Institutional Change in the Banking Union: The Case of the Single Supervisory Mechanism


 This article is about institutional change in the Banking Union. It has two related aims. The first is to engage with the law of the Single Supervisory Mechanism (SSM)—the first pillar of the Banking Union—and in this context to discuss tensions that have lately emerged between the case law of the Court of Justice of the European Union (CJEU) and that of the German Federal Constitutional Court. The second, but main, aim of this article is to put the law of the SSM as it was enacted in the SSM Regulation, and as it was interpreted by the CJEU and by the German court, in a broader perspective of institutional change. For this purpose, this article adopts an interdisciplinary approach that seeks insights on institutional change in the political science literature. In particular, the article seeks to shed light on the role played by courts. In short, it argues that whilst the SSM is a story of change following an exogenous shock (ie the sovereign debt crisis), it is also an account of change and contestation between courts made possible by the ambiguities and incompleteness of the SSM rules. It will show that the evolution of the SSM is by no means frictionless and that it is only by tracing change from the point of the enactment of the law to its interpretation by the courts that one gains a real appreciation of the dynamics and salience of change within the SSM.


I. Introduction
This article is about institutional change in the Banking Union. It pursues two aims. The first is to engage with the law of the Single Supervisory Mechanism (SSM)-the first pillar of the Banking Union. The SSM is about bank supervision and was established following the Eurozone crisis (or sovereign debt crisis). It has the European Central Bank (ECB) at its heart. The latter *Dr Pierre Schammo, Durham University, School of Law, Palatine Centre, Durham, DH1 3LE, UK. Email: pierre.schammo@durham.ac.uk. I would like to thank Birke Häcker, John Campbell, Lukas Simkus, Niamh Moloney, Kathleen Thelen, Johannes Ungerer and Steve Weatherill for offering encouragements, thoughts and comments. This article was written during my stay at the Institute of European and Comparative Law (IECL), Oxford University. was vested with bank supervisory tasks under a Council Regulation (the SSM Regulation), 1 which it carries out together with supervisory authorities of Member States that participate in the Banking Union (hereinafter, national competent authorities or NCAs). This article discusses the law of the SSM. In this context, it also examines tensions that have lately emerged between the case law of the Court of Justice of the European Union (CJEU) and that of the German Federal Constitutional Court (Bundesverfassungsgericht or BVerfG) following the latter's judgment on the Banking Union. 2 In short, it will show that the CJEU and the BVerfG do not agree on a basic question of competence in the SSM, that is, essentially whether NCAs, when carrying out supervisory tasks in relation to smaller credit institutions (so-called 'less significant' credit institutions) under the SSM Regulation, do so on the basis of their pre-existing national competence.
The second, but main, aim of this article is to put the law of the SSM as it was enacted in the SSM Regulation, and as it was interpreted by the CJEU and the BVerfG, in a broader perspective of institutional change. Hence, my interest in institutional change frames the approach that this article adopts. Specifically, the article grapples with several themes that have shaped debates on institutional change in the literature: how institutional arrangements change over time; what common mechanisms there are through which change occurs; what common factors act as triggers for change; and which actors participate in shaping change. To contribute to debates on institutional change, this article takes a cross-disciplinary approach and seeks insights on institutional change in the political science literature. The Eurozone crisis and the changes which ensued, have offered legal scholarship and scholarship in political science valuable material. Yet, despite being a theme of common interest, cross-fertilization between scholarship has been limited. This is unfortunate since disciplines such as law and political science have much to learn from each other. Indeed, a cross-disciplinary approach is arguably especially promising when what is at issue is the evolution of formal 'institutions' (that is for our purposes, written law). 3 In trying to understand and conceptualize the related dynamics, political science offers concepts and theories html. An English translation is available here: https://www.bundesverfas sungsgericht.de/SharedDocs/Entscheidungen/EN/2019/07/rs20190730_2bvr168514en.html. Note, however, that some key paragraphs (especially, paras [188][189][190][191][192][193][194][195][196] were not translated into English. 3 It is important to note that the meaning of 'institutions' in the literature on institutional change contrasts markedly with the way in which legal scholarship typically understands the term (that is, as organizations or as EU institutions, for example). The understanding of institutions in the institutional literature that resonates best with legal scholarship is that of written rules. However, note that scholars working in traditions such as sociological institutionalism can adopt a much wider understanding of institutions.
to account for change. Legal scholarship meanwhile offers the necessary understanding of the law, as well as its ambiguities and incompleteness.
To be sure, even if cross-fertilization between disciplines has been limited, the literature on the Eurozone crisis is voluminous. In political science, contributions have looked at the crisis from various theoretical angles: for example liberal intergovernmentalism, neo-functionalism, or new institutionalism, such as historical institutionalism. 4 Much has also been written in the legal literature, with authors seeking to draw lessons from a judicial and/or constitutional/administrative law perspective. 5 A theme of interest in both disciplines has been how best to account for institutional changes that have taken place in response to the Eurozone crisis. In the political science literature, this work fits with a long held interest in seeking to explain and account for institutional change. In the legal literature too, the Eurozone crisis has prompted assessments of patterns of change (eg in terms of a 'constitutional mutation' 6 ). Whilst naturally the emphasis is more on the substance of the law, assessments have led to contrasting conclusions with respect to the salience of these changes. 7 That said, the above literature has comparatively less to say on two important aspects. The first concerns a key response to the crisis: that is, the Banking Union and, given our area of interest, the Single Supervisory Mechanism. Here sector specific literature has more to say. 8 In law, this literature typically engages with the complexities and intricacies of the law itself. 9 However, the level of theorizing when accounting for change does vary. The second aspect concerns the evolution of the arrangements that were put in place in order to deal with the Eurozone crisis. A voluminous literature has focused on the EU's more immediate response to the crisis. Comparatively less attention has been given to the continuous evolution of these arrangements and what it tells us about institutional change. In this context, the role played by courts has also attracted less attention. 10 In particular, the judgment of the BVerfG on the Banking Union and its importance from an institutional change perspective remain under-appreciated. 11 By taking an inter-disciplinary approach and engaging with the law of the SSM and its interpretation by the courts, this article works towards addressing this gap. My basic contention is that whilst the SSM is a story of change following an exogenous shock (ie the sovereign debt crisis), shedding light on judicial decision-making demonstrates that the SSM is also an account of change and contestation between courts, made possible by the ambiguities and incompleteness of the SSM rules. Hence, I will show that the evolution of the SSM is by no means frictionless and that it is only by tracing change from the point of the enactment of the law to its interpretation by the courts that one gains a real appreciation of the dynamics and salience of change within the SSM.
This article proceeds as follows. Section II begins by examining the literature on institutional change and identifies a number of takeaways that will inform the subsequent discussion. Section III turns to the SSM Regulation in order to gain first insights on institutional change within the SSM. This section also seeks to prepare the ground for examining, in Section IV, the role of courts in shaping change in the SSM. Specifically, Section IV focuses on two key judicial decisions-the CJEU's decision(s) in Landeskreditbank Baden-Württemberg v ECB and the judgment of the BVerfG on the Banking Union. 12 It examines the contradictions between the findings of the CJEU and the BVerfG and demonstrates how each of these decisions contributes to the institutional change problematic. Section V summarizes the main lessons learned and reflects on the prospect of reconciliation between courts. Section VI concludes.

II. Institutional change: analytical background
As noted, this article seeks to assess institutional change in the SSM. This section takes the first steps towards this aim. It begins by introducing some of the literature on institutional change (Section II.A), and identifies the main takeaways for the subsequent analysis of the SSM (Section II.B).

