.25 In the worst case scenario, this may cause a harmful ‘race to the bottom’, when social and environmental standards are dismantled.26 Overall, this can lead to suboptimal competition for society as a whole.27 ‘Homogeneity’ of what: preferences or public space? But even if one acknowledges federalism theory’s plea for decentralized Member State competences, the theory builds on a notion of preferences orientation that may be contested from legal and political science perspectives. Economics, law, and political science converge at a pivotal juncture of EU governance research, that is, determining what jurisdictional level— local, regional, national, or EUwide— public authority should be exercised. Homogeneity and heterogeneity are both terms that have inspired economic, legal, and political scholarship and they may entail commonalities, but also divergences. Take the economic relevance of heterogeneity of preferences for the provision of public goods. Unlike private goods, public goods are hard to adjust to diverse preferences. Since all public goods are nonrivalrous, citizens must accept the same set of public good characteristics, whether they align with their preferences or not. Naturally, when the EU regulates a heterogeneous group of citizens with different norms, values, and habits, we are more likely to observe the emergence of disagreements over public goods and policies, such as how security should be supplied or what the preferences for environmental protection concern.28 The EU has not strictly followed this rationale of aligning homogeneity (heterogeneity) of preferences with harmonization (decentralization) of laws: harmonization 25 Palma Polyak, ‘The Silent Losers of Germany’s Export Surpluses. How Current Account Imbalances Are Exacerbated by the Misrepresentation of Their Domestic Costs’ (2024) 22 Comparative European Politics 31. 26 Catherine Barnard, ‘Social Dumping Revisited: Lessons from Delaware’ (2000) 25 European Law Review 57 (arguing against the ‘race to the bottom’ as a European problem at least in the area of social policies). 27 Albert Breton, ‘The Existence and Stability of Interjurisdictional Competition’ in Daphne A Kenyon and John Kincaid (eds), Competition among States and Local Governments (The Urban Institute Press 1991); Robert P Inman and Daniel L Rubinfeld, ‘Economics of Federalism’ in Francesco Parisi (ed), Oxford Handbook of Law and Economics (Oxford University Press 2017) 90. 28 Enrico Spolaore, ‘The Economic Approach to Political Borders’ (2022) CESifo Working Paper No 10165. 94 armin steinBach directly feed into the composition of governments and state leadership. The lack of input legitimacy has widely been criticized by lawyers and has led them to request that the EU should not further expand its competences.57 Others have emphasized the poor performance in terms of output legitimacy.58 The process of supranationalization led to a shift of governmental power from the politized national level, which was deeply rooted and connected to national citizens, to the technocratic EU level that defines itself through technocratic resolution of its competences.59 The merit of depolitization was seen in its ability to ensure better governance for the people without significant effects on governance by and of the people.60 Input legitimacy may find its economic corollary in what constitutional economics has described as citizens’ sovereignty. Constitutional economics suggests a normative design of these constitutions, as initially coined by William H. Hutt61 and elaborated by Viktor Vanberg62: an economic constitution that enhances consumer sovereignty, and a political constitution that enhances citizen sovereignty. While consumer sovereignty builds on the proposition that nondiscriminatory market competition and internalized externalities generate welfare maximizing outcomes, citizen sovereignty looks at the collective arrangements in the political arenas, allowing formation of political will through ‘cooperative ventures for mutual advantage’.63 In a manner similar to consumer sovereignty, citizen sovereignty places the individual at the heart of a democratic polity, in whose common interests the polity should be operated. Accordingly, the political process should be institutionally framed in a manner that makes citizens’ common interests its principal controlling force.64 In other words, citizen sovereignty requires that political institutions, domestic politicians, and bureaucracies as well as international organizations are made most responsive to citizens’ common interests— a requirement echoed by input legitimacy. Institutions, decisionmaking processes, fundamental rights protections, and adjudication must be implemented and respected in ways that maximize the prospect of the political process working to the mutual advantage of all citizens.65 57 Simon Hix and Andreas Follesdal, ‘Why is There a Democratic Deficit in the EU? A Response to Majone and Moravcsik’ (2006) 44 Journal of Common Market Studies 533; Dieter Grimm, Europa ja - aber welches? Zur Verfassung der europäischen Demokratie. (CH Beck 2016) 29. 58 Giandomenico Majone, Europe as the WouldBe World Power: The EU at Fifty (Cambridge University Press 2009) 10. 59 James Ferguson, The AntiPolitics Machine: Development, Depoliticization, and Bureaucratic Power in Lesotho (University of Minnesota Press 1994). 60 Schmidt (n 55) 69. 61 William H Hutt, Plan for Reconstruction: A Project for Victory in War and Peace (Kegan Paul 1943) 215. 62 Vanberg (n 14) 37. 63 John Rawls, A Theory of Justice (Harvard University Press 1971) 84. 64 Vanberg (n 14) 42. 65 ibid. centralization 95 In turn, welfare economics’ focus on outcomes can be likened with output legitimacy. Assessing whether policy decisions are adopted in line with citizens preferences is predicated on whether the policy outcome matches what citizens want. Output legitimacy focuses on whether public policies actually serve the common good in line with the community norms. Output legitimacy is empirically assessable: the extent to which the outcomes of a governing authority’s policies are effective in solving the polity’s problems and objectives.66 Some scholars have argued that the EU’s lack of input legitimacy could be compensated by its performance in output legitimacy. While the EU’s policymaking through nonmajoritarian institutions was criticized from an input legitimacy perspective, their comparative positive record in producing effective policies has been appreciated.67 Others have emphasized the benefits of the checks and balances in the EU’s multiple veto system to ensure appropriate outcomes.68 There is no empirical proof that EU decisions actually satisfy citizens’ preferences. One could draw on surveys highlighting the level of satisfaction with EU policies or participation in elections for the EU parliament, or refer to the perception of individual events taking place at the EU level— such as protests in France and elsewhere against the Commission’s initial services directive or the protests of farmers on the streets of Brussels.69 In conclusion, the legal and political science concepts revolving around homogeneity and legitimacy can be likened with economic concepts in different ways. Some lawyers refer to homogeneity of citizens as reference points for which and how many competences can be shifted to the EU level— according to some, only if there is a sufficient European identity, a sense of ‘peopleshood’, an identity predicating on people’s shared sense of belonging to a political community and acceptance of European practices. Economists in turn have used homogeneity of preferences as a concept for determining the scope of centralization versus decentralization. Economics also connects to the longstanding academic discourse on capturing the EU’s identity by reference to input and output legitimacy. The economic perspective may vary in this respect. On one hand, an institutional economic perspective of processorientation can be likened with input legitimacy by focusing on citizens being able to articulate their preferences to public policymakers. On the other hand, welfare economics, unlike constitutional economics, is concerned with specific outcomes.70 To the extent that constitutional economics emphasizes the acceptability of the process that leads to a decision rather than the 66 Schmidt (n 55) 32. 67 Giandomenico Majone, ‘Nonmajoritarian Institutions and the Limits of Democratic Governance: A Political TransactionCost Approach’ (2001) 157 Journal of Institutional and Theoretical Economics 57. 68 Andrew Moravcsik, ‘In Defence of the Democratic Deficit: Reassessing Legitimacy in the EU’ (2002) 40 Journal of Common Market Studies 603. 69 Schmidt (n 55) 61. 70 Vanberg (n 14) 27. 96 armin steinBach outcome, it connects to input legitimacy, while standard welfare economics with its outcome orientation can be likened with output legitimacy. effet utile and ‘implied powers’ We can further ask whether the principle of conferral is effectively applied and respected within the existing allocation of competences under EU Treaties. This perspective invites an economic eye to the dynamic enshrined in jurisdictional innovations developed by the ECJ such as ‘implied powers’ or effet utile, both of which can be understood as legal techniques innovated by the Court without approval from the ‘master of the Treaties’. More generally, a competence shift in the EU can be the result of a creeping erosion of competences.71 It is then not the deliberate rational choice of EU members to transfer a set of public policies to the EU, but it is Union institutions that use the Treatybased transferred competences to expand them informally through Treaty interpretation, which practically amounts to creeping Treaty amendments (though without complying with the substantive requirements for amendments). In theory, such an interpretation would amount to a breach of the Treaty duty and violate the principle of conferral, but if that breach is condoned or even committed by the ‘custodians of the treaties’ (the Commission and the ECJ), the Member States have limited recourse, particularly if both the executive and the judiciary of the EU agree on an extensive interpretation of EU law. When it is easy to circumvent the Treatybased distribution of competences by way of interpretation and Treaty application, the KompetenzKompetenz as a stronghold of national sovereignty loses its function of protecting the sovereignty of the Member States.72 The implied powers doctrine states that a principal competence expressly conferred on the EU must include those powers without which the principal competency cannot reasonably and expediently be exercised.73 This doctrine can operate in a narrow or broad manner, depending on whether an authority is deemed ‘necessary’ or only ‘reasonably necessary’ for the exercise of the expressly conferred powers.74 Closely conceptually related to this is the effet utile interpretation of the ECJ, according to which the Treatybased competence should be interpreted teleologically, with the aim of fully developing the meaning of the provision or of achieving its ‘full effectiveness’. The ECJ thus interprets the limited competence 71 Stephen Weatherill, ‘Protecting the Internal Market from the Charter’ in Sybe de Vries, Ulf Bernitz and S Weatherill (eds.), The EU Charter of Fundamental Rights as a Binding Instrument: Five Years Old and Growing (Hart Publishing 2016), 133. 72 Grimm (n 57) 112. 73 Piet Eeckhout, ‘The Doctrine of Implied Powers’ in EU External Relations Law (Oxford University Press 2011). 74 Trevor C Hartley, The Foundations of European Community Law: An Introduction to the Constitutional and Administrative Law of the European Community (Clarendon Press 1994) 110. centralization 97 according to the principle of effet utile in such a way that all conceivable Union powers can be fully exploited.75 Typically, under international law, courts that are established by Member States interpret the rules that were designed by Member States according to the will of Member States. This corresponds with a rational choice perspective— Members have no interest in an international organization turning against the will of the founders. Interpreting an agreement between states with reference to the historical and actual will of Member States ensures that international law is applied in line with the preferences of those who crafted it. Some even argue that only the original agreement to the constitution matters.76 Rational choice suggests that states creating an international agreement do so on the premise of retaining their full national sovereignty, even if they partially bind themselves to decisions of the international organizations they create (eg accepting compulsory WTO dispute settlement rulings). Likewise, and in line with the Westphalian logic of international law, conventional legal interpretation focuses on the will of the state, its agents, interpreted by reference to, for example documents drafted by government representatives.77 If the adjudicative bodies deviate from the founding states’ intention, we speak of a principal agent issue, with bargaining costs (predefining rules to limit deviation by adjudicative bodies) to be balanced with agency costs (caused by deviation of the courts as agent). From the early days of the European integration process, the ECJ established that EU law does not form part of regular international law, nor is EU law dependent on national acts. Rather, the EU evolved independently from its national origins and an autonomous legal order emerged, a supranational construct that stands on its own feet, an institutional organization sui generis.78 Consequently, the ECJ considered itself as not bound by conventional ways of interpreting EU law that would have subjected the interpretation of EU law to the will of the Member States and, in particular, respecting Member States’ national sovereignty. While a legal originalist interpretation of law partners with a rational choice perspective, the narrow focus on state perspective contrasts with a constitutional economic perspective, which is more concerned with the importance of ongoing agreement of citizens with the legal arrangement rather than the issue of original agreement. We recall that constitutional economics posits ‘citizen sovereignty’ and a procedural responsiveness of public policy choices to current members of society.79 The constitutional economic tenets state that it is the voluntary agreement 75 Michael Potacs, ‘Effet utile als Auslegungsgrundsatz’ (2009) 44 Europarecht 465. 76 Adrian Vermeule, Common Good Constitutionalism: Recovering the Classical Legal Tradition (Polity 2022) 91. 77 Grimm (n 57) 46. 78 William Phelan, ‘What is Sui Generis about the European Union? Costly International Cooperation in a SelfContained Regime’ (2012) 14 International Studies Review 367. 79 Vanberg (n 14). 98 armin steinBach of all members to the constitution that provides legitimacy. It is concerned with the ongoing voluntary acceptance of members’ citizens by its current members. Legitimacy to the constitution is drawn from the consent of the current democratic polity, not from some original agreement that may or may not have existed at the founding of the polity. Constitutional economics hence turns the ‘original intent’ legal interpretation upside down. Consequently, as Vivien Schmidt puts it, ‘if a constitution is no longer generally accepted by its current members, it can surely not be considered more legitimate than a constitutional arrangement that may have been imposed originally by outside force or by decree, but that in its current operation is met by general approval within the respective constituency’.80 With constitutional economics militating against legal interpretations attaching value to originality in interpretation and rigidity of Treaty application, it may even raise doubts as to whether the ‘principle of conferral’ effectively preserves the original allocation of competences. Constitutional economics looks favourably at an ‘implied power’ legal interpretation, one that dynamically adapts to evolving circumstances. To the extent that EU members consider a competence to be enshrined in a competence of the EU Treaty, even if not explicitly mentioned, they express their consent by practice— a sufficient indication for constitutional economists to assume this constitutes an arrangement of ‘mutual gains’. However, in practice it has been Union institutions— the Commission and the ECJ— who have been invoking extensive interpretations of EU competences— not Member States.81 One should accept mutual gains and consent with extensive Treaty interpretations only if they have been approved (unanimously) in the Council. Clearly, the supporters of literal interpretation and ‘original intent’ have a point when arguing that the initial Treaty was subject to (parliamentarian) approval, unlike the ‘implied power’ interpretation. The claim that current members of society are more relevant than the will of original members lacks empirical support because there is no active choice of current members, as no parliamentary approval nor referendum vote is exercised in relation to ‘implied power’ interpretation, while tacit acknowledgment of ECJ jurisprudence can hardly be taken as deliberate approval. Policy actions of specific EU policy decisions have never been brought to the citizen’s choice. Citizens’ choices form the basic idea of ‘competitive federalism’ (see above Chapter 4), yet Treaty amendments or implied powers are limited to Member States’ consent articulated through the Council rather than direct approval by citizens through elections or referenda. 80 Schmidt (n 55). 81 Aurelien Portuese, ‘Principle of Proportionality as Principle of Economic Efficiency’ (2013) 19 European Law Journal 612, 631; Gareth Davies, ‘Subsidiarity: The Wrong Idea, in the Wrong Place, at the Wrong Time’ (2006) 43 Common Market Law Review 63, 64; Martin Nettesheim, ‘Grundrechtliche Prüfungsdichte durch den EuGH’ (1995) Europäische Zeitschrift für Wirtschaftsrecht 106. centralization 99 b) Subsidiarity principle Article 5(1), sentence 2 and (3) TEU enshrines the principle of subsidiarity. From an economic perspective the principle of subsidiarity is a core element in a theory of system competition or fiscal federalism. Taking fiscal federalism as a benchmark, the subsidiarity principle can be economically operationalized with reference to externalities, economies of scale, and heterogeneity of preferences. Legally, it is a principle governing the exercise of competences and as such protects Member States from intrusive exercise of Union power. It precludes Union action when a matter can be effectively dealt with by the Member States at national, regional, or local level, and empowers the Union to exercise its powers when the objectives of a proposed measure are not sufficiently satisfied by the Member States.82 The subsidiarity test under EU law evolves in two steps. First, the policy action concerned must be one of shared competence between the Union and Member States. In the second step, the effectiveness test applies inquiring whether the policy action ‘cannot be sufficiently achieved by the Member States’. Concerning the nature of the competence, when the EU holds exclusive competences, or if the competence, in line with the principle of conferral, remains with the Member States, the subsidiarity test does not apply. The exclusive responsibilities of the EU (Article 3 TFEU) include customs policy, monetary policy, competition rules, and the common trade policy. Shared competence (Article 2(2) TFEU) includes, for example, internal market rules, agriculture, and environmental policies. The Treatybased limitation of subsidiarity to shared (and not exclusive) competence is not compelling from an economic perspective, because the considerations offered by the theory of system competition or fiscal federalism apply irrespective of the (artificial) legal delineation. Surely the competences accorded exclusively to the EU are those where economies of scale (customs administration), reduced transaction costs (internal market), or avoidance of negative externalities (single monetary policy) are rather intuitive. However, exclusive competences are carved into the Treaties without exception, which is not convincing for each and every case. Take as an example how euro members (exclusive competence) would have fared through the euro crisis with individual currencies, allowing them flexibility to devalue or appreciate their currencies, and thus facilitate economic adjustment. While hypothetical in nature, it is not obvious that a single currency in all circumstances is the adequate regime— Treatybased exclusive competence may not always meet the economic subsidiarity test. 82 On the economics of subsidiarity, see Roger van den Bergh, ‘The Subsidiarity Principle in European Community Law: Some Insights from Law and Economics’ (1994) 1 Maastricht Journal of European and Comparative Law 337; on the legal contours of subsidiarity, seed Alan Dashwood, ‘The Relationship between the Member States and the European Union/ European Community’ (2004) 41 Common Market Law Review 355, 356. 100 armin steinBach The first prong of the subsidiarity test— does the issue concerned fall under shared competence?— is a purely legal exploration, one that applies competence delineation techniques based on Treatybased competence allocation. EU lawyers typically assess the objective of the policy instrument in order to identify the corresponding competence.83 What matters for the ECJ in order to determine which competence is relevant is the objective pursued by EU legislators which motivates the use of a Treatybased instrument.84 The objective pursued may be controversial at times— think of the German Bundesverfassungsgericht questioning the ECB’s objective to pursue monetary policy. The Bundesverfassungsgericht considered ECB bond purchasing to be driven by economic policy objectives (Member States’ competence), not monetary policy (EU competence).85 The second and core element of the test can be conceptualized as costbenefit analysis informing whether the optimal level of allocation of a given shared competence lies at the Member State or EU level.86 If these conditions are met, the EU may pursue policies at the central level. Lawyers tend not to engage in a rigid economic assessment of costs and benefits. Rather they apply common sense reasoning by vaguely offering an account of how different considerations might support either EU or Member States competences.87 A more rigid economic approach is to characterize the benefits of centralization as the possibility of exploiting positive externalities, avoiding negative externalities, or leveraging economies of scale in the central allocation of policy responsibilities— on one side. On the other side, it would principally account for the costs of harmonization, as those are associated with overruling heterogeneity of preferences. Balancing the benefits with the varying preferences of the citizenry, the optimal degree of centralization should ensue: all functions where positive effects are dominant should be held at the central level, whereas all functions where high heterogeneity of preferences exceed the benefits should be kept at the local level.88 83 Sionaidh DouglasScott, Constitutional Law of the European Union (Longman Pearson Publishers 2002) 261; Stephen 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. 84 Case C370/ 12 Pringle (2012) EU:C:2012:756 [53]; Case C62/ 14 Gauweiler and others v Deutscher Bundestag (2015) EU:C:2015:400 [46]. 85 BVerfGE 154, 17, 119 PSPP (2020); following Case C493/ 17 Heinrich Weiss and others (2018) EU:C:2018:1000. 86 Jacques Pelkmans, ‘Testing for Subsidiarity’, Bruges European Economic Policy (BEEP) Briefing 13/ 2006 (2006). 87 Gracia Vara Arribas, ‘Subsidiarity and EU Added Value: The Difficulty of Evaluating a Legal Principle in a Pragmatic Way’ (2020) 3 European Court of Auditors Journal 30. 88 Parisi and Fon (n 19); Weatherill (n 12) 13. centralization 101 (Dis- )economies of scale Relevant economic categories for operationalizing the subsidiarity principle are economies of scale, diseconomies of scales, and economies of scope. Economies of scale are discernible from the logic of the Tiebout model, which states that decentralized provision of public goods is preferred where relatively small populations are sufficient for the efficient provision of the service. But federalism theory and its preference for small governance units does not square well with other government services, where economies of scale allow for costefficiency when centralizing the provision of the public good. Examples include national defence, electricity, and telecommunication networks; transnational infrastructure; monetary policy; research; and pandemic health action. Because of economies of scale (in addition to externalities), these goods and services can more efficiently be provided to large populations. Centralization may enhance efficiency.89 However, some policy functions may be associated with diseconomies of scale at the central level. This concept captures the idea that there are some things that are best done at the Member State level. There are different kind of costs that occasion the transfer of competences; not all are genuinely economic in nature. Political theory points at values of economic and procedural fairness, democratic participation, and the protection of personal rights and liberties. Regarding democratic participation, there is extensive literature dealing with the question of which level government activities should be placed in order to maximize the impact of constituents on actual decisionmaking. Giving more local selfgovernance increases participation and thus increases chances that public policies align with local preferences, which is the ultimate objective from an economic perspective.90 If overlapping jurisdictions exist, people with control over policy may adopt measures that affect other people who have no control over policy, and this may engender a transfer of wealth from the group with less control to the group in control.91 For example, citizens in subregion A may wish to be offered a different set of public policies than in subregion B. Their needs on infrastructure, security, and consumer protection may diverge. Offering onesizefitsboth services leaves citizens in both regions partially unsatisfied, as they have to bear costs for services offered to others. Government ‘by the people’ is thus not only a core idea of democratic governance applicable to states in their entirety, but also offers an economic rationale for lowerthanfederallevel governance. This concern is inherent in the sceptical views of legal and political science scholarship that lament the lack of EU 89 Inman and Rubinfeld (n 27) 88; Alberto Alesina, Ignazio Angeloni and Ludger Schuknecht, ‘What Does the European Union Do?’ (2005) 123 Public Choice 275, 53. 90 Robert Alan Dahl and Edward R Tufte, Size and Democracy (Stanford University Press 1973). 91 Posner and Sykes (n 46) 15. 102 armin steinBach democratic legitimacy.92 A recurrent argument is that supranational decisionmaking in Brussels decouples decisionmaking and legitimacy, with the double source of legitimacy from Member State representation in the Council and the representation in the European Parliament as directly elected body offering insufficient say to European citizens. Conceptually, this argument builds on the idea of citizens as a nationally determined group, as discussed above. A European ‘demos’ does not exist in that perspective, and as long as debates in media and public discourse remain predominantly national, European decisionmakers will remain notoriously deficient in terms of democratic legitimacy. This view is prominently represented, as mentioned, in the German Constitutional Court and in parts of legal scholarship. It serves as a counterargument opposing the transfer of competences that are particularly reliant on democratic input.93 This view connects easily to emphasizing the value of participation at a subfederal level. Diseconomies of scale also provoke the following question: which competences are nonamenable and should remain at Member State level? One could argue that the protection of personal rights and liberties should remain decentrally protected. Isaiah Berlin refers to ‘negative rights’, such as physical integrity, which should be secured well at local level through local police and local jurisdiction. The logic is that certain rights are best protected by local jurisdictions coupled with the free choice of citizens as to their place of residence.94 A connected view is the ‘transfer barrier’ that the German Constitutional Court erected as a red line for giving up national competencies. In this logic, the realization of fundamental rights, above all as regards intensive encroachments on fundamental rights such as the deprivation of liberty in the administration of criminal law, touches core democratic affairs. As a general rule, intensive encroachments on fundamental rights have to be based on national parliamentary decisions and should remain a prerogative of national parliaments to which citizens participate through election. Similarly, the French Constitutional Council stated that changes to the European Treaties in the sense of further competence transfers may be acceptable provided that they do not undermine the essential conditions for the exercise of national sovereignty. Among these ‘essential conditions’ are the fundamental rights and liberties of nationals.95 With the sensitivity of the relationship between legitimacy and fundamental rights being so high, decentralized participation thus ensures democratic safeguarding of fundamental rights. While this is convincing for negative rights, one may argue the contrary in relation to positive rights such as equal access to education, healthcare, 92 Grimm (n 57) 117; Crum and Merlo (n 85). 93 Joseph HH Weiler, ‘Does Europe Need a Constitution? Demos, Telos and the German Maastricht Decision’ (1995) 1 European Law Journal 219. 94 Isaiah Berlin, Four Essays on Liberty (Oxford University Press 1969). 95 Peter van Elsuwege, ‘The EU Constitution, National Constitutions and Sovereignty: An Assessment of a “European Constitutional Order” ’ (2004) 29 European Law Review 741, 746 with reference to the decision on 9 April 1992, Conseil Constitutionnel, décision No 92308 DC. centralization 103 and perhaps a safe physical environment. Equal access to achieve one’s full potential will be best protected within democratic federalism with responsibility assigned to a strong central government. Framing the legal concerns within our economic framework, one could say that there are switching costs associated with the transfer from Member States to the EU level. These costs, primarily political and social, vary depending on the nature of the competence concerned. Hence, Member States cannot ignore the barriers erected by constitutional courts blocking the transfer of sovereign rights. Pushing integration further may cause diseconomies of scale to the extent that the costbenefit assessment under the subsidiarity principle turns negative. Another source of diseconomies of scale may result from vanishing rules competition. The inclusion of rules competition as consideration for the subsidiarity principle militates against centralization. The logic is that rule competition (or interjurisdictional competition) serves as a discovery tool to identify regulatory paths of cost minimization that would remain untapped under the harmonization of rules.96 Unlike centralization, rules competition is not about harmonization of rules, but rather rules sharing. Rules competition leads ultimately to superior rules, because better rules in one Member State are imitated in countries with inferior rules. Consequently, successful and effective rules employed in one Member State have a higher chance of being adopted by other Member States by means of legal transplantation or by application of choiceoflaw rules. In these cases, it is not centralization, but rather Member States’ legislation that prevails. One may thus identify an additional diseconomy of scale of full centralization at the EU level given by the disappearance of virtuous rules competition, which could provide ways to minimize the costs of regulation at the local level but also, by extension, at the EU level. Rules competition thus makes it more difficult for a proposed centralization to pass the subsidiarity test. Applying a dynamic efficiency, that is, one that incorporates the cost advantages of rules competition through dynamic interaction between Member States, will add diseconomies of scale at central level. Incorporating the dynamic effect of rules competition in the subsidiarity test offers a factor that counterbalances the centralization push of the subsidiarity principle. Economies of scope While the rationale behind (dis)economies of scale is intuitive and wellresearched, economies of scope must be taken into account as well. Economies of scope refer to the added value of concentrating the combination of two or more policies at the same level of government (federal or subfederal). This is because creating and 96 Gerhard Wagner, ‘The Economics of Harmonisation: The Case of Contract Law’ (2002) 3 ERA Forum 77, 79; Weatherill (n 12) 13. 