A. The literature on institutional change
The literature on institutional change seeks to study and conceptualize change in relation to 'institutions'-a vague term that, depending on the theoretical lens, can be understood in diverse, more or less broad, ways. The understanding of institutions that resonates best with legal scholarship is that of formal rules. 13 Accordingly, this is the understanding of 'institutions' that I will adopt in this article. This sub-section reviews, for insight and inspiration, some of the literature on institutional change. There are at least three basic insights that can be gained from this literature. The first of these insights is about the dynamics and salience of change. To put it simply, institutional change can be more or less dramatic. It can come about suddenly or happen slowly over time. It can be either exogenously driven (eg as a result of an economic crisis) 14 or endogenously driven, that is 12 BVerfG (n 2) above; Case T-122/15, Landeskreditbank Baden-Württemberg-Förderbank v European Central Bank ECLI:EU:T:2017:337; Case C-450/17 P, Landeskreditbank Baden-Württemberg-Förderbank v European Central Bank ECLI:EU:C:2019:372. 13 As noted earlier (see (n 3)), the meaning of 'institutions' in the literature on institutional change contrasts markedly with the way in which legal scholarship typically understands the term (that is, as organizations or as EU institutions, for example). The understanding of 'institutions' that is adopted here is closest to rational choice or historical institutionalist approaches. On the different meaning of institutions, see J Campbell, Institutional Change and Globalization (Princeton, NJ: Princeton University Press, 2004) 3-4. 14 See eg ibid 174 (describing exogenous factors as including 'war, economic catastrophe, and other calamities as well as abrupt shifts in prices and transaction costs, changes in state policy, dramatic technological innovations, and the like'). Of course, the transition from problem (whether endogenous or exogeneous) to solution cannot be taken for granted. Campbell, for example, highlights a number of issues that might stand in the way of institutional change, such as actors failing to identify a problem; actors disagreeing on whether or how to address it; or actors deciding to address a problem without that this involves changing institutions (ibid 175).
'from within the workings of the institutions, rather than outside'. 15 Two models proved initially popular in order to conceptualize change: an evolutionary model where change 'proceeds in small, incremental steps along a single path in a certain direction', 16 and a punctuated equilibrium model where long periods of path-dependent stability are punctuated by brief periods of radical change during socalled 'critical junctures'. 17 These models proved popular in legal scholarship too. Grundfest, for example, described the development of US securities regulation as consistent with the logic of punctuated equilibria. 18 However, the usefulness of thinking about change in such strictly dichotomous terms has been questioned in the more recent past. Streeck and Thelen, for example, argue against being too deterministic. These authors seek to disentangle processes of change, which can be incremental or abrupt, from the outcomes of change, which can represent continuity or discontinuity. 19 In the process, they argue against a 'conceptual schema' that only allows for 'either incremental change supporting institutional continuity through reproductive adaptation, or disruptive change causing institutional breakdown and innovation and thereby resulting in discontinuity'. 20 Campbell too, is critical of a rigid conceptualization. For Campbell, institutional change is not a 'dichotomous variable'. Instead, it should be situated on a continuum: 'it is a matter of degree and needs to be examined as such '. 21 Besides this point about the nature and dynamics of change, Campbell's work offers two additional insights, which are more methodological, but which are critical for the discussion that follows. 22 The first is about the importance of precisely identifying what one seeks to track when accounting for change; the second is about the timeframe over which one seeks to track change. 23 Starting with the former, Campbell notes that to accurately account for change, one needs first to identify the dimensions of an 'institution' that matter. Typically, this question will be influenced by the theoretical perspective that one adopts. 24 22 See also Campbell who notes with respect to these insights that they 'can yield important theoretical insights into the nature of institutional change' (ibid 33). 23 Ibid 35. 24 Ibid 40. back to the point that political scientists and sociologists working on institutional analysis see the world through different 'lenses' (eg rational choice, historical, or sociological institutionalism) and their understanding of institutions varies accordingly. 25 As a result, they might seek to track change in formal written rules, informal rules or, say, cultural/cognitive frames. Nevertheless, the dimensions that one choses to examine must be 'salient'. As Campbell puts it, the 'danger is that researchers may confuse dimensions that are relevant to themselves with those that are salient to the actors they are studying'. 26 The second point is about the importance of 'time': change that appears small or radical at the stage of institutional creation might require an altogether different characterization if one extends the timeframe for analysis to the implementation or application stage of rules. This stage, 'carries its own dynamic of potential change', 27 say for example, because of contestation before a court or because of the way in which ambiguous rules are applied-or not applied-later on. Thus, 'avoiding one-shot game explanations' 28 by tracking changes over an appropriate timeframe can yield important insights.
The importance of selecting the right timeframe is apparent in several major works. 29 It is, for example, an important consideration in Farrell's and Héritier's work on interstitial institutional change, 30 which Bruno de Witte draws upon in his assessment of constitutional change in the wake of the Eurozone crisis. 31 Specifically, Farrell's and Héritier's work highlights the importance of examining the periods in between formal Treaty changes when accounting for change. They posit that Treaty ambiguities will offer opportunities to policy actors to maximize their competences in the periods in between Treaty change. Thus, interstitial institutional change is defined as 'change in institutions which occurs between Treaty reforms'. 32 What is more, they posit the existence of a sort of feedback loop: interstitial institutional changes may '[feed] back into processes of Treaty change'. 33 'Time' is also a key consideration in Mahoney's and Thelen's work on gradual institutional change. 34 The authors observe that institutional change based on endogenous developments that take place after the point of institutional creation are often 'overlook[ed]' by the literature because they require one to 'consider a somewhat longer time frame than is characteristic in much of the 25 Ibid. 26  literature'. 35 For the authors, institutions typically have 'power-distributional implications' 36 and actors, who benefit unevenly from the establishment of institutional arrangements, will as a result have differing interests to defend or contest these arrangements subsequently. Mahoney and Thelen go on to identify several mechanisms through which institutions may evolve over time. 37 They coin them 'displacement', 'drift', 'layering', and 'conversion'. Each mechanism or 'mode of institutional change' is presented as having different features and as describing a different logic of change. 38 Displacement, for example, refers to the 'removal of existing rules and the introduction of new ones', a process which according to Mahoney and Thelen can be abrupt/radical, but also 'slow moving'. 39 This may be so when actors who lost out under existing institutional arrangements are able to introduce new institutions which enter into competition with existing ones and which come to gradually eclipse the latter. 40 Drift refers to the situation where rules are left unchanged in the face of environmental shifts and this inaction causes the impact of the rules to change. 41 Layering and conversion are legally speaking more complex modes. Layering refers to the 'introduction of new rules on top of or alongside existing ones'. 42 According to the authors, examples of layering are common in situations where existing arrangements cannot be challenged, for example because of the presence of veto players. 43 Mahoney and Thelen claim that in such a case, adding a new institutional layer (eg new rules) on top of, or alongside, existing arrangements may be the way forward and still bring about 'substantial change' if the former is able to destabilize the latter. 44 Meanwhile conversion takes place where existing rules are interpreted in new and contrasting ways, allowing 'institutions'-in our context formal rules-to be redirected to 'new goals, functions, or purposes' 45 by actors who deliberately take advantage of the 'inherent ambiguities of the institutions'. 46 To be sure, the law cannot be interpreted at will. Most obviously, a rule cannot be interpreted in a way that plainly contradicts its wording. However, when seeking to ascertain the 35 Ibid. 36 Ibid 14. See also ibid 8 referring to institutions as 'distributional instruments laden with power implications'. 37 44 Ibid 17, noting that 'layering can, however, bring substantial change if amendments alter the logic of the institution or compromise the stable reproduction of the original "core"'. 45 Ibid 17-18. 46 Ibid 17. meaning or scope of a rule, there is often significant room to operate within the boundaries set by the wording of the rules. Indeed, for Mahoney and Thelen, change by 'conversion' is possible because of the ambiguity of rules. 47 Hence, they argue that incremental change often happens at the stage of the interpretation and enforcement of rules, emerging 'in the "gaps" or "soft spots" between the rule and its interpretation or the rule and its enforcement'. 48 Such 'gaps' may accordingly become contestation 'sites' over 'the form, functions, and salience of specific institutions . . .'. 49

B. Analytical takeaways
The above literature provides useful insights. It draws our attention to the different dynamics and salience of institutional change, but also warns us against being too deterministic when seeking to conceptualize change. It encourages us to be more explicit about the institutional dimensions that are relevant to ascertain the reality of change. Importantly, it alerts us to the importance of the timeframe when tracking change and draws our attention to the fact that whilst change can be a response to a problem that has an exogeneous cause (eg an economic crisis), it can also have as origin endogenous problems that emerge after the point of institutional creation. Finally, it invites us to examine the mechanisms (conversion, layering, drift, etc) by which institutions can change over time. Building on these insights, this subsection identifies three key takeaways that will inform the subsequent discussion on institutional change in the SSM.

Change and legal ambiguity/incompleteness
The message of Mahoney and Thelen on how legal ambiguities or 'gaps' in formal rules offer opportunities to shape 'institutions' over time is likely to resonate with legal scholarship. 50 Hereinafter, I will draw on the work of Mahoney and Thelen. I will proceed on the basis that legal ambiguities and legal incompleteness can serve as an intellectual and interpretative 'battle ground' where contests over the precise meaning and function of rules can take place, and where as a result institutional change can occur over time. The point has arguably special relevance in a context such as the EU where the law can often be vague and ambiguous. That is true of the Treaties, as Farrell and Héritier point out in their analysis of interstitial institutional change. 51 But it is also true of secondary legislation. There is a multitude of reasons for this state of affairs. It might be because of bounded rationality or imperfect information, or simply because it is the price to pay for reaching agreement over a piece of legislation. To conceptualize change, the institutional modes that Thelen and others identify help to gain ground. Recall that among these modes, 'conversion' describes how actors can attempt to take advantage of the 'inherent ambiguities of the institutions' 52 in order to redirect rules, by way of interpretation, to new outcomes. Hereinafter, these modes of institutional change will prove useful when considering the role of the courts in the evolution of the SSM. That said, when considering their usefulness, it is also worth bearing in mind that they are 'ideal types' in a Weberian sense. 53 They are conceptual tools that can help to make sense of a more complex reality. But they do not exhaust the ways in which institutions can change. 54 Moreover, to explain change, it may well be necessary to combine the distinct insights offered by different modes of institutional change. Thus, for example, in cases where new rules were added on top of existing ones (as described in the layering mode), the impact that the former have on the latter may be unclear and may accordingly need to be 'discovered' through interpretation, opening up the prospect of change through 'conversion'. Indeed, one might go a step further. If ambiguity is, as Mahoney and Thelen contend, something that is a 'permanent feature' of rules, 55 one might posit that it may well be unclear whether rules were merely layered on top of, or alongside, existing ones, or whether their impact is more dramatic, such as where they effectively replace existing ones. Under these circumstances, the outcomes to which rules can be directed and redirected, as described in the conversion mode, may well depend on whether the adoption of new rules is found to be consistent either with the sort of layering of rules that is characteristic of the layering mode, or with something else, such as a replacement of rules. I will return to the point later when considering the role of the courts in our account of institutional change. For the moment, it is sufficient to add that this issue does not arise in Thelen's conceptualization of modes of institutional change. However, questions about how rules interact with each other are arguably common in a two-level system such as the EU where regulatory activity takes place at both the national level and the EU level. 56 Indeed, ambiguity with respect to, say, the meaning or scope of EU rules may often be intended in order to 52 Mahoney and Thelen (n 17) 17. 53 I am very grateful to Kathleen Thelen for pointing this out to me. Max Weber developed the concept of 'ideal types', describing it as an 'analytical accentuation . . . of certain elements of reality' (quoted in R Swedberg, The Max Weber Dictionary: Key Words and Central Concepts (Stanford, CA: Stanford University Press, 2005), 119). The basic thinking is that, when trying to explain a complex reality, it can be useful to start with 'what is essential about a phenomenon'. Ideal types are meant to serve this purpose. See ibid 120. 54 For an example where the institutional modes were found not to fit and for a proposal to add a new mode (coined 'copying') to those identified above, see A Verdun, 'A historical institutionalist explanation of the EU's responses to the euro area financial crisis' (2015) 22 Journal of European Public Policy, 219-37, 232. 55 Mahoney and Thelen (n 17), 11. 56 Admittedly, EU rules benefit from primacy, but before national rules can be set aside, it might be necessary to first ascertain whether national rules are impacted by EU rules. This in turn may require the meaning or scope of EU rules to be established by way of interpretation. find political agreement at EU level. In such a case, the precise meaning or scope of EU rules may well ultimately have to be ascertained before the judiciary.

Contestation between courts
Hereinafter, I will examine how ambiguous and incomplete provisions of the Treaties and the SSM Regulation opened an interpretative space for the CJEU and the BVerfG to reach markedly contrasting findings on the SSM, and how as a result these courts were able to either amplify or contest institutional change within the SSM. Hence, in this article, I will conceive of courts as actors of change, which whilst relying on legal discourse, can shape outcomes by resolving ambiguities and incompleteness in formal written rules.
To be sure, contestation over the meaning and purpose of rules should find an easy resolution in an EU context. Under the preliminary reference procedure, it is the CJEU which has the final word on the interpretation and the validity of EU law. 57 Assessments of the facts of the case, as well as the need for a preliminary ruling in order to allow a national court to deliver judgment, are matters for national courts. 58 The relationship between courts is supposed to be based on cooperation and a distribution of functions. On the whole, the preliminary reference procedure has allowed the CJEU to become a very effective prointegrationist 'change actor'. 59 Nevertheless, the CJEU is not 'at the apex of an integrated judicial hierarchy' 60 and the CJEU's pro-active role in promoting integration has not been without causing strains in its relations with national courts wary of seeing the CJEU's integrationist case law encroaching on the autonomy of national legal orders and national institutions. 61 The literature identifies 57 Art. 267 TFEU. There is no need here to go into the detail of the separation of functions between EU and national courts. The only precision worth adding is that whilst national courts are prevented from ruling that an EU act is invalid, they are not prevented from ruling that it is valid. See  various ways in which these courts might seek to 'resist' or 'pushback' on the CJEU's rulings: for example, by failing to follow its interpretations, or by failing to refer a case to the European court. 62 Among Europe's most influential courts is the BVerfG. As Mayer points out, the BVerfG 'never quite accepted the ECJ's role as final arbiter'. 63 He notes that the BVerfG is effectively interpreting EU law in the context of its ultra vires doctrine, 64 which he describes as corresponding to a 'German constitutional law-based reserve of power over European acts'. 65 The BVerfG's recent judgment on the ECB's Public Sector Purchase Programme (PSPP) perfectly illustrates Mayer's point. 66 The PSPP is among the ECB's nonstandard monetary policy measures and is supposed to help the ECB fulfil its price stability mandate. 67 Since this programme concerns the ECB's role as guardian of the single currency and not its role as bank supervisor under the SSM, I will not dwell on this decision. Suffices to say that the judgment made history because the BVerfG found for the first time that the ECB and the CJEU, which in Weiss found the PSPP to be in accordance with EU law, 68 had acted ultra vires. Thus, the BVerfG held that Weiss had 'no binding force in Germany' 69 and that after a transitional period that could not exceed three months, the German central bank was prevented from participating in the implementation and execution of the relevant ECB decisions underpinning the PSPP, unless the ECB addressed the failings that the BVerfG had identified in its judgment. 70

Dimensions of change
Following Campbell's call to be explicit about the institutional dimensions that one seeks to track, my final point is to identify which dimensions have analytical relevance for studying change in the SSM. Hereinafter, I will introduce three institutional dimensions. I will coin them competence, powers, and governance. These dimensions are analytically salient because they are directly relevant to what is underpinning much of the wrangling about institutional choices in the field of financial market/banking supervision: that is, the extent of centralization and European integration. 71 They represent the most important variables in battles between the EU and Member States about the institutional shape of financial supervision. I will briefly elaborate on each of these dimensions.
'Competence' is the most powerful concept for our purposes. Competence determines which actor has the authority to act in a given field or sector. For our purposes, it is useful to think of 'competence' as having a Treaty sense and a legislative sense. Under the Treaties, the two main relevant concepts are shared and exclusive Union competence. Areas of exclusive Union competence are those where 'only the Union may legislate and adopt legally binding acts'. 72 Member States can only do so themselves 'if so empowered by the Union or for the implementation of Union acts'. 73 In areas of shared competence, both 'the Union and the Member States may legislate and adopt legally binding acts'. 74 However, Member States can only exercise their competence 'to the extent that the Union has not exercised its competence'. 75 However, there is a second sense-a legislative sense-in which competence can be understood. In this second sense, 'competence'-that is, the legal authority bestowed on an actor to act-is the outcome of choices made by the EU legislature when adopting an act under a legislative procedure in a field of shared (Treaty) competence. Thus, in the area of supervision, the EU legislature might decide to vest an EU actor (eg an EU agency) with supervisory competence in a given sector. In this context, it can also decide to confer this actor with exclusive competence to carry out supervisory tasks. This conferral of exclusive supervisory competence will, within the bounds set by the EU legislative act, result in a loss of national supervisory competence. In this sense, exclusive competence is to supervision what maximum harmonization is to rule-making. For example, in an 71 See also M Thatcher and D Coen, 'Reshaping European regulatory space: an evolutionary analysis' (2008) 31 West European Politics, 806-36, 809 who in their work on the evolution of the European regulatory space refer to 'strong debates about the extent of centralisation of powers at the European level and the respective roles of the European Commission and national and EU regulatory agencies, showing that implementing institutions link to wider political battles about integration and the form of the EU'. 72 Art. 2(1) TFEU. 73 Ibid. 74 Art. 2(2) TFEU. 75 Ibid. internal market context, the European Securities and Markets Authority (ESMA) is exclusively responsible (or competent, in my terminology) for the registration and supervision of credit rating agencies (CRAs). 76 Since the internal market is an area of shared competence, the choice to vest ESMA with this exclusive authority was made by the EU legislature when adopting an EU regulation under Article 114 of the Treaty on the Functioning of the European Union (TFEU). Moreover, since the principle of subsidiarity applies in areas of shared competence, the regulation was adopted in accordance with this principle. 77 Following its adoption, national authorities are prevented from unilaterally performing supervisory tasks that are within the scope of ESMA's competence. If a national authority carries out such supervisory tasks, it will do so on the basis of a delegation by ESMA. 78 'Powers' is about the means that an actor has to exercise its allocated competence. For our purposes, we can think of powers as a broad, rather diffuse, category. In our context, it includes the power to take individual decisions (eg a whole range of supervisory and investigatory powers). It also includes general decision-making powers (eg the power to issue instructions to a group of actors). The addressees of these powers are market actors, but also, depending on the power, NCAs. The important point is that in our conceptualization, powers and competence are essential aspects of authority. They are intrinsically linked: without adequate powers, competence is a meaningless concept; without competence, an actor lacks the authority to exercise its powers.
'Governance' comes into play where common structures are established: for example, a forum, a committee or an agency. Such common structures can have diverse tasks: for example, to improve cooperation or coordination, to draft rules, to ensure the implementation of rules by NCAs or their application by market actors. In an EU context, the establishment of such bodies raises the question of who has the say in such structures: that is, (i) representatives from NCAs who act as members of such structures, but who remain accountable for their actions at the national level and who are accordingly likely to 'see the world' primarily through a national lens; or (ii) full-time appointees who are employed by such structures and whose self-interest is accordingly more closely intertwined with the fate of the common structure and who can therefore be assumed to have greater incentives to defend its objectives. 79  two ideal-type composite actors that Scharpf identifies: collective actors and corporate actors. Scharpf defines collective actors as those that cannot determine their courses of action autonomously, but are instead 'dependent on and guided by the preferences of their members'. 80 Corporate actors meanwhile are more hierarchical structures under the control of a leadership. Members (if any) are 'not actively involved in defining the corporate actors' course of action . . .'. 81 According to Scharpf, these actors may allow a 'degree of effectiveness and efficiency that collective actors depending immediately on membership preferences could not achieve'. 82 For our purposes, the relevant point is that the say of NCAs will be significantly diminished in the case of corporate actors given that decision making is in the hands of appointed members and 'decoupled from the preferences of the membership'; 83 whereas it will be less so in the case of collective actors since NCAs' membership of the decision-making organs will ensure a continuing say over the actions of such actors. 84 To conclude, it is submitted that competence, powers, and governance are important institutional dimensions to account for when seeking to ascertain the reality of change in the field of bank supervision. They are important because they have proven to be salient variables in the wrangling between the EU and Member States about the institutional shape of financial supervision. To be sure, even after specifying institutional dimensions carefully, the question of whether change should be characterized as evolutionary or revolutionary, continuous or discontinuous, can be difficult to answer. Indeed, describing change in such simple dichotomous terms may often not do justice to a complex reality. Admittedly, thinking of change in terms of a continuum, as Campbell suggests, represents an improvement on this state of thinking. 85 Even so, one needs to remain alert to the risk of presenting impressionistic views when attempting to account for the reality of change. At any rate, more important than to insist on particular labels (evolutionary, revolutionary, etc) are for our purposes the mechanisms by which change occurs over time in the SSM and in this context the role which courts, as actors of change, have played in the continuous evolution of institutional dimensions. Before addressing this point, it is however first necessary to set the stage by examining the rules that were enacted in the SSM Regulation. 80 F Scharpf, Games Real Actors Play-Actor-Centered Institutionalism in Policy Research (Boulder, CO: Westview, 1997), 54. 81 Ibid 56. 82 Ibid 57. 83 Ibid 56. 84 Of course, the actual say will greatly depend on the decision-making procedures and especially the voting requirements (e.g. unanimity, qualified majority, simple majority). 85 See text to (n 21).

III. Institutional change and the creation of the Single Supervisory Mechanism
For most of the EU's existence, national supervision has been the approach in the financial sector. Today the picture is markedly more complex and this is especially obvious in the Eurozone where efforts to deal with the sovereign debt crisis led to the creation of the Banking Union. The SSM is the first pillar of the Banking Union. In brief, the SSM is about prudential supervision of credit institutions and to a lesser extent about macro-prudential supervision (that is, the supervision of the financial system as a whole). The aim of this section is to gain some initial insights on institutional change in the SSM by examining the rules that were adopted in the SSM Regulation. Specifically, this section focuses on the moment of institutional creation by examining the changes brought about by the SSM Regulation along the three dimensions that I introduced earlier: competence, powers, and governance. By doing so, this section will prepare the ground for an examination in the next section of the role played by courts in the evolution of institutional dimensions -that is, especially the 'competence' dimension, which has proved to be the most contested dimension. Hereinafter, I will focus on micro-prudential supervision, which is at the heart of the SSM Regulation and the area that has seen the most activity before the courts. 86 Each dimension is reviewed separately (competence in Section III.A, powers in Section III.B and governance in Section III.C), after which I will reflect on change during this phase of institutional creation (Section III.D).