110 armin steinBach EU exclusive competence that restricts the scope of action of Member States (eg in trade issues); and using the freedoms granted under international law that surpass the constraints of the EU community methods (eg through intergovernmental cooperation1). The degree of flexibility marks a significant difference between international law and European law. Soft law, vagueness, and reduced enforcement are recurrent patterns of international law.2 International law dispute resolution rarely binds in the same ways as the sanctionbased adjudication under EU law, and a ‘quasiexecutive’ like the EU Commission is unparalleled in international law. Finally, Treaty reservations are a lot more common in international law than under EU law.3 The legal literature has addressed flexibility of law through the lens of systemically analysing the contrast to the more rigid types of law in terms of bindingness, decisionmaking, and the sources of law employed.4 In turn, economic approaches to flexibility have asked why states retain flexibility through Treaty design in relation to international law, even if each state has an interest that all countries commit to and comply with a certain substantive standard.5 One straightforward answer is that states seek to retain scope for uncooperative and unilateral scopes of action, even if this flexibility comes at the cost of other Treaty members being lax in their commitments too, through reciprocity of commitments, which engenders lowering the overall gains from cooperation. The reciprocity principle governing international law thus poses a barrier to too much flexibility in the agreement. Beyond this rather intuitive case, parties may also prefer flexibility when they anticipate significant costs associated with a breach that a country is likely to make in the future (eg for domestic political reasons). Since these violation costs could undermine the cooperative gains on a more permanent basis, under international law a country may choose a less rigid commitment as a Treaty provision in the first place, even if this means foregoing the benefits of rigid application of 1 Sandrino Smeets, Alenka Jaschke, and Derek Beach, ‘The Role of the EU Institutions in Establishing the European Stability Mechanism: Institutional Leadership under a Veil of Intergovernmentalism’ (2019) 57 Journal of Common Market Studies 675; Sandrino Smeets and Derek Beach, ‘ “It Takes Three to Tango”: New InterInstitutional Dynamics in Managing Major Crisis Reform’ (2022) 29 Journal of European Public Policy 1414. 2 Chris Brummer, ‘Why Soft Law Dominates International Finance— and Not Trade’ (2010) 13 Journal of International Economic Law 623, 630; Michael Reisman, ‘The Concept and Functions of Soft Law in International Politics’ in Emmanuel G Bello and Prince Bola A Ajibola (eds), Essays in Honour of Judge Taslim Olawale Elias (Brill 1992) 135. 3 Andrew Guzman, How International Law Works (Oxford University Press 2008) 22, 33, 71, 119, 183; However, Koremenos’ empirical account highlights that reservations are a lot more common in human rights agreements than in economics agreements, Barbara Koremenos, The Continent of International Law: Explaining Agreement Design (Cambridge University Press 2016) 3. 4 CM Chinkin, ‘The Challenge of Soft Law: Development and Change in International Law’ (1989) 38 International and Comparative Law Quarterly 850; Michael Hahn, ‘Interesting Times: Soft Law in International Economic Governance’ in Manjiao Chi, Marc Bungenberg, and Andrea K Bjorklund (eds), Asian Yearbook of International Economic Law 2022 (Springer 2022). 5 Alan O Sykes, ‘The Economics of Public International Law’ (2004) John M Olin Program in Law and Economics Working Paper No 216. flexiBility 111 international law.6 This is likely to be different under EU law, which has, though flexibility exists, committed to a comparatively stricter regime with a more compelling enforcement mechanism in place. One obvious difference justifying more need for flexibility at the international law level than at the EU law level is that preferences of EU members are more homogenous than at the global level which makes exception and escape clauses less (often) necessary.7 Also, costs of retaliation are relevant: under international law, there is no centralized sanctioning entity that would either bear or mutualize the costs of retaliation between Treaty members, but sanctions are imposed largely through bilateral relationships and therefore also impose costs on the sanctioning party (not only on the sanctioned party)— rational sanctioning parties may want to avoid these costs and thus choose more flexibility in the Treaty design in the first place.8 Note that even under the comparatively advanced WTO dispute settlement mechanism, retaliation occurs bilaterally between the parties concerned, not through centralized sanctions. By contrast, in the EU there is an enforcement mechanism (put in place by the Commission and the ECJ) that assumes the costs of retaliation at the central level, hence Treaty parties face lower costs of enforcing rigid Treaty provisions compared to the predominantly bilateral sanction structure of international law. a) Exceptions and escapes clauses Focusing on specific forms of flexibility, Treaty designers must choose whether to craft vague Treaty language that gives leeway for discretion. Alternatively, the Treaty may foresee explicit flexibility clauses. Depending on the need for flexibility, adaptive or transformative clauses can be integrated into the Treaties.9 Adaptivity captures the notion of allowing flexibility within the existing agreement in order to allow for efficient nonapplication of the rules. Adaptive clauses exemplify flexibility by allowing members to respond to unanticipated political events, economic shocks, and special domestic circumstances while preserving existing institutional arrangements. Sticking strictly to the Treaty commitment in this case would incur high compliance costs if an efficient decision would be noncompliance, while blatantly breaching the rules would incur violation costs associated with disrespecting the binding agreement with detrimental longterm effects on the credibility of the rules. 6 Guzman (n 3) 135. 7 This is also in line with regional cooperation being generally more likely to succeed than at the global level, Armin Steinbach, ‘The Trend towards NonConsensualism in Public International Law: A (Behavioural) Law and Economics Perspective’ (2016) 27 European Journal of International Law 643. 8 Guzman (n 3) 140. 9 Barbara Koremenos, Charles Lipson, and Duncan Snidal, ‘The Rational Design of International Institutions’ (2001) 55 International Organization 761, 773. 112 armin steinBach Under these circumstances, adaptive clauses thus reduce the overall level of compliance by allowing deviations from the desired commitment level to take place in a lawful manner so as not to force a Treaty member to violate the rules. Adaptive clauses reduce the total costs incurred in the event of noncompliance with the substantive Treaty commitment. In other words, if compliance costs are so high that parties would breach the Treaty regardless of whether an escape clause exists, having an escape clause is better than not having one.10 However, from an economic perspective, adaptive clauses should be activated only in cases where it is efficient for a state to deviate from their obligations, and not, for example, give in to lobbying pressure by interested parties.11 We can further differentiate between types of adaptive clauses, specifically between exception and escape clauses. Exception clauses introduce flexibility to allow hardpressed states to avoid the full burden of their Treaty obligations on a decentralized basis. Within existing Treaty boundaries, unilateral invocations of flexibility, such as exception clauses, can be costly but necessary for countries to retain a positive payoff of commitment to EU rules. Situations may arise in which the benefits of domestic regulation curtailing EU law commitments may be high. In such situations a government may want to pursue restrictive measures. To allow such discretion, states may converge on accepting certain motives that reflect highcost domestic concerns that could allow restrictive measures. To that end, EU members agreed to accept measures that were ‘justified on grounds of public morality, public policy or public security’ (Article 36 TFEU). This provision is an exception clause allowing Member States to deviate from their internal market obligations. It is sensible to allow these exceptions (under the restrictions mentioned) as they can lead to efficient results. Take for example the import of an unsafe product— the efficiency question is whether the overall costs of allowing the import are smaller than the benefits of an import ban. The exclusion of such a product would normally be worth more to an importing state than the sale of an unsafe product would be to an exporting state.12 Hence, the existence of this escape clause increases the value of the internal market provisions in the Treaty. Likewise, take the exception ground to protect ‘order and security’— if risks occur to life and health, a negative externality would be attached to the imported good that the bearer of the basic freedom inflicts on the society as a whole. In such a case, the exception clause allows an efficiency assessment, since the costs to society are higher than the benefits of the individual service provider. Furthermore, Article 36 TFEU requires that prohibitions based on the grounds numerated in this provision must not be ‘a means of arbitrary discrimination or a disguised 10 Guzman (n 3) 152. 11 Alan O Sykes, ‘Protectionism as a Safeguard: A Positive Analysis of the GATT Escape Clause with Normative Speculations’ (1991) 58 University of Chicago Law School 255. 12 Guzman (n 3) 151. flexiBility 113 restriction on trade between Member States’. These impermissible trade restrictions would enhance the costs by further imposing unnecessary costs on trading partners. Note that a crucial institutional device is that the authority to determine what constitutes a justifiable ground for a Treaty exception must not lie with the state invoking the measure. Giving too much discretion to the state would invite abuse and protectionism13— in order to allow for objective and neutral legal assessment (ideally coinciding with an economic efficiency analysis), the EU Commission and the ECJ as representatives of Union interests must specify the terms and scrutinize states’ interpretations of the legal exception. Particularly, the independent ECJ has a pivotal role in defining the conditions, given that public choice literature suggests that exception clauses are prone to regulatory capture with rentseeking efforts focusing on invoking exception for distributive purposes. Exception clauses should be distinguished from escape clauses, the latter being a frequent instrument used by international organizations. Activation of escape clauses may lift obligations under the international Treaty on a temporary basis, while exception clauses offer punctual justification of Treaty violations. Take international trade agreements as an example. Trade agreements incorporate escape clauses (known as ‘safeguards’14) allowing states to enter into agreements they might not otherwise accept because of unforeseeable contingencies. Escape clauses indicate a lower degree of homogeneity in preferences between Treaty parties than is necessary for exception clauses. No EU Member State could temporarily lift the freedom of movement in its entirety. While EU Member States may escape the Treaty commitments altogether by taking recourse to Article 50 TEU, paving the way for exiting the Union if the Union’s obligations are perceived as infringing on sovereignty issues, EU Member States have accepted a higher level of reciprocal commitment by not allowing general escape clauses comparable to international law. While preferences may diverge between EU members depending on the issue area concerned, they do not fall back behind a minimum level of commitment. EU Member States have an incentive to invoke this flexibility in specific cases, but most take into consideration the reciprocal effect meaning that other members may invoke the exception as well. b) Transformative clauses While eventbased adaptivity clauses separate the outlying cases from the ordinary course of policy action, there are transformative clauses that allow the EU to 13 On the controversy on the right to selfdetermine security concerns under WTO rules, see Roger P Alford, ‘The SelfJudging WTO Security Exception’ (2011) Utah Law Review 697. 14 Krzysztof J Pelc, ‘Seeking Escape: The Use of Escape Clauses in International Trade Agreements’ (2009) 53 International Studies Quarterly 349. 114 armin steinBach respond to new policy challenges. Transformative clauses appear in the EU Treaties in different guises, but they respond to the same need. EU states are confronted with uncertainty about the future state of the world.15 This causes a dilemma when entering into EU Treaties: becoming locked into an institution may lead to unanticipated costs or adverse distributional consequences when the environment changes in the future. This dilemma may lead states to refrain from joining the EU at all, if uncertainty is high and anticipated benefits are low. Riskaverse states, in particular, will avoid committing themselves to rigid institutions.16 Assuming that gains from cooperation by committing to the EU Treaties are sufficiently high, how would states ensure that the EU has sufficient flexibility to adapt to changing conditions, while maintaining the will of EU members as the decisive reference point? One avenue is Treaty changes. Over time, Members have adapted to changing circumstances through Treaty changes including those made by the Maastricht, Amsterdam, Nice, and Lisbon Treaties. Treaty amendments are suitable when internal flexibility of the Treaty is insufficient to deal with a changed environment. For example, the enlargement of the EU by acceding new members or acquiring one single currency was not possible through internal flexibility. However, external flexibility through Treaty amendments has a downside. Renegotiation of Treaty terms is costly in terms of negotiation expenses, and it also invites strategic bargaining where states may ‘hold up’ the cooperation in an effort to increase their own gains from the renegotiated Treaty terms. EU law offers a number of transformative clauses. Take Article 352 TFEU as an example of a transformative clause— it allows the EU to take appropriate measures, if necessary, to attain one of the objectives set out in the Treaties. It is ‘designed to fill the gap where no specific provisions of the Treaty confer on the Community institutions express or implied powers to act, if such powers appear none the less to be necessary to enable the Community to carry out its functions with a view to attaining one of the objectives laid down by the Treaty’.17 Where the EU holds competence in one issue area, it should be flexible enough to extend its competence to issues that have sufficient links to this core competence. Take the EU’s competence for competition law as one example. The Treaty norms initially provided insufficient control over conduct that is incompatible with undistorted competition envisaged in the Treaty, and therefore through Article 352 TFEU, EU powers of action were extended with regard to concentrations.18 An economically sound approach taking account of the fact that crossborder implications of mergers should not be dealt with at the decentralized Member States level due to policy spillovers. 15 Koremenos (n 3) 39. 16 Koremenos, Lipson, and Snidal (n 9) 793. 17 Opinion 2/ 94— Accession of the Community to the European Convention for the Protection of Human Rights and Fundamental Freedoms (1996) I– 01759 [29]. 18 Council Regulation (EC) No 139/ 2004 of 20 January 2004 on the control of concentrations between undertakings (the EC Merger Regulation) [2004] OJ L24/ 1, para 7. flexiBility 115 Using transformative clauses may thus be aligned with allocative efficiency considerations under the legal constraint that new EU tasks remain within the boundaries of granted competences (complying with the principle of conferral). There are multiple other provisions in the EU Treaty that allow Member States space to bypass the Union’s constraints in order to develop the EU in a transformative way. The solidarity clause in Article 122 TFEU has played an important role in developing macroeconomic tools that produced welfare gains for the entire EU through means that were not explicitly foreseen by the Treaties. This legal provision was employed to furnish financial assistance during the pandemic to establish a risksharing instrument that permitted a coordinated macroeconomic response to the pandemic crisis (NGEU).19 The crisisinduced transformative character of this provision has been criticized for transgressing the competences accorded to the EU. In addition, it is not only Treaty provisions that offer suitable clauses granting space for exceptions and transformations of existing arrangements. The Commission and the Court play an important role too, especially in extending the scope of transformative EU competences. As mentioned before, effet utile or ‘implied powers’ are doctrinal innovations pushed by Union institutions to enhance flexibility within existing rules, hence giving an interpretation to EU law that ensures the greatest possible effect of EU law in Member States’ legal orders (see above Chapter 10 a)). With interests between Union institutions and Member States diverging on the desirable outreach of EU law into Member States’ legal orders, public choice insight informs us that flexibility enshrined in legal rules can be subject to misuse. Not only do individual states have incentives to freeride on an agreement by selfserving interpretations of flexibility clauses, but Union institutions can do the same. Driven by Community interest and the desire for prestige, the EU Commission and the ECJ have frequently been accused of being biased in favour of Community interests.20 From that perspective, the relevance and practical use of flexibility clauses and judicial doctrines under EU law must be viewed from a public choice angle (see below Part IV). c) Soft law The choice between hard law and soft law is one through which Treaty designers must balance a tradeoff between rigidity and flexibility.21 The emergence and 19 Council Regulation (EU) 2020/ 2094 of 14 December 2020 establishing a European Union Recovery Instrument to support the recovery in the aftermath of the COVID19 crisis [2020] OJ LI433. 20 Gareth Davies, ‘Subsidiarity: The Wrong Idea, in the Wrong Place, at the Wrong Time’ (2006) 43 Common Market Law Review 63, 64; Nicholas W Barber, ‘Subsidiarity in the Draft Constitution’ (2005) 11 European Public Law 197. 21 Francis Snyder, ‘The Effectiveness of European Community Law: Institutions, Processes, Tools and Techniques’ (1993) 56 Modern Law Review 19, 54. 116 armin steinBach characteristics of informal institutions and soft lawmaking in global governance— as well as the pressure they exert on the traditional modes of cooperation— have long been analysed with regard to international law22 as well as EU law.23 Soft law generally captures rules of conduct which, in principle, have no legally binding force but which nevertheless may have legal implications and compliance pull.24 On a theoretical level, the most extreme form of soft law cooperation in terms of reduced binding character and formality could be seen in the renouncing of an explicit agreement completely. Legal obligations may be selfenforcing in the sense that parties may not consider it necessary to agree on a formalized commitment. More generally, whether an issue is dealt with through soft or hard law, a written or oral agreement, or agreed at all, depends on the issue at stake. In a coordination game, in which EU members share a common interest in not pursuing certain conduct, hence in which there are only benefits and no costs, and where defection is not attractive, there may be no need to specify an obligation in formal law, but it could be arranged tacitly or through soft law. This may explain why certain fundamental obligations in interstate conduct are found in public international law, in which the international community is more heterogenous and interests among states are less aligned than in the EU, but not in EU law. The EU Treaties are less explicit than international law on core rules such as the nonintervention principle, the principle of nonviolation, and the equality of sovereign states. Lawyers may argue technically that these core principles are enshrined in the EU Treaty through Article 3(5) TEU and other provisions, but their lack of salience and explicitness may also be explained with reference to these obligations being of such a selfenforcing and obvious nature that EU members have not considered elaborating them in the Treaties— they are tacitly presumed to form the basis of cooperation. Other comparative patterns of soft law (and the corresponding absence of hard law) support the view that international law relies more heavily on soft law than EU law does. International soft climate law contrasts with European hard climate law— under international law, states fear the freeriding of other countries to an extent that it undermines accepting binding rules or because countries outside Europe (especially developing countries) consider costs of CO2 reduction as too harmful to their economies to enter into an ambitious reduction path; under EU law, states accept binding reduction commitments because freeriding within the EU is less probable due to rigid monitoring, possibly also with a view to reaping firstrunner gains from technological advances in climate technologies seizing global market shares. Competition law is another example of soft international law and hard EU law: antitrust rules at the international level are limited to informal 22 See, eg, Kenneth W Abbott and Duncan Snidal, ‘Hard and Soft Law in International Governance’ (2000) 54 International Organization 421. 23 Mareike Kleine, Informal Governance in the European Union: How Governments Make International Organizations Work (Cornell University Press 2013) 54. 24 Snyder (n 21); Guzman (n 3) 142. flexiBility 117 exchanges between antitrust authorities short of binding substantive and procedural rules and largely decentralized.25 There is no cooperation although global welfare would be enhanced under a uniform standard of antitrust governance ensuring a worldwide level playing field curbing monopoly rents and competitive biases. However, several issues related to jurisdictional and sovereignty claims may comprise a fundamental reason not to surrender national competences. Moreover, the uncertainty surrounding the design of a universal standard of antitrust governance and the scope of discretionary practice of national authorities form another barrier. All these factors translate into significant sovereignty and monitoring costs, thus rendering hardlaw consensualism an unattractive option.26 At the EU level, the contrary is the case— Member States have rendered their antitrust competences and passed them to the EU, motivated by welfare gains— even if in singular cases national champions and even national welfare may suffer. Besides greater homogeneity in national antitrust practices, it is the absence of a hegemonic mode of action which is embodied in the extraterritorial action of antitrust agencies, leading international power players such as the United States to be reluctant to relinquish their hegemonic position in economic governance, as losing extraterritorial outreach of their unilateral competition policies would be costly.27 While less frequent than under international law, soft law has nevertheless been a popular mode of governance throughout European integration. Provided in the Treaties under Article 288(5) TFEU, the advantages of EU soft law as being fast, flexible, easy to issue, and thus able to adapt to rapid evolutions and changes in policies, contrast with important legitimacy drawbacks, as soft law is hardly justiciable, and its legal effects are blurred.28 Since the 1990s, there seems to be a growing preference for procedural frameworks over substantive prescriptions, with different forms of coordination emerging in fields such as social and economic policy, employment, the environment, education, and research. The open method of coordination (OMC) has become a frequent mode of coordination favouring new forms of soft governance.29 From an economic perspective, we are interested in identifying the conditions under which governments prefer soft laws to the formal hard ones, and vice versa.30 It is true that from a rational choice perspective, a binary distinction between 25 Steinbach (n 7). 26 Yane Svetiev, ‘The Limits of Informal International Law: Enforcement, NormGeneration and Learning in the International Competition Network’ in Joost Pauwelyn, Ramses Wessel, and Jan Wouters (eds), Informal International Lawmaking (Oxford University Press 2012). 27 Charles P Kindleberger, ‘Dominance and Leadership in the International Economy: Exploitation, Public Goods, and Free Rides’ (1981) 25 International Studies Quarterly 242. 28 Snyder (n 21) 54. 29 Burkard Eberlein and Dieter Kerwer, ‘Theorising the New Modes of European Union Governance’ (2002) 6 European Integration online Papers. 30 From the perspective of international relations, see, eg, Armin Schafer, ‘Resolving Deadlock: Why International Organisations Introduce Soft Law’ (2006) 12 European Law Journal 194; From an economic perspective, see Stefan Voigt, ‘The Economics of Informal International Law— An Empirical 118 armin steinBach ‘binding’ law and ‘nonbinding’ law does not make sense, because both are chosen by states to facilitate cooperation and both impose some costs on a noncompliant party through the enforcement modes of reputation, retaliation, and reciprocity.31 A great amount of work has emphasized the functionality of informality, notably the flexibility it offers to parties implying lower sovereignty costs as well as its ability to integrate a broader range of actors and stakeholders than would be possible under formal approaches.32 Indeed, soft law often emerges in areas where sovereignty costs are high and serves to avoid (even higher) costs implied in formal and binding cooperation in these issue areas, which explains for instance the different degree of softness in competition policy cooperation under international and EU law mentioned above. Also, while informality and soft law are conceptually different, they often coincide. For example, where domestic political costs associated with EU legislation are high, Member States may seek informal ways of decisionmaking at the EU level.33 By contrast, hard law is more restrictive as it generates reliance and often exerts (though not always) a greater compliance pull.34 Hence, formal law restrains policy spaces more significantly than informal conduct which allows for a broader range of conduct and cooperation mechanisms.35 It is not only sovereignty costs that are higher when hard law entails the countries’ acceptance of an external authority over political decisions (eg the Commission and the Council sanctioning an EU member for noncompliance with fiscal rules).36 Soft law also incurs comparatively less negotiation costs as it offers flexibility and speed both in the conclusion and the endorsement of nonbinding guidelines.37 Accordingly, flexibility due to the changing nature of circumstances may be a valid concern implying the desire for low modification costs.38 Therefore, if states prefer greater adaptability in an agreement, they may take recourse to soft terms in order to make rights and obligations under an agreement more flexible. For example, coordination of national economic policies on the basis of Broad Economic Policy Guidelines (Article 121 TFEU) or the practice of EU Assessment’ in Thomas Eger, Stefan Oeter, and Stefan Voigt (eds), The Economics of Informal International Law (Mohr Siebeck 2011) 33. 31 Guzman (n 3) 160. 32 Charles Lipson, ‘Why Are Some International Agreements Informal?’ (1991) 45 International Organization 495, 495. 33 Kleine (n 23). 34 Dinah L Shelton, ‘Introduction’ in Dinah L Shelton (ed), Commitment and Compliance: The Role of NonBinding Norms in Commitment and Compliance: The Role of NonBinding Norms in the International Legal System (Oxford University Press 2000) 8. 35 Brummer (n 2) 632. 36 Michael C Dorf, ‘Dynamic Incorporation of Foreign Law’ (2008) 157 University of Pennsylvania Law Review 103, 133. 37 Jacob Gersen and Eric A Posner, ‘Soft Law: Lessons from Congressional Practice Soft Law: Lessons from Congressional Practice’ (2008) 61 Stanford Law Review 573, 589. 38 Janet Koven Levit, ‘A BottomUp Approach to International Lawmaking: The Tale of Three Trade Finance Instruments’ (2005) 30 Yale Journal of International Law 125, 179. flexiBility 119 employment policy based on Commission guidelines (Article 148(2) TFEU) leave Member States with ample flexibility and only require them to take these guidelines into account. The OMC is the dominant governance mode in these areas, with accountability being more horizontal than vertical, meaning that Member States are required to take seriously, and to answer the preferences, objections, and counterproposals of other governments.39 This kind of adaptability is most practical where parties to an agreement are faced not only with high sovereignty costs but also with significant uncertainty over factual future developments and where, in economic terms, the costbenefit balance of the agreement remains to some extent unpredictable. Member States can identify the effects of rules and guidelines in practice by avoiding formal legality in assessing their benefits.40 In this event, soft law offers strategies for learning processes, in which parties can eventually resolve their problems.41 Taber has pointed at instances where EU hard law has been softened to give more manoeuvring space to policymakers. Some instruments of environmental policy have shifted from the traditional approach based on the setting of uniform, legally binding norms to a less coercive and more flexible approach.42 For instance, the Directive on Integrative Pollution Prevention and Control, enacted in 1996, introduced soft, nonbinding targets and a strong procedural component through the delegation of policy formulation to participatory, coregulatory networks, in a field where legally binding emission limit values on air, land, and water used to be applied to several industrial sectors.43 Clearly, informal law typically provides the executive branch with more space.44 In particular, informal laws with a lower profile are more strictly monitored by the government bureaucracies that negotiate and implement the agreements and are less exposed to intrusion by other agencies or parliamentary bodies. Hence, recommendations, benchmarking, and best practices are based on the desire of participants to agree, through collective deliberation, on procedural norms, forms of regulation, and shared political objectives, while preserving a diversity of solutions and local measures. Finally, the reputational costs of informal law are comparatively low. Consider that noncompliance with EU hard law typically triggers scrutiny by the Commission and eventually leads to infringement procedures. Since 39 Fabien Terpan, ‘Soft Law in the European Union— the Changing Nature of EU Law’ (2015) 21 European Law Journal 68, 81. 40 Abbott and Snidal (n 151) 442. 41 Brummer (n 2) 633. 42 Katharina Holzinger, Christoph Knill, and Ansgar Schäfer, ‘Rhetoric and Reality? “New Governance” in EU Environmental Policy’ (2006) 12 European Law Journal 403. 43 Terpan (n 39) 90. 44 See also Andrew T Guzman, ‘The Consent Problem in International Law’ (2012) 52 Virginia Journal of International Law 747, 763 (comparing the international sphere with the domestic setting in which power is shared between legislation and administration. Given that the administration, not the legislator, domestically deals with all issues, not all issues at the international level can be managed with international agreements). 