A. Competence dimension
The most apparent change which the establishment of the SSM brought about was a transfer of prudential supervisory competence over credit institutions from the national to the EU level. By vesting the ECB with bank prudential competence, the ECB was given an entirely new role to play. No longer was it just the guardian of the euro. It was given an important role to play in a core area of national economic activity, that is, banking.
The ECB's new role was enacted in the SSM Regulation and further fleshed out in an ECB regulation (the SSM Framework Regulation). 87 The Treaty basis on which the SSM Regulation was adopted is Article 127(6) TFEU. According to this provision, the Council can 'confer specific tasks upon the European 86 As far as macro-prudential tasks are concerned, suffices to say that the role of the ECB is more limited. The ECB was vested with some powers to apply EU macro-prudential tools, but the primary responsibility for the macro-prudential toolkit continues to rest with national actors (see Art. 5, SSM Regulation). 87  That said, the transfer of bank prudential competence was not without complications. It rests horizontally on a complex interplay between different provisions of the SSM Regulation. Vertically, it rests on a potentially uneasy relation with the Treaty legal basis on which the SSM Regulation is based. The extent of these complexities ought to be examined since resolving them, one way or another, has important consequences for determining the nature and scope of the ECB's competence in the prudential field. Starting with the Treaty basis, it is fair to say that there has been much debate about what Article 127(6) entails in terms of conferring prudential competence on the ECB. For one thing, Article 127 (6) is in a Treaty chapter on monetary policy, which is an area of exclusive Union competence and not an area of shared competence. 88 On the other hand, however, the literature has made very good arguments why Article 127(6) TFEU did not transform prudential supervision-the subject matter of the SSM-into an exclusive Union competence, 89 where Member States are excluded from acting unless authorized by the Union. Questions have also been raised with respect to the wording of Article 127(6) and especially whether the language of 'confer[ring] specific tasks upon the European Central Bank concerning policies relating to the prudential supervision . . .' offers a suitable basis for vesting bank prudential supervision in the ECB. 90 According to a narrow reading of this provision, the term 'specific' seeks to limit in substance what the ECB can be vested with. Thus, according to this reading, Article 127(6) does not permit a full transfer of bank supervision to the ECB. However, according to an alternative reading, the term 'specific' affects form rather than substance. Under this reading, 'specific' merely requires tasks to be clearly identified, so as not to be 'generic'. 91 These questions were largely left unresolved, but it is apparent that the text of the SSM Regulation was drafted with the wording of Article 127(6) in mind. Thus, the ECB was given a list of 'tasks', which are found in Article 4(1) of the SSM Regulation. However, these tasks are typical of any bank (prudential) 88  supervisor. 92 Simplifying, they are about overseeing credit institutions, applying relevant rules, and making sure that they are complied with.
The second complication concerns the interplay between certain provisions of the SSM Regulation which are crucial for determining the scope of the ECB's competence under the SSM Regulation. The political negotiations over the ECB's competence have been extensively examined in the literature. It is well known that the Commission proposal to vest the ECB with exclusive competence to perform supervisory tasks over all credit institutions established within the Banking Union faced resistance by a group of Member States led by Germany. 93 The latter sought to maintain national supervision for smaller banks 'in line with lobbying from the German Landesbanken (regional banks) and Sparkassen (savings banks)'. 94 As Merino notes, in the end it was agreed to differentiate between 'significant' entities, which are basically large credit institutions, and 'less significant' entities, which are in essence smaller banks. 95 The political compromise is found in Article 4(1) and Article 6 of the SSM Regulation. 96 As noted, the former provision specifies the bank prudential tasks which the ECB was vested with. Specifically, it states that the ECB is 'exclusively competent' to carry out those tasks in relation to 'all credit institutions' that are established in Member States that participate in the Banking Union. However, Article 4(1) must be read together with Article 6. The latter is an especially convoluted provision. For our purposes, it is sufficient to note that because of the scheme set out in Article 6, 97 NCAs generally supervise less significant entities. Thus, NCAs carry out Article 4(1) tasks 98 in relation to less significant entities-except for a couple of tasks that are not dependent on the significance of a credit institution and that are a matter for the ECB also in the case of less significant entities (eg authorizing credit institutions or approving a qualified holding acquisition in a credit institution 99 ). Significant entities are meanwhile directly supervised by the ECB. 100 Article 6(4) sets out criteria that allow differentiating between significant and less significant entities.
That said the relationship between Article 4(1) and Article 6 is ambiguous and this ambiguity is at the origin of uncertainties about the allocation of competence 92 See eg Wymeersch (n 8), 13 noting that the ECB's tasks constitute the 'core activity of any prudential supervisor'. 93  with respect to less significant entities. Witte notes in this context that the arrangement between Article 4(1) and Article 6 can be read in different ways. 101 On the one hand, one can give importance to Article 4(1) and the fact that it talks of the ECB as having exclusive competence for the tasks in Article 4(1) in relation to 'all credit institutions'. Thus, according to this reading, the ECB is exclusively competent for the supervisory tasks in Article 4(1)-both in relation to significant entities and in relation to less significant entities; NCAs continue supervising less significant entities merely because they were delegated certain responsibilities in accordance with Article 6. 102 On the other hand, however, one can put greater weight on Article 6 and the opening wording of Article 4(1) which specifically refers to Article 6 ('[w]ithin the framework of Article 6, the ECB shall . . . be exclusively competent . . .'). Under this reading, one might take the view that the scope of the ECB's exclusive competence is subject to the scheme of Article 6. Accordingly, one might argue that the supervision of less significant entities is, because of the reference to Article 6 in Article 4(1), excluded from a transfer of competence to the EU level. 103 Hence, one might conclude that NCAs continue to carry out their tasks in relation to less significant entities on the basis of pre-existing national competence. 104 Arguably such a reading is dovetailed by Article 127(6) TFEU, which only envisages that tasks be conferred on the ECB, and not on NCAs.

B. Powers
To carry out its prudential tasks under the SSM Regulation, the ECB was given extensive powers. These include supervisory and investigatory powers that are explicitly set out in the regulation. 105 They also extend, pursuant to Article 9(1) of the regulation, to the powers conferred on competent authorities under relevant EU law (eg under capital requirements legislation). 106 Indeed, where necessary for the ECB to carry out its tasks and where the regulation does not vest such powers in the ECB, the SSM Regulation also authorizes the ECB to act by way of instructions and require NCAs to make use of their powers 'under and in accordance with the conditions set out in national law'. 107 Besides these supervisory and investigatory powers, the ECB was also given powers to manage the division of responsibility with NCAs that supervise less 101 Witte (n 96), 652-3. 102 Ibid. 103 Ibid. That said, recall that in relation to a couple of tasks, such as the authorization of a credit institution, the distinction between significant and less significant entities does not matter, and accordingly these tasks remain with the ECB regardless of significance (see text to n 99 above). 104 Ibid. 105 Arts 9-18 SSM Regulation. See especially, Art. 10 (investigatory powers); Art. 11 (general investigations); Art. 12 (on-site inspections); Art. 14 (authorization); Art. 15 (assessment of acquisitions of qualifying holdings); Art. 16 (supervisory powers); Art. 18 (administrative penalties). 106 Art. 9(1) sub para. 2. 107 Art. 9(1) sub-para. 3. significant entities. Thus, the ECB can issue regulations, guidelines, or general instructions to NCAs. 108 It is empowered to ask the latter for information on the performance of their tasks. 109 It can make use of its investigatory powers. 110 The ECB also has a general oversight role to play. 111 Crucially, the ECB can decide to directly supervise a less significant entity 'when necessary to ensure consistent application of high supervisory standards'. 112 This latter power is especially noteworthy. It means that the ECB can take over the supervision of a less significant entity from a national authority either on its own initiative or following a request from an NCA. 113 In sum, the SSM Regulation puts the ECB at the heart of the SSM. Member States agreed to confer important powers on the ECB which were previously exercised by NCAs. The ECB was also given direct powers over NCAs that are involved in bank supervision under the SSM. The distinctiveness of the SSM arrangements is further underscored by the fact that the SSM Regulation not only requires the ECB to apply Union law, but also national laws that implement directives. 114 Moreover, where Member States benefit from options under EU regulations, the ECB must also apply the national laws which exercise those options. 115 These arrangements which see the ECB applying national law, have been described as 'rather unprecedented in the law of the Union'. 116 That said, the exercise by the ECB of its powers can nevertheless be subject to complications. Questions have for example been raised with respect to the powers that are available to the ECB under Article 9(1) of the regulation, a provision that I singled out above and which enables the ECB to benefit from powers vested in competent authorities under Union legislation (eg the Capital Requirements Directive). 117 Complications also arise because of the different 108 Art. 6(5)(a). 109 Art. 6(5)(e). 110 Art. 6(5)(d). 111 Art. 6(5)(c). See also Art. 6(1) which states that the 'ECB shall be responsible for the effective and consistent functioning of the SSM'. 112 Art. 6(5)(b). It can also use this power in case of financial assistance by the European Stability Mechanism (or the European Financial Stability Facility). 113 Art. 6(5)(b). 114 Art. 4(3). 115 Ibid. of law that the ECB must apply when performing its tasks, and indeed because of the continuous role of NCAs in the SSM. The latter not only supervise less significant entities; they are also involved at an operational level in the exercise by the ECB of its SSM tasks. They prepare, for example, draft decisions proposing to the ECB to grant authorization to a credit institution; 118 they also participate in so-called 'joint supervisory teams' (JSTs) which supervise significant entities or significant supervised groups. That said, despite the involvement of NCAs at this operational level, the ECB has not shied away from asserting its role. 119

C. Governance
The final dimension to consider is the governance dimension. At the outset, it is worth acknowledging that any pooling of authority at EU level is bound to bring about a reduction of authority for individual Member States and their NCAs. But as already suggested, the extent of this reduction will nevertheless vary depending on the 'collective' or 'corporate' nature of a composite actor. 120  a day-to-day supervisor, but only in a few fields. These fields are, in contrast to banking, not at the heart of Member State economies and have no true legacy of NCA supervision. 123 The ESAs were, like their predecessors, conceived as collective actors. 124 As a reminder, according to Scharpf, collective actors are actors that 'are not autonomous in the choice of the preferences that guide their actions'. 125 Rather they depend on member preferences. It is plain that the ESAs are akin to collective actors in the sense that the ESAs' main decision-making bodies-the boards of supervisors-are dominated by NCAs. The latter hold voting rights and have the final say on the ESAs' courses of action. 126 Moreover, the latest revisions of the ESAs' founding regulations testify to the importance that Member States attach to this state of affairs. Thus, Member States rejected attempts by the Commission to vest some decision-making powers in newly appointed independent members. 127 To be sure, the influence of the ESAs' chairpersons-a full-time, independent professional-was strengthened following the latest 123 ESMA was vested with supervisory responsibility for Credit Rating Agencies and Trade Repositories. Following the recent reforms of the founding regulations of the ESAs, ESMA will in addition have responsibility for data reporting services providers and certain financial benchmarks There are variations depending on the decision to be adopted (e.g. an Art-17 decision on breach of Union law might be taken by a written procedure and considered adopted unless rejected by a simple majority, see ibid, Art. 44(4)). Moreover, in the case of EBA, additional voting requirements were introduced in 2013 in order to accommodate the concerns of some Member States (chiefly, the UK) which were not part of the Banking Union. See in this context, Schammo (n 79). 127 Of particular relevance were the Commission's proposals to entrust these appointed members, as members of a new so-called Executive Board, with (inter alia) the power to make decisions in the case of disagreements between NCAs or alleged breaches of Union law by NCAs (for the proposals, see Commission (n 123) 20 and 22-3). If adopted, this would have been a significant change which would have undermined the authority of national representatives, acting collectively within the supervisory boards of the ESAs, over these procedures. revisions. On some matters, s/he will now have a right to vote in the ESAs' supervisory boards. 128 However, adding the Chairperson to the voting members of the Board of Supervisors evidently does not call into question the dominant role of NCAs within the Board of Supervisors of each ESA.
Decision-making within the ECB offers a contrasting picture. The transfer of supervision to the ECB entailed agreement on a governance set-up that has much less fit with the description of a collective actor. To illustrate this point, it suffices to focus on the broad lines of this set-up in the micro-prudential field. 129 The 'planning and execution' 130 of the ECB's tasks is according to the SSM Regulation in the hands of the Supervisory Board, an internal organ, which is made of national representatives of participating Member States and appointed members, which are the chair and vice-chair of the Supervisory Board and four ECB representatives. 131 The Supervisory Board prepares draft decisions, but the final decision-making authority rests with the ECB's Governing Council which is made of the governors of central banks in the euro area and appointed members (six members of the ECB's Executive Board, including the President of the ECB and its Vice-President). 132 Thus, from the present perspective, a key difference between the ESAs and the ECB (in its role as bank supervisor) is the place and number of appointed members, with voting rights, in the governance set-up of the ECB. As already noted, full-time appointees can, in contrast to national representatives, be assumed to have better incentives to defend the interests and objectives of the EU institution to which they belong. That is not to say that the ECB is a 'corporate actor' in the sense in which Scharpf uses this term-that is, a hierarchical structure where 'strategy choices are decoupled from the preferences of the membership'. 133 . Nevertheless, in comparison to the ESAs, appointed members play a much more prominent role in the ECB's governance set-up. As a result, the ECB is a much more hybrid actor than the ESAs. 134 Thus, in deciding to vest bank prudential supervision in the ECB, Eurozone states agreed to share decision-making authority in a way that they continue to resist in relation to the ESAs. Granted, the Treaty provisions governing the ECB acted as an effective constraint on the type of decision-making arrangements that political leaders could envisage. Membership of the Governing Council, the ultimate decision-making organ of the ECB, is anchored in the Treaties. 135 However, the same is not true of the Supervisory Board, which is an internal organ, and whose membership was decided at the point of institutional creation of the SSM. Moreover, even if vesting prudential supervision in the ECB was a somewhat natural choice, 136 it was not the only conceivable choice. Placing a collective actor, such as an EU agency, at the heart of the SSM was among the options, although implementing this choice would have been legally more challenging.