126 armin steinBach work when states determine their policy course, which is relevant for differentiated integration. On a theoretical level, nonconsensualism for our purposes captures the difference between the number of states and membership, in our case between EU states and those who actually participate in integration while being EU members. As Keremone and others emphasize, the number of states refers to the exogenous feature of the policy issue concerned.5 The number of states refers to the set of states interested in participating in cooperation. In many, not all, issue areas, the natural default would be to consider cooperation among all EU members. By being EU members, there is a presumption that all members will participate in the full integration programme of the Treaties as well as in enhanced cooperation. However, we discuss various forms of nonconsensualism, that is, where EU members do not consent to the same pace of integration. This can be determined through EU primary law by way of Treaty optouts; it can occur through Treatybased variability in integration, notably through ‘enhanced cooperation’ between a subgroup of EU members as ‘a last resort’ when cooperation is not feasible among the EU as a whole (Article 20 TEU); variable pace can also be channelled through interse agreements sidelining EU law with EU members acting in their traditional international law sovereignty outside of the supranational architecture; and differentiation finally is a common feature during crises, with the euro crisis as an illustrative example of variable forms of cooperation. Our hypothesis is that nonconsensualism varies according to multiple factors, mostly due to interests, preferences, variable gains associated with cooperation, but also due to relative power between EU members enabling powerful countries to craft agreements in their favour. These differences in terms of preferences, implementation, and power imbalances determine the degree of asymmetry between parties, a relevant factor in the gametheoretic analysis of parties to cooperation. a) EU Treaty optout In the context of international agreements, the practice of Treaty reservations is widespread under international law.6 By contrast, the founding Treaty of Rome foresaw few special protocols and derogations for individual countries. It was only the Treaty of Maastricht that gave differentiated integration— that is, the adoption of EU norms that do not bind all EU members— a Treaty basis. It introduced the innovation of allowing some Member States not to fully participate in the Economic 5 Barbara Koremenos, Charles Lipson and Duncan Snidal, ‘The Rational Design of International Institutions’ (2001) 55 International Organization 761, 777. 6 Eric Neumayer, ‘Qualified Ratification: Explaining Reservations to International Human Rights Treaties’ (2007) 36 Journal of Legal Studies 397; John King Gamble, ‘Reservations to Multilateral Treaties: A Macroscopic View of State Practice’ (1980) 74 American Journal of International Law 372. non-consensual eu law 127 and Monetary Union. Optouts were granted to the UK and Denmark, but these were intended as only temporary exceptions.7 Likewise, Ireland and Denmark invoked exceptions in the domain of core state powers in Justice and Home Affairs, releasing these states from the binding effect of some but not all EU policies, such as the directives defining the rights of asylum seekers.8 Original signatory states crafting a new Treaty often face substantial costs in the process of Treaty negotiation and drafting whereas the costs of accession to an existing Treaty can be ambiguous: accession to an existing Treaty may incur less bargaining costs (in the EU because the acquis communautaire is not negotiable upon accession), while the sovereignty costs of the acceding state may be substantial due to the indispensable requirement of implementing the acquis communautaire into domestic law, that is, incorporating the entire body of EU law. Acceding states are not able to submit reservations to the vast amount of preestablished written and unwritten rules. Potentially, transitional applications of the acquis communautaire through an agreed timetable have been granted but the general principle is that no discriminatory membership in the EU is permitted. The few special protocols and derogations for single countries introduced with the EEC Treaty notwithstanding,9 optouts do exist but are scarce. This contrasts with the practice under international law: Article 21 of the Vienna Convention is the international law metric on how content and participation of Treaties can deviate from the principle of equal content applicable to all signatories. Reservations to a Treaty at the time of accession or ratification create multiple bilateral implications and transform the international agreement between the reserving state and nonreserving states. The reservation reduces the commitment level of the Treaty to the level agreed to by these two states. The economic perspective on these flexibilityincreasing Treaty elements is mixed. It has been argued that reservations, just as exit clauses and escape clauses, increase states’ willingness to enter into agreements.10 However, an increase in one state’s willingness to enter an agreement corresponds to a decrease in the value of the agreement to the other party accepting the reservation, lowering that party’s incentive to join the agreement. The law and economics literature has devoted attention to Article 21 of the Vienna Convention, and some of this analysis is of interest to the EU. It was shown that Article 21 tilts the balance in favour of highcost and lowbenefit states who can take advantage of the reciprocity mechanism. Highcosts refer to Treaty 7 Thomas Beukers and Marijn van der Sluis, ‘Differentiated Integration from the Perspective of NonEuro Area Member States’ in Thomas Beukers, Bruno de Witte, and Claire Kilpatrick (eds), Constitutional Change through EuroCrisis Law (Cambridge University Press 2017) 14. 8 Bruno de Witte, ‘The Law as Tool and Constraint of Differentiated Integration’, in Mark Dawson and Markus Jachtenfuchs (eds), Autonomy without Collapse in a Better European Union (Oxford University Press 2022). 9 Dominik Hanf, ‘Flexibility Clauses in the Founding Treaties, from Rome to Nice’ in Bruno de Witte, Dominik Hanf, and Ellen Vos (eds), The Many Faces of Differentiation in EU Law (Intersentia 2001). 10 Laurence R Helfer, ‘Exiting Treaties’ (2005) 91 Virginia Law Review 1579. 128 armin steinBach implementation and compliance costs and lowbenefits to the gains from fulfilment of the international obligations— poorer or developing states often face a highcost and lowbenefit combination.11 In line with that literature, poorer countries with high implementation costs have a comparatively strong position in multilateral Treatysetting, because they can ask for costreducing side concessions (provided that power imbalances are not at play).12 The practice of international Treaty law suggests that international agreements accommodate this situation by implementing various levels of Treaty obligations: for example, under WTO law, developing countries enjoy a preferential scheme; under climate change agreements, developed countries must undertake greater climate change mitigation efforts and offer sidepayments to less developed countries; and under the Montreal Protocol on OzoneDepleting Substances, developed countries had to subsidize the developing countries’ compliance costs. One plausible explanation for the absence of treaty reservations as common features under EU law may be the higher homogeneity among members in the EU compared to the level of international treaty law. The theory suggests that the more symmetric countries are in terms of preferences, costs, and implementation, the more similar a country’s payoff function of cooperation is. As a result, there is less need for reservations and they are more likely to align at the same level of Treaty obligations.13 In the gametheoretic perspective, optout strategies under EU law, as well as reservations under international Treaty law, imply the same strategic behaviour. During Treaty negotiation, in general, rational states anticipate that other states may potentially draw on Article 21 of the Vienna Convention, and by doing so lower the content level of the Treaty. This provision leads to matching reservations through reciprocity. The reciprocity mechanism in Treaty reservations disincentivizes strategic unilateral reservations, because they would engender mutual losses for all states involved, which is not a dominant strategy given a country’s payoff function. With reciprocity prevailing, each state would only lower their Treaty commitment leading to an overall inferior outcome for all parties. Hence, in the case of symmetric states with homogenous interests, full Treaty commitment becomes the equilibrium strategy for both states.14 The matching effects of Treaty reservations furthers levels of Treaty ratification that are higher than those that they would otherwise adopt in a Nash equilibrium.15 11 Francesco Parisi and Vincy Fon, The Economics of Lawmaking (Oxford University Press 2008) 52, 268. 12 This contrasts with bilateral Treaty settings such as bilateral investment treaties, in which power imbalances between larger and smaller states are significant, see Deborah L Swenson, ‘Why Do Developing Countries Sign BITs?’ in Karl P Sauvant and Lisa E Sachs (eds), The Effect of Treaties on Foreign Direct Investment (Oxford University Press New York 2009). 13 Andrew T Guzman, ‘The Consent Problem in International Law’ (2012) 52 Virginia Journal of International Law 747; Parisi and Fon (n 11) 246. 14 Jack L Goldsmith and Eric A Posner, The Limits of International Law (Oxford University Press 2006) 23, 243. 15 Parisi and Fon (n 11) 245. non-consensual eu law 129 In turn, incentives to request Treaty reservations are different when participating states are different. The reciprocitybased matchingreservation mechanism does not work when states face asymmetric incentives. Asymmetry— that is, diverse preferences, different implementation costs, significant power imbalances— invites unilateral reservations. Asymmetry implies heterogeneity causing states to have different costbenefit ratios from Treaty implementation. The difference in Treaty bargaining among symmetric states is that asymmetry of states leads to different payoff functions which feeds into varying preferences for levels of Treaty ratification. The commitment level under the Treaty is determined by the marginal benefit under matching reservations and the marginal cost at full ratification. Heterogeneity in payoff functions may, for example, play out in the EU when some acceding Eastern European members find it difficult to comply with the high ruleoflaw standards in the EU, which imply costly domestic policy changes. Asymmetry may also be caused by power imbalances, where powerful countries can more impose their preferences, because their outside options are simply wider, and they can pursue their objectives through unilateral action or intergovernmental agreements outside of EU law. Importantly, asymmetry causes a lower Treaty content level, and it is the highestcost states, the ones for whom Treaty implementation is particularly costly, who are responsible for this. Parisi and Fon have shown that the highestcosts states, unlike averagecost states, do not need to accept undesired reservations from other states. Through the reciprocity logic of Article 21 of the Vienna Convention, the highestcost states effectively determine the Treaty commitment level and attain their optimal Treaty terms in all bilateral Treaty relations with other states.16 This bears consequences for the preferred mode of cooperation. While some multilateral approaches introduce flexibility through differentiation (eg WTO rules), the trend towards regionalization and bilateralization highlights that Treaty membership is increasingly sought among homogenous states. Nonmultilateralism means the downsizing of participation towards greater homogeneity in payoff functions (and thus higher Treaty commitment levels). In turn, the reciprocity mechanism under Article 21 of the Vienna Convention ensures the social optimum only in cases where the members face homogenous payoff functions, or when all states prefer full ratification. What implications can we draw from this literature for the EU? The core ‘Westphalian’ legal premises of reciprocity and sovereign rights to cooperate are the same principles undergirding both EU and international law. The logic of Article 21 of the Vienna Convention also applies to the EU, yet reservations are empirically more widespread in international law. This difference appears largely as a result of asymmetry— for example, diverse preferences or differences in costs 16 ibid 265. 130 armin steinBach of compliance with the agreement. With greater homogeneity across the EU in terms of economic performance and cultural proximity, there is less likelihood for optouts to occur. However, cases exist where discrepancies are insurmountable and lead to optouts. The UK abstention from social policies, laid down in a social protocol attached to the TEU and prompted by the dogmatically partisan view of the British Conservative government reflected deeply divergent views on social matters.17 Similarly, currency matters concern core sovereign issues and the UK and Danish optout can be interpreted as emphasizing the fundamental division between the euroarea and noneuroarea Member States as reflected in both the institutional and substantive norms of EMU constitutional law. Likewise, Justice and Home Affairs is widely viewed as the domain of core state powers. A further particularity of the EU compared to the wider international community is the EU’s ability to bridge heterogeneity in one Treaty area by crosscompensation. From an economic perspective, there may be an effective sidepayment mechanism in place in the EU that would not be available for fieldspecific international agreements. ‘Sidepayments’ allow Member States to account for a crosssubsidization of treaties by offering concessions to a reluctant state in an area other than the one from which the country wishes to optout. Compensatory sidepayments may level out heterogenous preferences or may compensate highcost states. The availability of sidepayments as an instrument for levelling out a country’s different payoff function in relation to one area of cooperation (eg EU agricultural policy, internal market obligations) highlights a difference between EU law and international law: international agreements are in most cases made for specific subject areas (eg international trade, climate mitigation, and human rights). Their subjectspecificity limits the scope for sidepayments. Sidepayments may exist in some areas— think of the WTO ‘single undertaking’ approach allowing crosssectoral sidepayments— but for international agreements they are typically not only limited in substance but also institutionally, because international organizations rarely deviate from the onecountryonevote principle in terms of voting rights or member representation. This is different in the EU: the scope of policy fields in the EU is much larger and gives leeway to the coupling of unrelated concessions as sidepayments. For example, a special protocol attached to Ireland’s accession to the EU ensures that it will benefit from financial transfers, while Finland was granted favourable treatment under the European Regional Development Fund in order to provide support for its Arctic agricultural region.18 In addition, the EU, as a supranational construct, is built on an institutional architecture that accommodates structural differences between EU members. One obvious example of institutionally favouring highcost smaller states 17 Françoise de La Serre and Helen Wallace, ‘Flexibility and Enhanced Cooperation in the European Union: Placebo Rather than Panacea?’ (1997) Research and Policy, Papers No 2, 9. 