D. Assessing institutional change
In approaching the question of institutional change and how best to characterize the changes brought about by the SSM Regulation, it is worth beginning by acknowledging the complexities of the institutional set-up. On the one hand, NCAs are part of the SSM. They continue supervising less significant entities and remain involved at the operational level for the supervision of significant entities. Decision making within the ECB also continues to depend on national input. These elements point to a level of continuity. On the other hand, however, one should not underestimate the magnitude of some of the changes. The ECB was vested with direct supervisory competence over the largest credit institutions (significant supervised entities) in the states that participate in the Banking Union. Together, these account for almost 82 per cent of banking assets in these states. 137 By agreeing to transfer prudential supervision from the national to the ECB level, Eurozone states agreed to hand over control over key actors of their national economies to the ECB. As Howarth and Quaglia point out, this transfer 'could only ever have been achieved in exceptional circumstances-which the 133 Scharpf (n 80), 56. 134 Schammo (n 132), 27. Petit describes the mix of national representatives and appointed members as reflecting a mix of transnationalism and supranationalism (see Petit (n 129), 113). 135 See Art. 283 TFEU. 136 Merino (n 89), para. C.2.1. who also examines why the ECB was in the end chosen over alternatives such as eg a regulatory agency established pursuant to Art. 114 TFEU. 137 https://www.bankingsupervision.europa.eu/about/thessm/html/index.en.html. sovereign debt crisis created'. 138 Agreement represented a major shift in policy thinking-a quid pro quo for contemplating a direct future recapitalization of failing banks. 139 Indeed, Eurozone states not only agreed to pool authority. By deciding to vest supervisory authority in the ECB, they agreed to share authority with independent appointees within the ECB's decision-making structures, something that Member States have resisted in relation to the ESAs and which as a result continue to better fit the description of collective actors. Moreover, as was shown above, the ECB was given significant powers to carry out its duties, including the power to replace an NCA as supervisor of a less significant entity and the power to apply national laws.
Hence, it is possible to identify significant changes along all three dimensions (competence, powers, and governance). If change is, as Campbell suggests, best conceived as a continuum, the overall picture is arguably one that is somewhat closer to being discontinuous than continuous. To be sure, even after specifying each dimension carefully, there is room to disagree on how best to characterize change overall. Indeed, as shown, there are elements of discontinuity and continuity in the above description, notwithstanding that the original trigger of the reform process was an exogeneous shock of the type described in the punctuated equilibria literature. 140 However, whether we label the overall state as amounting to continuity or discontinuity is not in the end the most relevant point. The more important point for the present purposes concerns how much of the changes that were enacted in the SSM Regulation were left open to interpretation and therefore open to future change and contestation. This brings me to the question of the supervision of less significant institutions and what it tells us about change in the SSM.
As pointed out earlier, the relationship between Article 4(1) and Article 6 is ambiguous and can be subject to contrasting readings. One reading supports the view that NCAs retain their original competence to perform SSM tasks in respect of less significant entities. 141 A second, alternative, reading supports the ECB's exclusive competence with respect to all credit institutions, including less significant entities. The first reading downplays the salience of change in the competence pillar. Change is akin to what Thelen and others describe as layering: 'an existing institution is not replaced, but . . . new institutional layers-these might for instance be rules, policy processes or actors-are added to it'. 142 Thus under this reading, the SSM Regulation does mostly not call into question the competence of NCAs in respect of less significant entities. However, it adds on top of this national layer of supervisory competence, a new role (or new 'layer') for the ECB, which is vested with, inter alia, an oversight role and powers under Article 6 which it can exercise with respect to NCAs that supervise less significant entities. 143 Under the second reading, national competence is replaced by ECB competence in respect of all supervised entities, including less significant entities. NCAs act as mere agents of the ECB when exercising their tasks in respect of less significant entities. Change in the competence dimension is further amplified and pushed further along the discontinuous path.
That said, it is of course true that, whatever reading one adopts, the ECB's powers with respect to NCAs remain unchanged. They are those that are vested in the ECB under the SSM Regulation. However, besides impacting national sovereignty differently, these contrasting readings may have important implications for the way in which the ECB exercises its powers. That is to say, an ECB whose competence excludes less significant entities can be expected to behave differently to an ECB whose exclusive competence extends to all entities. In the first case, NCAs continue to be subject to national accountability lines (eg national parliaments) when carrying out their SSM tasks, and the legitimacy of their actions continues to be closely intertwined with the national level. Accordingly, one would expect the ECB to behave more consensually towards NCAs. However, in the second case, the ECB's exclusive competence over all supervised entities ensures that it has the necessary legitimacy to behave in a more hierarchical manner towards NCAs and to apply its powers in a more pro-active way. 144 NCAs will have no choice other than to acknowledge the ECB as their principal. To be sure, the question of how in practice the ECB behaves towards NCAs is one that is best addressed empirically. However, in the absence of robust empirical work on this issue, the above assumptions are nevertheless useful and reasonable.
Arguably, a reading which puts the ECB front and centre in the competence dimension, would also offer strong legitimizing reasons to those who wish to strengthen the ECB's powers over NCAs in the future. Although the ECB does have significant powers, including the power to remove an NCA from the prudential supervision of a less significant entity, its powers have some noticeable limitations. For example, whilst the ECB can issue general instructions on how NCAs 142 J van der Heijden, 'Institutional layering: a review of the use of the concept' (2011) 31 Politics, 9-18, 11. 143 For details, see Section III.A. and B. above. 144 See also the assessment by Witte (n 96), 653.
should perform their SSM tasks, it cannot currently address individual instructions to an NCA. 145 A pro-ECB interpretation of the competence mandate will offer ammunition to actors who wish to see these restrictions removed in the future. Indeed, arguably, a reading that supports ECB competence at the expense of Member State competence might in the future also offer support to those who favour a stronger role for appointed ECB members in the Supervisory Board since arguably in such a case the rationale for national membership of the Supervisory Board is diminished. Admittedly, implementing such changes through legislative amendment of the SSM Regulation would not be without complications since any amendment would need to meet the unanimity hurdle of Article 127(6) TFEU. However, change might still come incrementally through future interpretations of the provisions of the SSM Regulation. In this context, a reading that puts the ECB front and centre in the competence dimension might in the future come to support interpretations of other SSM rules that strengthen the ECB's position, say, with respect to the scope or range of powers that are available to the ECB. 146 Accordingly one may posit that change in one dimension (the 'competence' dimension) may over time contribute to change in other dimensions (the 'power' dimension or the 'governance' dimension) through a sort of feedback dynamic.
To sum up, the more we zoom in on issues of competence, the more apparent that the way in which we characterize change-or the extent of change-depends importantly on how we resolve the ambiguities that are inherent in the SSM Regulation. Accordingly, in order to appreciate the extent of change, it is necessary to look beyond the point of institutional creation and to consider the role of courts in addressing legal ambiguities. Since the ambiguities in question stem from an EU regulation, our focus naturally shifts to the CJEU as the ultimate authority to interpret EU law in the EU legal order. Thus, in our conceptualization, the question of how to characterize change (eg as layering or a more radical replacement) is itself a question of interpretation. However, given that the CJEU is not at the top of an integrated judicial system, this characterization is not immune from contestation. Hence, from our vantage point, institutional change may not only be subject to contestation before courts, but also between courts: that is, between the CJEU and national courts. It is to this point that I turn now.

IV. Institutional change and the courts
This section turns to the role played by courts in the evolution of the SSM, that is, especially in the evolution of the first of the three institutional dimensions that 145 Art. 6(5)(a) SSM Regulation, referring to 'general instructions' only. This power to instruct NCAs should not be mixed up with the one under Art. 9(1). Recall that according to Art. 9(1), the ECB can instruct national authorities to use their powers where this is necessary for the ECB to carry out its tasks under the SSM Regulation and where the latter does not confer such powers on the ECB. 146 See eg the uncertainties with respect to Art. 9(1) which I highlighted above (see n 117 above). I examined above: the competence dimension. For this purpose, this section examines the decision(s) of the CJEU in Landeskreditbank Baden-Württemberg v ECB (or L-Bank) and the judgment of the BVerfG on the Banking Union. 147 It will draw on Mahoney's and Thelen's insights on modes of institutional change in order to explain how each of these rulings contributed to the change problematic. I will begin by discussing the CJEU's decision(s) in L-Bank (Section IV.A), after which I will focus on the BVerfG's judgment on the Banking Union whose importance from an institutional change perspective remains largely underappreciated (Section IV.B). Before proceeding, it is worth noting that the judgment of the BVerfG also deals, besides the SSM, with the Single Resolution Mechanism, which is the second pillar of the Banking Union. My focus here is solely on the SSM.

A. The CJEU's decision(s) in L-Bank
In L-Bank, the General Court (GC) was asked to decide an annulment action. L-bank wished to be supervised by the German NCA (the Bundesanstalt für Finanzdienstleistungsaufsicht or Bafin). Thus, L-Bank-a 'significant entity' according to the ECB-asked the ECB to agree to its reclassification based on a specific provision of the SSM Regulation. Under this provision, the ECB can decide that a credit institution, which would ordinarily be considered 'significant', should-because of 'particular circumstances' 148 -be reclassified and accordingly be supervised nationally. However, the ECB refused to re-classify L-Bank. The latter challenged the ECB's decision before the GC. The GC sided with the ECB and dismissed the action. The GC's decision was subsequently upheld on appeal.
Admittedly, there was little at the outset to suggest that this annulment action was destined to become a case of major significance for the SSM. The main issue was relatively innocuous from an institutional point of view. However, the end result was markedly different. L-Bank has become a landmark case on competence in the SSM. This is because in L-Bank, the GC took the opportunity to make sweeping statements about the allocation of prudential competence to the ECB. In short, the GC interpreted the provisions of the SSM Regulation in a way that put the ECB front and centre in competence terms. Specifically, the GC found that the ECB had exclusive competence in respect of the tasks set out in the SSM Regulation (Article 4(1)) and that NCAs carried out their SSM tasks with respect to less significant entities as a result of a 'decentralized implementation' of the ECB's exclusive competence 149 and not as a result of 'the exercise of a national competence '. 150 In coming to this finding, the GC examined the interplay between two key provisions which we discussed earlier, that is, Article 4(1) which determines the ECB's prudential tasks and provides that the ECB is exclusively competent to exercise those tasks in relation to all credit institutions within the Banking Union; and Article 6 pursuant to which NCAs are generally responsible for the Article 4(1) tasks in relation to less significant entities. 151 The GC found that this interplay did not support a distribution of competence between the ECB and NCAs. 152 In reaching this conclusion, the GC placed greater weight on Article 4(1). For the court, it was 'at the stage of the definition of the tasks' vested in the ECB by Article 4(1) that competences were distributed. 153 Article 6 was merely there to enable a 'decentralized implementation' of the ECB's exclusive competence by NCAs in respect of less significant entities. 154 Moreover, according to the GC, its findings were supported by both the text and the recitals of the SSM Regulation. Thus, the court inferred from the recitals that direct supervision by NCAs under the SSM was merely envisaged as a 'mechanism of assistance to the ECB' and not as an 'exercise of autonomous competence'. 155 The ECB's powers over NCAs (ie the power to issue regulations, guidelines, and general instructions, as well as the power to take over the supervision of a less significant entity), were for the GC also 'indicative of the subordinate nature of the intervention by the NCAs in the performance of [the Article 4(1)] tasks'. 156 The fact that the Council, which had adopted the SSM Regulation, decided to associate NCAs in the implementation of the tasks under Article 4(1) did 'not allow any conclusions to be drawn as to maintaining prudential supervisory competence for the national authorities' with regards to the tasks in Article 4(1). 157 Remarkably, the opening wording of Article 4(1), which stipulates that it is '[w]ithin the framework of Article 6' that the ECB is exclusively competent to perform the tasks of Article 4(1), played no part in the GC's assessment of the scope of the ECB's competence under the SSM Regulation.
The decision of the GC was subsequently appealed before the Court of Justice (CJ). 158 The latter was arguably a bit more circumspect. 159 But the CJ confirmed the decision of the GC entirely. 160 Indeed, at no point, when considering the interplay between Article 4 and Article 6, did the CJ suggest that the GC was 151 By tasks I mean those found in Art. 4(1)(b), (d) to (g) and (i). See Art. 6(6). 152 Case T-122/15 (n 12), para. 54. 153 Ibid, para. 56. 154 Ibid, para. 63. 155 Ibid, para. 58. In fact, the court made this inference based on the way in which the recitals were arranged in the regulation. 156 Ibid, paras 59-61. 157 Ibid, para. 64. 158 Case C-450/17 P (n 12). 159 For example, the CJ was careful to avoid referring to the existence of an exclusive Union competence as the GC had done (see Case T-122/15 (n 12), para. 72). It simply referred to 'exclusive competence'. For further details on this point, see below. 160 See especially para. 49, in which the CJ referred to paras 54, 63, and 72 of the GC's decision. wrong and that NCAs carried out their SSM tasks on the basis of pre-existing (national) competence. Instead, the CJ confirmed the GC's views when noting that the latter's interpretation of the ECB's competence could not be invalidated by L-Bank's arguments which were 'based on the postulate that the national authorities retain, under Article 6(4) of [the SSM Regulation] their competence . . .' for performing the tasks of Article 4(1). 161 AG Hogan was more forthright on this point in his opinion when stating that he could not agree 'with the claim by the appellant that [national] authorities retain their pre-existing competences in respect of [less significant] entities'. 162 Referring to the GC's decision, the CJ concluded that NCAs were acting within the scope of a decentralized implementation of an exclusive competence of the ECB, which was enabled by Article 6. 163 Pizzolla argues that in approaching the SSM Regulation, one should differentiate between competence, powers, and responsibilities. 164 The distinction appears to agree with the GC's and CJ's approach. NCAs are responsible for carrying out certain tasks under Article 4(1) in relation to less significant entities. They can exercise relevant powers and are authorized to adopt all relevant supervisory decisions in relation to these entities. However, when it comes to the competence question, matters are different. According to the CJEU, NCAs act within the scope of a decentralized implementation of an ECB exclusive competence and not on the basis of pre-existing national competence when performing their SSM tasks in relation to less significant banks.
To sum up, in a decision that 'surprised practitioners and academics alike' 165 the GC tipped the balance of competence between the ECB and NCAs resolutely in the ECB's favour. When deciding that NCAs acted within the scope of a decentralized implementation of an ECB exclusive competence and not on the basis of national competence, the CJEU rejected a reading of the SSM Regulation that would have been consistent with a 'layering' of rules and which would have preserved the original (national) competence of NCAs in respect of less significant entities. Instead, the GC 'deployed' the provisions and recitals of the SSM Regulation in support of the ECB and legal integration within the SSM. 161 Ibid, para. 36. See also para. 49 which refers explicitly to (inter alia) paras 54 and 72 of the GC's judgment. In the latter paragraph of the GC's judgment, the GC states that '. . . under the SSM the national authorities are acting within the scope of decentralised implementation of an exclusive competence of the Union, not the exercise of a national competence' (emphasis added). 162 See AG Opinion, Case C-450/17 P (n 12) ECLI:EU:C:2018:982, para. 53. 163 Case C-450/17 P (n 12) para. 49, noting that the 'the Council conferred on the ECB exclusive competence, the decentralised implementation of which by the NCAs is enabled by Article 6 of that regulation, under the SSM and under the control of the ECB, in relation to less significant credit institutions'. 164  By the same token, it amplified change in the competence dimension. Enabling this decision were the ambiguities and incompleteness of the SSM Regulation. Hence, L-Bank perfectly illustrates how institutional change can be based on endogenous developments that take place after the moment of institutional creation, and how in this context legal ambiguities can lead to change by way of conversion.
To be sure, L-Bank is not the only decision of the CJEU on the SSM where important institutional developments have taken place in the 'malleable' space between rule enactment and rule interpretation. 166 L-Bank was followed by Berlusconi, a case in which the CJEU ruled in favour of its own jurisdiction to review (national) preparatory acts in so-called composite administrative procedures. 167 Composite procedures entail decision making by both national administrative authorities and EU authorities and can be found across a range of policy areas. 168 The judicial review of these procedures is an area where the Treaties fail to offer straightforward answers and where developments have accordingly taken place incrementally at the hand of the CJEU. Under the SSM, one such composite procedure concerns the assessment of the acquisition of qualifying holdings in credit institutions. 169 It is one of a very few procedures under the SSM Regulation where the distinction between significant and less significant entities does not matter. 170 Whilst the procedure involves NCAs which receive applications for an intended acquisition of a qualifying holding and carry out assessment work, it is the ECB that is solely competent to decide the fate of the application, following a (non-binding) proposal by the relevant NCA. 171 In Berlusconi, the national referring court asked the CJEU whether national or EU courts had jurisdiction to review the legality of (national) preparatory acts in the above procedure. 172 The CJEU held that where national preparatory acts did not bind an EU institution having 'final decision-making power', 173 national courts had to be denied jurisdiction to review national preparatory acts. 174 It was for 166 See also Mahoney and Thelen (n 17), 14 who speak of change that emerges 'in the "gaps" or "soft spots" between the rule and its interpretation'. 167 170 The other such procedure concerns the authorization, and the withdrawal of authorization, of a credit institution. See Art. 14 and text to (n 99) above. 171 See Art. 4(1)(c) and Art. 15 SSM Regulation. There are additional provisions under the CRD, but they can be left aside for the present purposes. 172 Case C-219/17 (n 167). 173 Ibid, para. 43. 174 Ibid, paras 44-51, especially para. 47. The CJEU also held that matters were different where in a composite procedure, EU institutions had 'only a limited or no discretion, so that the national act is binding on the EU institution'. It went on to note that '[i]t then falls to the national courts to EU courts to examine any defects vitiating the national preparatory acts that could taint the validity of EU decisions. 175 The Court sought support for its views in previous case law, as well as in the Treaties-especially the Treaty provision on the action for annulment which it interpreted in light of the general principle of sincere cooperation. 176 Thus, in the case in point, since it was the ECB who had final decision-making authority, it was a matter for the EU courts to review whether the ECB decision was tainted by 'any defects rendering unlawful' the national preparatory acts. 177 That jurisdiction, according to the CJEU, 'exclude[d] any jurisdiction of national courts in respect of those acts . . . '. 178 Berlusconi is widely seen as a case whose significance extends beyond the SSM context. 179 It was also a pillar of the CJEU's recent decision in Iccrea. 180 Since this latter case concerns the Single Resolution Board (SRB)-the EU body at the centre of the Single Resolution Mechanism, the second pillar of the Banking Union-suffices to say that the CJEU applied Berlusconi and its findings on its own jurisdiction when considering the role played by national courts in the context of an SRB composite procedure. 181 Before concluding this subsection, it is worth returning a final time to the CJEU's decisions in L-Bank in order to highlight one additional point about rule on any irregularities that may vitiate such a national act-making a reference to the Court for a preliminary ruling where appropriate-on the same terms as those on which they review any definitive act adopted by the same national authority which is capable of adversely affecting third parties and moreover, in the light of the principle of effective judicial protection, to regard an action brought for that purpose as admissible even if the national rules of procedure do not so provide'. See ibid, paras 45 and 46. 175  , and that national resolution authorities provided only 'operational support' to the SRB (ibid), the CJEU found, in accordance with its Berlusconi case law, that only EU courts had jurisdiction, when reviewing the legality of the SRB's decision, to determine whether a preparatory act of a national resolution authority was affected by defects that could taint the decision of the SRB (ibid 48): 'no national court can review that national act' (ibid). The CJEU also considered the role of national resolution authorities and national courts in the part of the procedure that followed the adoption by the SRB of its decision on bank contributions. At this stage, national resolution authorities are required to notify banks of the contributions that they must make to the SRF, as well as to raise the relevant contributions from them. The CJEU noted that neither the national resolution authorities, nor national courts which were asked to review the actions of the former, could take decisions that 'conflict[ed] with decisions of the [SRB] on the calculation of the ex ante contributions to the SRF and which, in practice, deprive the latter decisions of their effects, by impeding the raising of those contributions' (paras 55-60, especially para. 60). these rulings. It is plain that the decisions of the GC and the CJ were far from resolving all ambiguities and uncertainties. Indeed, they created altogether new ones. 182 Fundamental notions for delimiting the scope of the ECB's exclusive competence-think of the notion of a 'decentralized implementation' of an ECB exclusive competence-were not fleshed out and left open to future interpretation. 183 The question of how the L-Bank rulings square with the Treaty competence construct and especially with Article 127(6) TFEU (the Treaty article on which the SSM Regulation is based) were also not addressed. Recall that the GC's findings on the ECB's exclusive supervisory competence were purely a matter of statutory interpretation: that is, of the binding text of the SSM Regulation and of the text's recitals. Nevertheless, even if one looks in vain for a mention of Article 127(6) TFEU, the GC, when examining a question about subsidiarity, approached subsidiarity as if the ECB benefited from an exclusive Union-that is, Treaty-competence. In such a case, subsidiarity considerations are excluded. Matters are different where, in the absence of an exclusive Union competence, choices about the allocation of supervisory competence are made by the EU legislature when adopting an act under a legislative procedure in an area of shared competence. 184 Subsidiarity ought to apply. Indeed, in the case of the SSM Regulation, it is plain that the regulation was adopted with subsidiarity in mind. 185 Remarkably, however, the GC found in L-Bank that subsidiarity considerations were excluded since 'under Article 5(3) TEU, [the principle of subsidiarity] applies only in areas which do not fall within exclusive EU competence'. 186 Unfortunately, the GC did not offer any additional insight. It moved on quickly without elaborating further.