18 Ibid. non-consensual eu law 131 over larger states is the allocation of voting rights in the Council and the allocation of seats in the European Parliament, which are biased in favour of smaller EU members.19 Highcost countries can be enticed into the EU, despite heterogeneity, through structural advantages in the decisionmaking architecture. What is scarce under international Treaty law— the availability of sidepayments as a mechanism for smoothing the heterogeneity of preferences and levelling out differences between highcost and lowcost states— is plenty in a European Union that allows side deals and concessions across a wide portfolio of issue areas. From a behavioural economics perspective, we can further say that ‘ambiguity aversion’ can be used to explain why optouts are chosen, especially when a rational choice analysis suggests that an equilibrium should favour multilateral cooperation due to cooperation gains. This is because actors are ambiguityaverse when probabilities cannot be easily predicted, hence they prefer known outcomes over unknown ones. The degree of uncertainty in EMU matters during the crafting of the Maastricht Treaty was considerable, and the sovereignty costs for the UK were high, as it was asked to abandon a stable currency. While optingout from the common currency upheld heterogeneity and higher transaction costs due to currency exchanges, any substantial change (going beyond informal exchanges) would create uncertainty about the applicable concept of price stability; hence, the effect of loss aversion is exacerbated. This loss aversion may play out differently in other situations. While Germany feared the threat of losing a stable and dominant currency, sidepayments in other fields of EU integration (eg an extended internal market and fiscal rules) attenuated this effect. b) Differentiating integration Treaty optouts are one form of nonconsensualism, but with the Amsterdam, Nice, and Lisbon conferences, the prospect of enlargement with a large number of new states engendered a broader debate on differentiated or ‘multispeed’ integration. At the end of the nineties, a general conviction among EU members emerged that the Treaty framework should be further developed in order to allow for ‘institutional flexibility’ or ‘closer cooperation’. The Nice Treaty text allowed for what was now named ‘enhanced cooperation’ to be launched by a qualified majority vote in the Council, except in common foreign and security policy. However, it was only with the adoption of the Lisbon Treaty that a formal framework for variable geometry through the enhanced cooperation mechanism was established.20 19 Jonathan Rodden, ‘Strength in Numbers?’ (2002) 3 European Union Politics 151. 20 Daniela A Kroll and Dirk Leuffen, ‘Enhanced Cooperation in Practice. An Analysis of Differentiated Integration in EU Secondary Law’ (2015) 22 Journal of European Public Policy 353. 132 armin steinBach Enhanced cooperation has been implemented five times to date; it is rare but feasible. The first instance concerned the regulation of transnational divorces. The second was the creation of a new EU patent system, agreed following a controversy in which Italy and Spain opposed EUwide participation due to disagreement over the language regime. The third instance concerned the twin regulations on judicial cooperation in matters of matrimonial property and registered partnership property. The fourth was the financial transaction tax, which obtained Council approval for enhanced cooperation (but was never put in effect). The fifth instance saw twentytwo Member States supporting the creation of the European Public Prosecutors Office. While enhanced cooperation takes place within the EU legal order, there are outside options available to EU members by resorting to international law agreements sidelining the EU institutional structure. Traditionally, legal scholarship has looked with mistrust on (bilateral) treaties between Member States, perceiving them as possible threats to the EU legal order, at least when they addressed issues that formed part of or were related to the EU legal order, applied to some (not all) Member States (‘partial agreements’), and did not use the EU institutions.21 However, Member States resorting to international agreements outside the EU legal order is in itself nothing new.22 Past inter se agreements include instruments such as the Schengen framework or the Prüm Convention, in the area of justice and home affairs, and the Social Policy Agreement. In these cases, a group of Member States decided to push integration further through an instrument of international law.23 In a sense, one could say that this form of flexibility has existed since the early days of the European integration process. Hundreds of bilateral and multilateral international treaties have been concluded between Member States of the European Union since the 1950s in areas such as tax law, environmental protection, defence, culture, and education. They typically occur in areas in which the European Union has no lawmaking competence at all, but also occur in areas in which the EU possesses shared lawmaking competences, where a set of Member States prefer to use their ‘share’ to conclude an agreement among themselves rather than acting within the framework of the European Union. These inter se agreements become a true alternative form of variable geometry when they serve the purpose of allowing a group of Member States to move European integration forward in the face of opposition from other Member States. More recently, inter se agreements between Member States reflect a general trend of intergovernmentalism, propelled by the euro crisis.24 While a number 21 Steve Peers, ‘Towards a New Form of EU Law? The Use of EU Institutions Outside the EU Legal Framework’ (2013) 9 European Constitutional Law Review 37, 40. 22 See also Bruno de Witte, ‘Using International Law in the Euro Crisis: Causes and Consequences’ (2013) Arena Working Paper No 4. 23 Alicia Hinarejos, The Euro Area Crisis in Constitutional Perspective (Oxford University Press 2015) 90. 24 ibid 111; Kenneth A Armstrong, Governing Social Inclusion (Oxford University Press 2010) 67. non-consensual eu law 133 of reforms to the architecture of EMU have been carried out in the framework of EU law, the Member States have decided to act to a large extent outside the EU legal order, tightening budgetary constraints, establishing new mechanisms of financial stability, and setting up a framework for economic adjustment for countries in fiscal trouble. This has been done notably through the Treaties establishing the European Financial Stability Facility (EFSF) and the European Stability Mechanism (ESM), both of which (like the stability Treaty) aim to supplement the EU law measures on EMU.25 There was controversy surrounding these ‘peripherical agreements’, particularly regarding the role of intergovernmentalism as the driving force behind them.26 This strategy is consistent with an intergovernmental model for the management of the euro crisis, which has stressed the centrality of national governments (in the European Council) and their freedom to act through agreements outside EU law, rather than the centrality of the EU institutional machinery and the potential of EU law to address the crisis.27 However, the limits of intergovernmentalism in managing the euro crisis effectively and legitimately have been repeatedly emphasized.28 Our inquiry aims to explain the use of differentiated integration. Scholars in the international relations29 and international law30 literature have emphasized that reducing complexity is a rationale to explain nonmultilateralism in international relations. Similar considerations apply for differentiated integration in the European Union. From an economic perspective, the act of abandoning full participation of EU members in favour of ‘enhanced cooperation’ or even extraEU intergovernmental approaches responds to the state’s individual benefits and costs involved in the bargaining process. In order to achieve welfare gains through cooperation, a welfare improvement requires parties to negotiate on how to compensate for losses by balancing out the gains and losses incurred by different parties. With the increasing number of participating states, the transaction costs on determining a compensation solution increase, impeding compensation solutions. Exempting those states who have a payoff function that does not allow sufficient 25 Steve Peers, ‘Towards a New Form of EU Law? The Use of EU Institutions Outside the EU Legal Framework’ (2013) 9 European Constitutional Law Review 37, 39; Fabian Amtenbrink and Menelaos Markakis, ‘Never Waste a Good Crisis On the Emergent EU Fiscal Capacity’ in Alicia Hinarejos and Robert Schütze (eds), EU Fiscal Federalism (Oxford University Press 2023), 183. 26 Bruno de Witte, ‘Using International Law in the Euro Crisis: Causes and Consequences’ (2013) Arena Working Paper No 4. 27 Federico Fabbrini, Economic Governance in Europe (Oxford University Press 2016) 110. See also Michele Messina, ‘Strengthening Economic Governance of the European Union through Enhanced Cooperation: A Still Possible, but Already Missed, Opportunity’ (2014) 39 European Law Review 404, 404. 28 Edoardo Chiti and Pedro Gustavo Teixeira, ‘The Constitutional Implications of the European Responses to the Financial and Public Debt Crisis’ (2013) 50 Common Market Law Review 683, 683. 29 James A Caporaso, ‘International Relations Theory and Multilateralism: The Search for Foundations’ (1992) 46 International Organization 599, 611. 30 Andrew Guzman and Beth A Simmons, ‘To Settle or Empanel? An Empirical Analysis of Litigation and Settlement at the World Trade Organization’ (2002) 31 Journal of Legal Studies 205. 134 armin steinBach matches with the group of states pursuing closer integration facilitates an agreement. Accordingly, ‘enhanced cooperation’ allows for varied membership enabling likeminded states to work together separately from a more heterogenous group. Differentiation can likewise be suitable for dealing with the sensitivity of a given policy subject. The integration of core state powers— such as defence, interior, monetary, and fiscal policies— has led to durable differentiated integration. Distributional and sovereignty concerns have been particularly salient drivers of variable geometry in the EMU (euro membership, ESM, and the Fiscal Compact).31 In this sense, differentiation serves to overcome the incompatibility of heterogeneity of preferences and unanimity in decisionmaking, as collective decisions become infeasible unless qualified majority voting prevails.32 The Schengen experience is still referred to as another model offering an example of both the potential of such agreements to overcome a blockage within the Union’s decisionmaking system, and the possibility for their later reintegration within the EU legal system. With the pooling of homogenous preferences leading to multispeed integration, negotiations and compensations can more easily be bartered in differentiated settings. As illustrated above with Treaty optouts, asymmetric countries as parties to the Treaty bear the risk of Treaty reservations, leading to a lower commitment level to the Treaty content. This supports the idea that countries with more aligned policy preferences are likely to agree on more ambitious Treaty commitments. On this basis, one may, though with an element of speculation, classify past instances of differentiated integration: ‘enhanced cooperation’ in transnational divorces was only possible between eighteen EU members because some countries had concerns that this law would lead to the recognition of samesex marriages as wellregistered partnerships, in countries where they do not exist— clearly a culturally sensitive issue involving high political costs. Enhanced cooperation on patents was supported by neartofull membership of twentyfive members, while cooperation on the European prosecutor concerned the sensitive issue of criminal law, a core sovereignty issue. In turn, the inter se agreements on EMU affairs are more illustrative of the very heterogenous effects of the sovereign debt crisis on euro members. In any case, in all these areas, likeminded countries have agreed to differentiated but higher Treaty commitments and were thus able to achieve equilibria that were easier to attain than a full membership agreement. It is not surprising that, while the initial EEC Treaty with six founding members contained some protocols, it 31 Frank Schimmelfennig and Thomas Winzen, Ever Looser Union? (Oxford University Press 2020); Philipp Genschel, Markus Jachtenfuchs, and Marta Migliorati, ‘Differentiated Integration as Symbolic Politics? Constitutional Differentiation and Policy Reintegration in Core State Powers’ (2023) 24 European Union Politics 81. 32 Philipp Genschel and Markus Jachtenfuchs, ‘Introduction: Beyond Market Regulation. Analysing the European Integration of Core State Powers’ in Philipp Genschel and Markus Jachtenfuchs (eds), Beyond the Regulatory Polity? (Oxford University Press 2013). non-consensual eu law 135 was only in the aftermath of Maastricht and with ensuing rounds of accessions, that heterogeneity in the EU raised inevitably to a level where enhanced cooperation and inter se agreements became more frequent. Differentiation is often also the outcome even if overall welfare gains would propagate an agreement with full membership. As discussed above, failure to reach undifferentiated integration may be multifold, even if less prevalent in the EU compared to international law. The positive correlation of bargaining costs and number of participants is straightforward, as is the correlation between bargaining and length of an agreement;33 heterogeneity leads to divergent preferences that engender different payoff structures, which may be hard to reconcile through compensation in unrelated policy fields (even in the EU34); the complexity of a matter may increase bargaining costs to a level exceeding cooperation gains;35 but the opportunity costs of pursuing ‘enhanced cooperation’ (compared to full participation) are not too high as any nonparticipating EU member is able to join the cooperation at any stage (Article 331 TFEU). More generally, in order for European cooperation to be successful, the outcome of two games must overlap— within the logic of Putnam’s twolevel game involving complexity. At the national level, domestic groups pursue their interests by pressuring the government to adopt favourable policies, and politicians seek power by constructing coalitions among those groups. At the international level, national governments seek to maximize their own ability to satisfy domestic pressures, while minimizing the adverse consequences of foreign developments.36 The negotiators thus need to address the concerns of domestic interest groups and, at the same time, reach an agreement that is acceptable for the parties to the international Treaty. As every EU Member State government has to cope with a different domestic policy situation, the success of full participation of all EU members is far from obvious. To this one may add behavioural economic idiosyncrasies in domestic policy contexts. Consider that behavioural economics adopts a perspective emphasizing that ‘framing’ of the policy decision is important. This means that individuals may exist in a framework as employees, not as consumers (as suggested by economists to capture consumer rents); or as national citizens with selfish interests, not as Europeans with a common interest; or as ‘netpayers’, not viewing the positive spillovers from EU transfers for the EU as a whole. Framed accordingly, loss aversion may play an important role, given that some groups may lose from policy changes while the benefits are distributed more broadly. Loss aversion plays out forcefully 33 James D Fearon, ‘Bargaining, Enforcement, and International Cooperation’ (1998) 52 International Organization 269. 34 Joel P Trachtman, The Economic Structure of International Law (Harvard University Press 2008) 72, 187. 35 Eric Posner and Alan O Sykes, Economic Foundations of International Law (Harvard University Press 2013) 23. 36 Robert D Putnam, ‘Diplomacy and Domestic Politics: The Logic of TwoLevel Games’ (1988) 42 International Organization 427, 434. 238 armin steinBach and taxpayers in every subfederal jurisdiction and on every level always perfectly overlapped. No one could live at the expense of their neighbours or could be forced to support their neighbours. Today, the US is an example where similar incentives are offered by a federal state. Instead of sharing liability as in a union state system, the central government offers no such safety net.15 Instead, there is a great deal of uncertainty. For example, in 2010 California’s request for a bailout from the United States government was rejected. This is still true even today despite the extensive debt taken over (in fact or implicitly) by the Federal Reserve and the US government in the course of various bailouts throughout the financial crisis. While liability for state debt thus remains individualized limiting shared liability, this does not mean that other fiscal policies are in place that implement interstate solidarity with a strong economic equalization function. Interregional or vertical fiscal transfers designed to smooth macroeconomic shocks (in line with OCA theory) are to some extent a substitute for shared liability, though one that is more efficient because it engages in shock response without biasing incentives towards more debt inclination. For example, the United States has significant fiscal transfers in place to absorb asymmetrical regional shocks, with unemployment insurance as a channel for fiscal transfers. About 30 per cent of the states’ expenditures (including Medicaid) consist of federal grants. 40 per cent of local government revenues come from the state governments (state grants). The risk of a bailout is thereby significantly mitigated by the existence of significant vertical financial assistance. The eurozone has no automatic stabilizer in place that would be responsive to fluctuations in business cycles. The established EU transfer mechanisms seek to promote economic support of the economically weakest regions in Europe without however fulfilling an effective function regarding the business cycle (eg the Structural Fund and the Cohesion Fund).16 Only more recently, transfer schemes such as SURE and NGEU have not only ensured a macroeconomic function of smoothing asymmetric shocks, specifically to counter the effects of the pandemic.17 The transfer schemes have also introduced elements of union states by allowing the EU to take on debt that is ultimately secured by its Member States and which implied transfers between EU members.18 15 The US experience of enforcing a nobailout regime at the federal level visàvis the US states has been controversially discussed, with more recent developments in US fiscal federalism being an indication of this provision being relaxed. See Peter ContiBrown and David A Skeel, When States Go Broke: The Origins, Context, and Solutions for the American States in Fiscal Crisis (Cambridge University Press 2012); David Schleicher, In a Bad State: Responding to State and Local Budget Crises (Oxford University Press 2023). 16 Berend Diekmann, Christoph Menzel, and Tobias Thomae, ‘Konvergenzen und Divergenzen im “Währungsraum USA” im Vergleich zur Eurozone’ (2012) 92 Wirtschaftsdienst 27; Cristina Fasone, ‘EU Budget and Spending Powers’ in Alicia Hinarejos and Robert Schütze (eds), EU Fiscal Federalism (Oxford University Press 2023) 256. 17 Paul Dermine and Ana Bobic, ‘Of Winners and Losers: A Commentary of the Bundesverfassungsgericht ORD Judgment of 6 December 2022’ (2024) 20 European Constitutional Law Review 163, 175. 18 Clemens Fuest, ‘The NGEU Economic Recovery Fund’ (2021) 22 CESifo Forum 3. economic anD monetary union 239 By contrast, in the US, a federal state, each state pays debt costs in line with its individual level of debt. To guard against debt costs rising too high, many states in the US have on their own initiative amended their constitutions to include debt limits of various kinds. The levels of government debt among the states in the US illustrate the results. While balanced budget rules have not prevented states running high deficits on a temporary basis, the overall debt of states has been contained. State debt in 2009 was as low as 7.3 per cent (California) and 4.9 per cent (Arizona).19 That stands in stark contrast to Germany, a union state, where the level of debt ranged from about 7 per cent in Bavaria to almost 70 per cent of GDP in Berlin.20 State debt in union states By contrast, the eurozone, while legally conceptualized as a federal entity, as Staatenverbund, which stipulates economic market accountability of each of its Member States, actually exhibits the features of a union state.21 By definition, Member States of unions can take on their own debts, and those debts remain first and foremost their own liability. But as a guarantee, they can rely either on the centralized government or on other states in the union state that can be held liable for each other’s debts. Whether this is channeled through formal joint and several liability or only by implicit liability of the state does not make a difference from an economic perspective (what matters are markets’ beliefs). Whatever the governments do, there is some certainty that they can rely on other union members of a union state to step in and bail them out. Contemporary Germany is an example of a union state. While formally no joint liability for debt means exists, in practice the Bundesländer can count on the federal government to step in if fiscal hardship occurs. While this bailout is subsidiary only and tied to strict conditions, the mutualizing effect is reflected in the debt costs paid by the Bundesländer in Germany’s union system, as they remain largely independent of their individual debt levels. Their ratings ultimately depend on the rating of the federal Bund. In a union state, the liability regime is protective of lenders. While shared liability in union states incentivizes debtors to take on excessive debt, the lender is well protected because their repayment claims can ultimately be satisfied 19 Randall Henning and Martin Kessler, Fiscal Federalism: US History for Architects of Europe’s Fiscal Union (Bruegel 2012) 19. 20 Kai A Konrad and others, ‘Wege aus der Europäischen Staatsschuldenkrise’ (2010) 90 Wirtschaftsdienst 783, 802. 21 Armin Steinbach, ‘Markets as an Accountability Mechanism in EU Economic Governance’ in Mark Dawson (ed), Substantive Accountability in Europe’s New Economic Governance (Cambridge University Press 2023). 240 armin steinBach by several (indirect and direct) debtors, not only the individual direct debtor. Through implicit joint liability the creditor’s claims will be satisfied through a cascade of claims, starting with the primary debtor country but eventually extending to other members of the union or through institutionalized agencies channeling mutualization. The common debt assumed by the EU in the context of NGEU is illustrative: while the EU is the primary and direct debtor entering into contracts with creditors on capital markets, the repayment is backed by the EU budget (financed by Member States’ contributions), and if one Member State fails to make its contribution, additional contributions can be drawn from other Member States (on a pro rata basis).22 This contrasts with a federal state where each subfederal government takes on debts according to its individual debt demand curve, because lenders set their lending conditions based on each subfederal government’s likelihood to honour debts. In a union state, a pooling of the demand curves leads to a greater risk of default that may occur in subfederal governments.23 This is because in a union state the regional governments are expected to share their risk, making the total demand for debt actually larger than the sum of the subfederal governments’ demand for debt.24 The incentive for debt accumulation is increased because the financing costs are lower, at least for many states within the union state. This is different for fiscally sound countries who face a deterioration of the financing conditions if markets perceive them as potentially bailing out other members of the union, because costs are determined by the solvency risks of all state entities who could possibly be held liable, not by the individual government debt level. This also makes fiscal consolidation efforts more unlikely, as due to the mutualization of solvency, each regional government’s fiscal contribution to consolidation has only a limited effect on improving financing costs because the positive effects of consolidation are socialized across all states. Freeriding occurs— it may be rational from a member’s perspective to wait for other countries to contribute to financial stability in the union, thus freeride on their commitments by employing the additional debt space offered through other members’ fiscal efforts. In sum, shared debt liability increases moral hazard.25 What has been described corresponds to the situation in Germany (particularly before it introduced subfederal constitutional debt brakes aiming at preventing excess debt taking). Regional governments in Germany have no risk in the case of insolvency and can count on a bailout from the central government. In 1992, the German Federal Constitutional Court declared that the central government and 22 Sebastian Grund and Armin Steinbach, ‘European Union Debt Financing: Leeway and Barriers from a Legal Perspective’ (2023) Working Paper 15/ 2023, Bruegel. 23 William Oates, ‘On the Evolution of Fiscal Federalism: Theory and Institutions’ (2008) 62 National Tax Journal 313, 324. 24 Blankart and Fasten (n 14) 45. 25 ibid. economic anD monetary union 241 the regional governments in Germany are bound by a kind of community solidarity and are obligated to support one another.26 As a result, the Court decided that the Bund and the Bundesländer must help and offer financial support when any member of the German state community was facing an extreme budget crisis. At that time, the judgment was criticized for incentivizing regional governments to deliberately precipitate a budget crisis in order to force other members to pay for their fiscal policy misconduct.27 The Court’s decision resulted in the regional state governments of the Bundesländer Bremen and Saarland receiving extensive aid from 1994 to 2004. In 2004, Bremen’s debt level was still higher than that of any other regional state government. In Saarland as well, budget cuts did not have a longlasting positive effect. But the Federal Court’s decision of 1992 had clearly established that the central government and regional governments were jointly liable for each other’s debts. The rating agencies’ scoring of the local state governments shows that shared liability decisively influences the assessments of market participants. The Federal Constitutional Court confirmed the liability of the federal state as ultima ratio in its ‘Berlin decision’ in 2006. However, the Court sought to strengthen the regional state governments’ responsibility by requiring a high level of budgetary emergency and imposing budgetary consolidation measures on the regional government as a precondition for budgetary aid. Notwithstanding the fiscal conditionality as a precondition for the solidarity principle to apply, the rating agencies maintained for both the central government and the regional state governments their highest rating indicating that Bundesländer ultimately enjoy a bailoutoption. Since insolvency becomes less probable when bailouts are plausible options, even states with extremely high levels of debt (or nonsovereigns like the EU) can retain access to the credit market over an extended period. Under this condition, marketbased refinancing interest rates do not accurately reflect a state’s solvency, nor do markets function as a means of enforcing fiscal discipline.28 Incentives for debt consolidation in federal and union states Organizing states as federal or as union states within the characteristics discussed above raises questions about the tools available for dealing with the public good dilemma that the EMU is facing— that is, to incentivize members of the common currency to engage in minimizing negative spillovers from fiscal conduct and to 26 For the following, see Kai A Konrad and Holger Zschäpitz, Schulden Ohne Sühne? Warum der Absturz der Staatsfinanzen uns alle Trifft (CH Beck 2010) 186. 27 Konrad and Zschäpitz (n 26) 187; Alexander Schulz and Guntram B Wolff, ‘The German SubNational Government Bond Market: Structure, Determinants of Yield Spreads and Berlin’s Forgone BailOut’ (2009) 229 Jahrbücher für Nationalökonomie und Statistik 61. 28 Konrad and others (n 26) 802; For a different view see Blankart and Fasten (n 14) 51. 242 armin steinBach contribute to financial stability as a public good.29 A core rationale for fiscal policy coordination is thus to mitigate externalities on public goods or ‘clubgoods’ (like the stability of the euro currency). Since the ‘pooling’ of demand curves and default probabilities are a typical feature of federal states (where no credible nobailout exist), all members of the currency area bear the costs of high state debt and raising interest rates— the effects of fiscal policy on the public good are socialized. We focus on debt consolidation, as this has been in the past the major source of financial instability, while it is admitted that a sole focus on debt runs the risk of disregarding the necessary balancing with growth impulses and thus leeway for debtbased fiscal impulses. With this incentiveoriented focus on managing fiscal incentives, there are multiple ways of managing debt such as ceilings on deficit and debt, or methodologies computing debt sustainability— these instruments rely on a predefined set of thresholds or criteria, which are assessed and enforced by an (independent) agency. Rather than relying on the market’s actual response to deficit and debt levels, this rulesbased enforcement structure is implemented by a nonmarketbased enforcement of rulemakers’ (discretionary) decisions. Statutory debt ceilings may try to reflect economically sustainable debt levels, but they are definitely different from relying on the market’s verdict.30 A different approach is to either credibly exclude, by law, or at least create uncertainty that governments will be bailed out, with the aim of inducing markets to base their solvency judgements solely on an individual country’s performance, hence using marketdetermined interest rates as disciplining tools.31 Thus, the first approach is a discretionbased, nonmarket enforcement of fiscal policy and the latter relies on marketbased enforcement, while both rely on the existence of a credible legal framework by determining the assessment and enforcement structure. In the EU, both approaches have been laid down in the EU Treaties: a deficit and debt ceiling approach with a rulesbased mechanism of enforcement (Article 126 TFEU), more recently complemented by a regime determined by debt sustainability analysis, as well as a marketoriented nobailout clause (Article 125 TFEU). The first approach limits regional governments’ ability to take on debt put in place through statutory regulation and administrative enforcement (by the EU Commission). This approach does not necessarily alter the debtor’s preference to take on debt (rather than increasing state revenue through taxation), but it limits the debt space through the decisions of an externally enforced limitation. In contrast, the second approach is aimed at reducing the debtor governments’ preference 29 Isabelle Joumard and Mathis Kongsrud, ‘Fiscal Relations across Government Level’ (2003) OECD Economics Department Working Papers No 375, 29. 30 Daniel Kelemen and Terence Teo, ‘Law, Focal Points, and Fiscal Discipline in the United States and the European Union’ (2014) 108 American Political Science Review 355, 363; Geoffrey Woglom, ‘Do Credit Markets Discipline Sovereign Borrowers? Evidence from US States’ (1995) Journal of Money and Credit, and Banking 1046. 31 Mark Hallerberg, ‘Fiscal Federalism Reforms in the European Union and the Greek Crisis’ (2010) 12 European Union Politics 127, 130. economic anD monetary union 243 for debt by leveraging market prices (interest rates) to influence the lender’s willingness to loan.32 The latter is a marketbased strategy that uses the risk of insolvency to link the cost of credit as closely as possible to the risk of default. Consequentially, the effectiveness of marketbased nobailout provisions critically depends on functionality of markets and the extent to which marketbased interest rates are determined by fundamentals of the country concerned rather than being influenced by irrational market sentiments. The history of the EMU has shown that reliance on a statutory nobailout clause in Article 125 TFEU is an insufficient tool to constrain debt incentives: over a decade since the beginning of the euro, markets did not view any divergence between Member States that was of relevance to members’ solvency, as bond spreads were largely convergent. This market ignorance engendered underestimating different degrees of solvencies between EU members, and with the outbreak of the crisis, the bond spreads overreacted in a way that was not justified by the fundamental economic data (thus drastically overestimating divergences in solvency).33 In turn, the effectiveness of enforcementbased approaches under Articles 121 and 126 TFEU depends on how existing rules deal with political economy and strategic behaviour, in particular how enforcement can be implemented effectively. Since its inception, the Treatybased enforcement structure, implemented through the Stability and Growth Pact, has been subject to criticism due to its lack of effectiveness. Fiscal policy arrangements Market-based instruments to lower deficit inclination Nobailout rule Insolvency rules SGP Debt brakes Union state (shared liability is possible) Federal state (no shared liability) Non-market based but fixed deficit/debt limitation Fig 23.1 Fiscal policy arrangements Source: Armin Steinbach, Economic Policy Coordination in the Euro Area (Taylor & Francis Ltd 2014) 83. 