B. The BVerfG's judgment on the Banking Union
Having examined the decisions of the CJEU, I turn next to the judgment of the BVerfG in order to contrast the findings of the latter court with those of the CJEU. As already noted, the BVerfG rejected the constitutional complaint directed against the Banking Union. However, the more relevant point is that in reaching this conclusion, the BVerfG came to markedly different findings on the scope of the ECB's competence under the SSM Regulation. The essential bone of contention concerns the question of competence over less significant institutions. Recall that according to the CJEU, NCAs carry out their tasks under Article 4(1) as a result of a 'decentralized implementation' of an ECB exclusive competence 187 and not as a result of the exercise of pre-existing national competence. 188 However, for the BVerfG this is not so: NCAs exercise their supervisory tasks with respect to less significant entities on the basis of national competence and not on the basis of ECB competences that were delegated back to them (Rückdelegation). 189 The decision of the BVerfG is densely argued and accordingly one is at risk of failing to see the wood for the trees. Hence, I will only focus on some key points of contrast with the decisions of the CJEU. Specifically, I will show that with respect to the allocation of competence over less significant entities, the BVerfG's findings are broadly consistent with layering: the original (national) layer of competence of NCAs is for the most part preserved; the fact that the SSM Regulation vested the ECB, as far as less significant entities are concerned, with a new role and a range of powers, did not call this finding into question. Hence, I will show hereinafter that by using legal reasoning, which offers legitimacy and an impression of objectivity, the BVerfG was able to 'redirect' the relevant rules in support of a new outcome -that is, one that rejects the idea of a 'decentralized implementation' of an ECB exclusive competence by NCAs and supports a distribution of competence between NCAs and the ECB. In making these findings, the BVerfG contested the extent of institutional change, as 'discovered' by the CJEU in its L-Bank rulings. Enabling this process of conversion were, as in the case of the L-Bank rulings, ambiguous and incomplete legal provisions.
It is plain that the BVerfG approached the question of competence differently to the CJEU. According to the BVerfG, a redelegation of ECB tasks to NCAsas opposed to the case where NCAs act on the basis of pre-existing national competence-would presuppose that bank supervision had been fully transferred to the ECB. 190 However, according to the German court, taking this view would fall foul of Article 127(6) TFEU, the Treaty legal basis on which the SSM Regulation was founded. 191 Hence, unlike EU courts, the BVerfG began by examining the competence question through the lens of the Treaties and the Treaty competence construct. Recall in this context that Article 127(6), and what it precisely entails, has been subject to debate and controversy. Earlier we saw 187 Case T-122/15 (n 12), para. 63; Case C-450/17 P (n 12), 49. 188 Case T-122/15 (n 12), para. 72. 189 See especially BVerfG, (n 2), paras 183-197. This part of the judgment should be read together with paras 172-182. Note that in the English translation of the judgment, paras 188-196, which are a key part of the BVerfG's reasoning on the issue of competence allocation with respect to less significant entities, were not translated into English. I therefore mostly relied on my own translation of the German text. 190 Ibid, para. 185, noting that 'a re-delegation of EU administrative tasks would require that all supervisory tasks had been fully conferred on the ECB. . .'. See also para. 187 noting that '[a] redelegation of EU administrative tasks would require that all supervisory tasks had fully been conferred on the ECB. . .'. 191 Ibid, that the provision can be read in different ways. 192 Recall also that the L-Bank rulings did not offer any new insights on Article 127(6). The EU courts' findings on the ECB's exclusive competence in respect of the SSM tasks, and on a decentralized implementation of this exclusive competence by NCAs, were based purely on an assessment of the SSM Regulation. The BVerfG on the other hand focused extensively on what Article 127(6) permits and-importantly-what it does not permit. 193 It adopted a narrow reading of the wording of Article 127(6) and stressed that Article 127(6) did not transform prudential supervision into an area of exclusive Union competence that could entail a redelegation. 194 Crucially, the SSM Regulation had, as an instrument of secondary law, to be interpreted in light of the higher order (Treaty) legal basis, and by taking account of the subsidiarity principle. 195 Turning to the SSM Regulation, the BVerfG found support for its views on a division of competence between the ECB and NCAs. It examined both the recitals and the binding text of the SSM Regulation, 196 including the interplay between Article 4(1) and Article 6. Recall that Article 4(1) sets out the prudential tasks for which the ECB is exclusively competent in relation to all credit institutions, but because of Article 6, NCAs generally supervise less significant credit institutions. When considering the interplay between these provisions, the BVerfG clearly did not find, as the GC had found, that the 'sole purpose' 197 of Article 6 was to enable a decentralized implementation of the ECB's exclusive competence. Instead, the BVerfG placed greater weight on Article 6 when finding that the ECB was-only within the framework of Article 6-competent for the tasks listed in Article 4(1). 198 Nor did the BVerfG find, as the GC had found, that the recitals of the SSM Regulation supported the view that national supervision was merely a 'mechanism of assistance to the ECB' rather than the 'exercise of autonomous competence'. 199 Instead, the BVerfG effectively downplayed the importance of the SSM Regulation for the supervision of less significant entities, when suggesting, after examining the recitals of the SSM Regulation and those of the SSM Framework Regulation, that the supervision of less significant entities was essentially not the subject of Union regulation. 200 Crucially, the ECB's most potent power under Article 6(5)-that is, the right to take over the supervision of a less significant entity-did not call into question the BVerfG's conclusions on a division of competence. To be sure, the BVerfG acknowledged that the ECB had an oversight role to play and benefited from a range of powers over NCAs, including the power to take over the supervision of a less significant entity. Recall that for the GC, the existence of this latter power (alongside the power to issue regulations, guidelines, or general instructions) was indicative of the subordinate role played by NCAs. 201 Ultimately, it contributed to the GC's finding that the 'sole purpose' of Article 6 was to 'enable a decentralised implementation' of the ECB's exclusive competence. 202 In the BVerfG's analysis, the existence of this power had no such consequence. For the BVerfG, the power of the ECB to take over the supervision of a less significant entity (the ECB's Selbsteintrittsrecht) was a measure of last resort and as such merely an exception which did not call into question the role of NCAs with respect to less significant entities pursuant to the SSM Regulation. 203 Hence, it is submitted that in contrast to EU courts, the BVerfG adopted an altogether different understanding of the relevance of the ECB's Selbsteintrittsrecht for determining questions of competence. In the case of the GC, the existence of this power supported the GC's findings on a decentralized implementation of an ECB exclusive competence. However, according to the BVerfG, it was only in concreto, when exercised, that the ECB's Selbsteintrittsrecht mattered for determining questions of supervisory competence. Thus, as long as the ECB did not exercise its right to take over the supervision of a less significant entity, which according to the BVerfG could only be a measure of last resort, the ECB did not have supervisory competence over a less significant entity 204 -except, of course, for a few supervisory tasks for which the ECB is solely competent irrespective of the 'significance' criterion. 205 The BVerfG concluded with what could be seen as a red flag for the future: the SSM Regulation would have to be deemed ultra-vires if it had fully transferred bank supervision to the ECB. 206 Somewhat unexpectedly, the BVerfG also found that the L-Bank decisions of the CJEU were perfectly reconcilable with its views. To be sure, there were points of agreement between the courts. These concerned the role of the ECB in relation to significant entities 207 and the point that supervisory tasks which are outside the SSM Regulation were within national competence. 208 Moreover, there was agreement on the point that the ECB also had sole competence over a few tasks in respect of less significant credit institution (eg the authorization of a credit institution). 209 Recall that for these (few) tasks, 210 the significance criterion does not matter. However, on the crucial question of whether NCAs carry out their SSM tasks on the basis of national competence or on the basis of a decentralized implementation of an ECB exclusive competence, 211 the BVerfG's findings contrasted markedly with those of the CJEU. Yet, as pointed out, the BVerfG concluded that the CJEU's findings in L-Bank did not stand in its way. 212 It is plain that reaching this conclusion required some legal gymnastics. The BVerfG, for example, deemed the general findings on competence allocation in L-Bank to be of limited relevance to the ruling. 213 In the literature, Gentzsch and Brade talk in this context of obiter dicta by the CJEU. 214 Whether the CJEU's findings on the scope of the ECB's exclusive competence should be disregarded on this ground is, however, far from obvious. 215 At any rate, what is essential or not in the CJEU's jurisprudence is ultimately a question for the CJEU and not for the BVerfG to decide. 216 The BVerfG also interpreted the recitals of the SSM Regulation in an unusual way. Specifically, the BVerfG referred to recital (5) of the SSM Regulation to support its views on competence. 217 The first sentence of recital (5) states that '[c]ompetence for supervision of individual credit institutions in the Union remains mostly at national level'. However, upon reflection, it is doubtful that this recital actually supports the BVerfG's views on Member State competence. The recital in question is among a few recitals which essentially present the main considerations that prompted the Council to adopt the SSM Regulation. 218 It is submitted that the first sentence of recital (5), which the BVerfG highlighted, refers to the supervisory landscape prior to the establishment of the SSM. 219 In other words, all this sentence does is to underscore that hitherto banks were supervised nationally. It is essentially 'backward-looking', not 'forward-looking', and has as such little to offer in order to determine questions of competence under the SSM Regulation.
To sum up, according to the BVerfG, NCAs, when carrying out their SSM tasks in relation to 'less significant' entities, continue to do so on the basis of preexisting (national) competence and not as a result of a redelegation by the ECB to NCAs. In coming to this conclusion, the BVerfG relied both on Article 127(6) TFEU and on the SSM Regulation which according to the BVerfG supports a distribution of competences between the ECB and NCAs. Importantly, the fact that 216 See in this context M Jacob, Precedents and Case-Based Reasoning in the European Court of Justice: Unfinished Business (Cambridge: Cambridge University Press, 2014), 81 noting that the Court is 'quite happy' to seek support in statements that would be treated as obiter dicta according to common law doctrines. 217 BVerfG (n 2), para. 192. 218 After the first sentence, rec (5) goes on to say that '[c]oordination between supervisors is vital but the crisis has shown that mere coordination is not enough, in particular in the context of a single currency. In order to preserve financial stability in the Union and increase the positive effects of market integration on growth and welfare, integration of supervisory responsibilities should therefore be enhanced. This is particularly important to ensure a smooth and sound overview over an entire banking group and its overall health and would reduce the risk of different interpretations and contradictory decisions on the individual entity level'. 219 Note that the sentence in question appears to originate from the Commission Proposal for the SSM Regulation (COM (2012) 511 final), which further underscores that it refers to the supervisory landscape prior to the establishment of the SSM. See proposed rec (4) which states that '[c]ompetence for supervision of individual banks in the Union remains mostly at national level. This limits the effectiveness of supervision and the ability of supervisors to reach a common understanding of the soundness of the banking sector throughout the Union. In order to preserve and increase the positive effects of market integration on growth and welfare, integration of supervisory responsibilities should therefore be enhanced'. The reference to 'mostly' in the first sentence of recital (5) should be seen in light of the establishment of the European System of Financial Supervision (ESFS) and the European Supervisory Authorities (ESAs), as part of the ESFS, in 2011. Although the European Banking Authority (one of the ESAs) was not vested with day-today supervisory powers, it was given various tasks and powers which are relevant for the supervision of credit institutions at the national level (promoting convergence, coordination, and cooperation between NCAs, using intervention powers, etc.). the SSM Regulation vested the ECB with an oversight role and with a range of powers over NCAs, did not warrant a different conclusion. As far as the allocation of competence over less significant entities is concerned, the BVerfG accordingly interpreted the rules in a way that is broadly consistent with layering: national competence is not replaced, but 'new institutional layers'-in our case an ECB layer-'are added to it'. 220 As a result, the BVerfG was able to reach an outcome on competence that markedly contrasted with the findings of the CJEU.
V. Assessing institutional change: from contestation to reconciliation?
Having examined the decisions of the EU courts and the BVerfG, it is now possible to pull together the different parts of the argument. In Section III, I began by examining the changes that the SSM Regulation introduced along three key institutional dimensions: competence, powers, and governance. Rather than to insist on a single 'label' to describe these changes, I submitted that analytically the more interesting question concerned the extent to which the changes introduced by the SSM Regulation were left open to interpretation and therefore open to future change and contestation. I zoomed in on the ambiguities affecting the competence dimension-the most important institutional dimension-and examined the decisions of the CJEU and the BVerfG. I explained how the BVerfG's very own interpretation of EU law contrasted with that of the CJEU, and how as a result the decisions of these courts failed to agree on the scope of the ECB's exclusive competence under the SSM Regulation. I described this episode as one of change and contestation between courts: where the ambiguities and incompleteness inherent in the SSM rules enabled change to be amplified by the CJEU and to be contested by the BVerfG, through processes of 'conversion'. I showed how in this conceptualization of conversion, conversion and layering were linked: the outcomes to which ambiguous rules could be directed and redirected (that is, by amplifying legal integration or by affirming national competence) depended on whether or not, these rules were interpreted in a way that is consistent with layering. Hence, it was only by tracing change from the point of the enactment of the law to its interpretation by the courts that it was possible to gain a real appreciation of the dynamics and salience of change within the SSM. However, there is more. In the episode that I describe above, attempts at 'conversion' were not only directed at enacted rules. They extended to the interpretation of judicial decisions. Recall that according to the BVerfG, the L-Bank rulings did not stand in the way of its views on the allocation of competence in the SSM. As a result, the BVerfG's judgment was not-for the BVerfG at least-at 220 J van der Heijden, 'Institutional layering: a review of the use of the concept' (2011) 31 Politics, 9-18, 11. the origin of an open conflict between the German and EU courts. In this way, the BVerfG's judgment on the Banking Union is also different to the BVerfG's decision on the PSPP which I mentioned earlier. 221 Recall that in this latter judgment, the BVerfG found that both the ECB and the CJEU had acted ultra vires. It is common to view this judgment as the first where the BVerfG made good on its threat to openly oppose the CJEU or, as some commentators have put it, to not just 'bark' at the CJEU but to 'bite'. 222 In contrast, in the Banking Union judgment, the BVerfG pretended that all was well between courts. However, to come to this conclusion, the BVerfG read the L-Bank rulings in a way that trivialized their findings on the scope of the ECB's exclusive competence. The reality is that in the Banking Union judgment, the BVerfG disputed in a very real sense the scope of the ECB's exclusive competence, as 'discovered' by the CJEU in its L-Bank decisions. To use the 'bark or bite' metaphor, in the Banking Union judgment, which preceded the BVerfG's judgment on the PSPP, the BVerfG was not just 'all bark and no bite'. It bit; it is just that it pretended not to. 223 That being said, I am left in this section with a final point to consider: that is, the prospect of reconciliation following the decision of the BVerfG on the Banking Union. Generally speaking, two common mechanisms or strategies can be envisaged. The first is reconciliation by political agreement. In our context, it involves resolving ambiguities in the legislative text by way of legislative amendment. The second mechanism is reconciliation by way of judicial 'agreement'. It involves the preliminary reference procedure. I will explain why both are problematic and conclude by examining a third possibility, which I will call 'unilateral' reconciliation.
Starting with the strategy of legislative amendment, it is of course true that it is always open to the Council to amend the SSM Regulation in accordance with the requirements of Article 127(6) TFEU. Specifically, the Council could decide to offer clarification on the scope of the ECB's exclusive competence by amending the provisions of the SSM Regulation. However, if reconciliation is the aim, this strategy is an unlikely contender. Seeking clarification would almost certainly fail the unanimity hurdle of Article 127(6) TFEU. As noted earlier, the question of who supervises smaller credit institutions was a major issue for Germany during the negotiations of the SSM Regulation. As Merino notes: [i]t is certainly not by coincidence that Germany has been the country that most actively advocated for decentralisation of supervision, either during the negotiations at the Maastricht Treaty and at the moment of activating the ECB's supervision more than 20 years later. 224 By affirming the primary competence of NCAs over less significant entities, the BVerfG also affirmed the competence of the German financial supervisory authority (BaFin) over numerous smaller German credit institutions. Unsurprisingly, the decision of the BVerfG was welcomed by the German government which also noted in relation to bank supervision that it would continue to make sure that a significant part of the tasks and responsibilities remained with NCAs. 225 Hence, attempts to clarify the competence provisions of the SSM Regulation in a way that would benefit the ECB at the expense of NCAs would almost certainly be vetoed by Germany. Likewise, any amendment that diminished the role of the ECB would face the unanimity requirement of Article 127(6).
That said, even if salient amendments are likely to fail because of the unanimity hurdle, there are still other ways to make changes -for example, by taking advantage of the interplay between the SSM Regulation and other legislative acts. Thus according to Article 1 of the SSM Regulation, institutions referred to in Article 2(5) of the Capital Requirements Directive (CRD) are not subject to supervision by the ECB. It follows that by amending Article 2(5), it is possible to exclude institutions from the ECB's supervisory competence. The point is not just academic. Indeed, it is precisely because of a recent amendment to Article 2(5) that L-Bank is no longer subject to ECB supervision. As Smits put it, in the end, L-Bank was able to achieve 'through political means' what it could not achieve through judicial means. 226 A second strategy is formal judicial dialogue between courts. The formal method of judicial dialogue is the preliminary reference procedure. Under this procedure, national courts address questions about the interpretation or the validity of EU law to the CJEU. Gauweiler was the first ever preliminary reference of the BVerfG to the CJEU. 227 The CJEU was asked to decide on the legality of the ECB's Outright Monetary Transaction programme which authorized the predictive orientations, which would also require a more thorough assessment of the judicial behaviour of the CJEU and the BVerfG. 237 Suffice to say here that whatever incentives the CJEU may have had to adopt a conciliatory tone, prima facie, these incentives are likely to be significantly diminished following the BVerfG's PSPP judgment in which the BVerfG did not hold back in its criticism of the CJEU, 238 and ruled that the latter had acted ultra vires. 239 Arguably, in the wake of this decision, the CJEU is more likely than ever to wish to take a stand and affirm its role as the ultimate authority on European law. 240 That said, my aim here is not to make predictions or advise on what is advisable. It is simply to note-in broad and general terms-that lack of legal clarity can offer actors openings to seek reconciliation. Thus, if legal ambiguity or legal 'gaps' offer actors an interpretative space to engage in contests over the meaning, purpose, or function of rules, 241 it prima facie also offers an interpretative space to seek reconciliation. As noted earlier, the L-Bank rulings are not devoid of ambiguity-for example, think in this context of the meaning of a 'decentralized implementation', a notion that the EU courts in L-Bank largely failed to flesh out.
Last but not least, unilateral reconciliation may prima facie also occur because of the way in which rules are practised. To be sure, for legal scholarship, the study of 'formal institutions' in the sense of enacted rules is of obvious interest. In this article, I examined three such formal institutional dimensions: mainly competence, but also powers and governance. However, it is important to acknowledge that the way in which such 'institutions' are applied in practice can be important for how we think of change in each dimension. Hence, we must remain alert to the difference between 'law in action' and 'law on the books'. Thus, even if the ECB can rely on the authority of L-Bank to insist on its exclusive competence over the tasks in Article 4(1), the ECB may in practice show greater deference to NCAs when the latter supervise less significant entities. Of course, one could make a similar point with respect to BaFin, which might in practice act in a way that acknowledges the ECB's exclusive competence, notwithstanding the findings of the BVerfG on primary national competence. As noted earlier, the question of how rules are practised is one that is best addressed empirically. This leaves open an important space for future research.