32 Timothy D. Lane, ‘Market Discipline’ (1993) IMF Staff Papers No 53, 55; Fabrizio Balassone, Daniele Franco and Raffaela Giordano, ‘MarketInduced Fiscal Discipline: Is there a FallBack Solution for Rule Failure?’, Public Debt (Banca d’Italia 2004), 410. 33 Paul de Grauwe, Yuemei Ji, and Armin Steinbach, ‘The EU Debt Crisis: Testing and Revisiting Conventional Legal Doctrine’ (2017) 51 International Review of Law and Economics 29. 244 armin steinBach Figure 23.1 shows how different state organization principles may have different methods of restricting sovereign debt accumulation. State organization principles can be distinguished by reference to shared liability as well as by reference to their market orientation. The suitability of each approach depends on the form of state organization that is legally implemented. A credible nobailout provision and state insolvency procedure are better suited in a federal state setting building on a marketbased system to contain debt. In federal unions, debt is individualized and shared liability is precluded, and the markets assess only the solvency of an individual state— a logic that best suits a bailoutclause and state insolvency framework, that is, procedure to restructure debt.34 The American experience with Chapter 9 of the American Bankruptcy Code shows that a formalized bankruptcy procedure can provide an incentive for fiscal consolidation. Certainly, marketbased approaches could also be implemented in a union state, where shared liability determines the regional government’s ability to take on debt, and one may argue that the EU is precisely such a case, as the Treatybased bailout clause stands next to institutions such as the ESM, SURE, and NGEU, all of which have established some kind of shared liability. But marketbased approaches are less effective in union states because shared liability undermines the market signal, not allowing market interest rates to be credible reflections of an individual state’s solvency risk. Again, the EU is illustrative in this regard. During the first ten years of the eurozone, the strong convergence of sovereign bond interest rates suggested that markets assumed joint liability in the EU, as spreads were largely unconnected from underlying economic data. The market’s belief in some kind of shared liability— despite the nobailout clause— may have been at work. In turn, in a union state, in which market forces on an individual country’s budgetary conduct are comparatively weak by design, there is a greater need for alternative debt containment in order to mitigate external effects. Without the disciplining force of markets that sanctions a state’s solvency, there is a need for additional instruments containing negative external and destabilizing effects that may result from fiscal conduct— the EU has implemented the logic and established fiscal governance from the beginning of the euro. Revenue generation through EU bonds that are backed by Member States’ contributions and ultimately secured by other members’ solvency (as implemented in SURE and NGEU) is alien to the system in a pure federal organization due to the associated mutualization effects— while bonds allowing shared liability may be implemented for sound macroeconomic reasons to stabilize the currency area,35 this instrument 34 Jonathan P Thomas, ‘Bankruptcy Proceedings for Sovereign State Insolvency and Their Effect on Capital Flows’ (2004) 13 International Review of Economics & Finance 341. 35 Gabriele Giudice and others, ‘A European Safe Asset to Complement National Government Bonds’ (2019) MPRA Paper No 95748. economic anD monetary union 245 weakens the marketbased principle established through the nobailout clause. Joint liability bonds engender a ‘pooling’ of demand curves.36 In turn, for a union state, shared liability bonds are the suitable tool to stabilize the character as a union state, because shared liability expenditure is an appropriate way of funding projects that address common concerns such as transnational public goods. The financing of transnational public goods (or European public goods, see above Chapter 21 d) is not obvious in a federal state with individual liability. Strictly speaking, members must contribute to European public goods according to the benefits they derive from them (fiscal equivalence). Common debt may facilitate financing public goods. b) Debt limitation in the eurozone The EU is not easily characterized as either a union or federal state. The EU is not a pure federal state, nor a union state. From a legal perspective, the EU has been characterized by the German Constitutional Court as an association of sovereign states (Staatenverbund), which highlights that the EU is more than just a loose group of states tied together through an international law Treaty, but is not a sovereign state in its own right. Formally, the EU has established a nobailout principle (Article 125 TFEU) as well as a ban on monetary financing (Article 123 TFEU), both of which incorporate the idea of fiscal responsibility for each individual Member State of the EU— there is agreement in legal scholarship, both among scholars as well as in jurisprudence (of the ECJ and the German Constitutional Court), that the rationale of these core provisions of the EMU is to let market pressure be the disciplining force on Member States, thus emphasizing the EU’s character as a federal state in which all union members remain responsible for their own fiscal fate.37 Experience has shown that the anticrisis measures have put in question the binding force of these fiscalmonetary pillars of selfdependence. Financial turmoil threatening the financial stability of the eurozone has been seen to lead to detrimental effects (for all members of the eurozone) and injecting considerable flexibility to the rules as the only way to reconcile comprehensive financial assistance with statutory rules. Both a fiscal nobailout provision as well as a monetary ban on state financing were turned away from a strict interpretation: a move that was motivated to divert negative effects that could result from state insolvency within the euro currency area. Hence, while a nobailout clause builds the EU on the premise of a federal state, the statutory ban has incrementally softened towards a union state implementing elements of shared liability character. This leads 36 Brady Gordon, The Constitutional Boundaries of European Fiscal Federalism (Cambridge University Press 2022) 350. 37 ibid 122– 130, 345. 246 armin steinBach markets to work only partially in constraining fiscal policy. The institutionalization of the crisis instruments in the EU Treaty, such as the legal basis for the ESM (Article 136 TFEU), have further established a permanent mechanism to provide financial aid when financial stability requires it.38 Despite weakening market pressure through these crisis instruments, the EU sought to offset the weakening of the bailoutclause by imposing fiscal conditionality as a quid pro quo for financial assistance. The consequence was a change from market accountability towards political accountability. As long as Article 125 TFEU was fully intact, Member States were held by markets accountable for their policies that affected their solvency, with markets sanctioning members for unsound fiscal policy. The term ‘economic accountability’ captures the EU Treaties’ choice of free market rules that subject the financing needs of private actors and states to the judgement of markets.39 Economic accountability has been supplanted by subjecting receiving Member States to the judgement of other sovereigns’ demands, with accountability to markets being incrementally replaced by accountability to other sovereigns or EU institutions. Creditor Member States, the EU Commission, and the European Central Bank act as providers of financial aid in return for conditional structural, economic, financial, and social reforms to be carried out by the debtor states. What has previously been pressured by markets is now (under the ESM) subject to political discretion of nonmajoritarian institutions as well as from other sovereigns. It is against this background that a vast body of literature in political science emerged that focused on the problematic aspects of this shift from a legitimacy and acceptance perspective.40 From an economic perspective, one can be neutral regarding who should be responsible for containing external effects— if the goal is to prevent Members to act in a fiscally imprudent fashion (Article 125 TFEU), a conditionality programme set up by nonmajoritarian institutions is no worse than market pressure, provided that it is equally effective. A precondition for conditionality to be effective is that it is carefully calibrated so that it does not create bailout expectations.41 Taken together, there is no instrument in the EU that sufficiently minimizes the propensity to borrow. Since the eurozone was founded, sovereign bond interest rates have factored in the mutual commitment among eurozone countries, encouraging borrowing by offering relatively low rates. It was only during the sovereign debt crisis that rates reacted with hypersensitively, moving beyond the underlying 38 Fabian Amtenbrink and Menelaos Markakis, ‘Never Waste a Good Crisis on the Emergent EU Fiscal Capacity’ in Alicia Hinarejos and Robert Schütze (eds), EU Fiscal Federalism (Oxford University Press 2023) 197. 39 Steinbach (n 21). 40 Ben Crum and Deirdre Curtin, ‘The Challenge of Making European Union Executive Power Accountable’ in Simona Piattoni (ed), The European Union (Oxford University Press 2015) 63; Deirdre Curtin, ‘Challenging Executive Dominance in European Democracy’ (2014) 77 Modern Law Review 1. 41 Gordon (n 36) 354. economic anD monetary union 247 economic fundamentals.42 In addition, there is also no insolvency regime that would restructure debt and require a plan for resolvency. If the EU wants to strengthen its federal element by reducing its Members’ propensity to incur debt, a credible nobailout and the possibility of insolvency for states would be suitable forms of coordination that aim at strengthening preventive market incentives. Insolvency proceedings create legal certainty for debtors and creditors. Having an insolvency procedure with restructuring requirements in place would further induce creditors to assess debtors’ individual creditworthiness rather than counting on shared liability of union members. The introduction of collective action clauses (CAC) has been a major development in this regard. The main purpose of the euro area model CAC is to introduce a ‘singlelimb’ voting mechanism. Restructuring decisions are thus binding on all holders of a series of debt securities aggregated in one voting group if the proposed modification is approved by holders of a majority of all securities aggregated in the voting group. Facilitating debt restructuring not only speeds up insolvency proceedings but also has a chilling effect on lending practices. EU law has never relied exclusively on the bailout prohibition. In the absence of a credible ban, EU law sets out a comprehensive network of procedural and substantive fiscal obligations, led by the overarching goal to have ‘sound public finances and monetary framework conditions’ (Article 119 TFEU). This objective guides the fiscal policy regulations specified in Article 126 TFEU, the Stability and Growth Pact (SGP) specifying the implementation of numerical deficit and debt criteria, which was supplemented by the Fiscal Compact, an international Treaty requiring EU members to introduce national debt brakes.43 National debt brakes fulfil a similar function to the SGP: they do not use market incentives for budget discipline, but rather impose (more or less) arbitrary debt and deficit limits set and administered by authorities. The SGP and the Member State debt brakes are therefore more suitable for a union state in which market forces are conceptually abandoned on strict terms. The same applies to shared liability bonds as implemented under the ESM, SURE, and NGEU. They amount, at least in part, to joint liability, even if— depending on their design— they mimic market pressure through conditionality. On an economic account, it is questionable to what extent EU fiscal rules with their application of the rigid ‘onesizefitsall’ approach using numerical benchmarks will promote the sustainability of its members’ debt. Even if the distinction between ‘good debt’ and ‘bad debt’ is an unplausible market definition, there are instances when debt can be used for growthenhancing purposes (thus stabilizing fiscal sustainability), and EU rules have shown certain flexibility for expenditures 42 Paul de Grauwe and Yuemei Ji, ‘SelfFulfilling Crises in the Eurozone: An Empirical Test’ (2012) 34 Journal of International Money and Finance 15. 43 Steinbach (n 21) 84. 254 BiBliograPhy Cecchini P, The European Challenge 1992: The Benefits of a Single Market: European Challenge— Benefits of a Single Community (Gower Publishing Ltd 1998) Chalmers D, ‘The European Redistributive State and a European Law of Struggle’ (2012) 18 European Law Journal 667 Chayes A and Handler Chayes A, ‘On Compliance’ (1993) 47 International Organization 175 Checkel JT, ‘Why Comply? Social Learning and European Identity Change’ (2001) 55 International Organization 553 Chilton A and Linos K, ‘Preferences and Compliance with International Law’ (2021) 22 Theoretical Inquiries in Law 247 Chinkin CM, ‘The Challenge of Soft Law: Development and Change in International Law’ (1989) 38 International and Comparative Law Quarterly 850 Chiti E and Teixeira PG, ‘The Constitutional Implications of the European Responses to the Financial and Public Debt Crisis’ (2013) 50 Common Market Law Review 683 Claeys G and Sapir A, ‘The European Globalisation Adjustment Fund: Easing the Pain from Trade?’ in L Paganetto (ed), Capitalism, Global Change and Sustainable Development (Springer 2020) 97 Claeys G and Steinbach A, ‘A Conceptual Framework for the Identification and Governance of European Public Goods’ (2024) Working Paper 14/ 2024, Bruegel Coase RH, ‘The Problem of Social Cost’ (1960) 3 The Journal of Law and Economics 1 — — , ‘The Institutional Structure of Production’ (1992) University of Chicago Law Occasional Paper Colliard JE, ‘Optimal Supervisory Architecture and Financial Integration in a Banking Union’ (2019) 24 Review of Finance 129 ContiBrown P and Skeel DA, When States Go Broke: The Origins, Context, and Solutions for the American States in Fiscal Crisis (Cambridge University Press 2012) Conway G, ‘Recovering a Separation of Powers in the European Union’ (2011) 17 European Law Journal 304 Cooper I, ‘National Parliaments in the Democratic Politics of the EU: The Subsidiarity Early Warning Mechanism, 2009– 2017’ (2019) 17 Comparative European Politics 919 Cooter R and Gilbert M, Public Law and Economics (Oxford University Press 2022) Coutts S, ‘Supranational public wrongs: The limitations and possibilities of European criminal law and a European community’, 54 Common Market Law Review (2017), 771– 800 Cox GW and McCubbins MD, ‘The Institutional Determinants of Economic Policy Outcomes’ in Stephan Haggard and Mathew D Mccubbins (eds), Presidents, Parliaments, and Policy (Cambridge University Press 2001) Cozzi AO, ‘The Italian Constitutional Court, the Plurality of Legal Orders and Supranational Fundamental Rights: A Discussion in Terms of Interlegality’ (2022) 1 European Law Open 606 Craig P, ‘Pringle: Legal Reasoning, Text, Purpose and Teleology’ (2013) 20 Maastricht Journal of European and Comparative Law 3 — — , ‘Institutions, Power, and Institutional Balance’ in Paul Craig and Gráinne de Búrca (eds), The Evolution of EU Law (Oxford University Press 2021) 46 Craig PP and Markakis M, ‘Gauweiler and the Legality of Outright Monetary Transactions’ (2016) 41 European Law Review 4 Croley SP, ‘Theories of Regulation: Incorporating the Administrative Process’ (1998) 98 Columbia Law Review 1 Crum B and Curtin D, ‘The Challenge of Making European Union Executive Power Accountable’ in Simona Piattoni (ed), The European Union (Oxford University Press 2015) 63 — — , ‘The Challenge of Making European Union Executive Power Accountable’ in Simona Piattoni (ed), The European Union (Oxford University Press 2015) Crum B and Merlo S, ‘Democratic Legitimacy in the PostCrisis EMU’ (2020) 42 Journal of European Integration 399 Curtin D, ‘Challenging Executive Dominance in European Democracy’ (2014) 77 The Modern Law Review 1 Cutler C, ‘Critical Reflections on the Westphalian Assumptions of International Law and Organization: A Crisis of Legitimacy’ (2001) 27 Review of International Studies 133 [Document text truncated for crawler view.]