VI. Conclusion
The aim of this article was to engage with the law of the SSM, with a view to contribute to the literature on institutional change. It contended that whilst the SSM was a story of change following an exogenous shock (ie the sovereign debt crisis), shedding light on judicial decision-making demonstrated that the SSM was also an account of change and contestation between courts, made possible by the ambiguity and incompleteness of the SSM rules. This ambiguity and incompleteness served as an interpretative 'battle ground' between courts, and ultimately enabled change to be amplified or to be contested through processes of rule 'conversion'. Hence, it was only by tracing change from the point of the enactment of the law to its interpretation by the courts that one could gain a real appreciation of the dynamics and salience of change within the SSM. To carry out the analysis, this article adopted a cross-disciplinary approach. It drew on literature in political science in order to gain a better understanding of the mechanics of institutional change, and adopted a legal lens in order to make sense of the legal intricacies of the SSM Regulation and the judicial rulings. Finally, this article reflected on the prospect of reconciliation. It showed why common strategies were out of reach. However, it did not dismiss the possibility of reconciliation, either because of the way in which rules are practised or following future judicial rulings. In this latter context, it stressed that legal ambiguity not only offered an interpretative space that enabled actors to contest change. Prima facie, it also offered a space that could enable reconciliation.
Before bringing this article to a close, it is worth repeating that the BVerfG does not have jurisdiction over an EU institution such as the ECB. Notwithstanding this, there are arguably good reasons for the ECB not to take a strictly hierarchical approach in its relations with NCAs when these supervise less significant entities. The SSM is a complex construct which, to function effectively, requires cooperation between the EU and the national level. This is especially so in relation to the supervision of less significant entities, where the ECB greatly depends on NCAs which are on the front line.
Lastly, it is also worth underscoring a final time that from an EU law perspective, it is the CJEU and not the BVerfG that decides questions about the interpretation and the validity of EU law. Nonetheless, there are reasons to be critical of both courts. As far as the BVerfG is concerned, it is the legal gymnastics that especially disappoint. In particular, the BVerfG's reading of the L-Bank decisions is open to criticism. Meanwhile, it is unlikely that the GC's decision will be remembered as a decision that was well reasoned. As noted, the GC was at times sloppy (eg with respect to subsidiarity). The gaps and ambiguities in the legal text offered the court significant leeway. Moreover, in ruling on the annulment action, the GC's broad-brush approach with respect to the scope of the ECB's exclusive competence arguably went beyond what was necessary to decide the case. The subsequent appeal before the CJ made no difference. None of this can be said to build trust between courts. And relationships, including relationships between courts, work better with trust.