Shared Tasks, but Separated Controls: Building the System of Control for Shared Administration in an EU Multi-Jurisdictional Setting

In order to address “wicked problems”, complex, multi-level governance structures must be established. These structures in turn require sophisticated systems of controls over public power to safeguard the rule of law. This seems to have been ignored in EU legislative practice and relevant research. This article argues that future research and legislative design of controls over public power in the EU need to be guided by the principle of connecting, aligning and making interplay between relevant concepts, institutions, procedures and scopes of different types of control belonging to the many jurisdictions, whose actors are involved in the executing of (shared) tasks in the EU. Connecting the disciplines that study these issues is a necessary prerequisite to this endeavour.


I. INTRODUCTION
In order to address "wicked problems", complex, multi-level governance structures and decision-making procedures must be established; the EU is a primary example of such a system. These complex, multi-level governance structures and procedures in turn require sophisticated systems of controls over public power to safeguard and uphold the rule of law. The sophistication comes at least from two perspectives: connecting various types of controls (political, judicial, financial audit, etc); and aligning the systems of controls belonging to different jurisdictions (EU-national, national-national) for the exercise of (shared) tasks by executive actors belonging to those different jurisdictions. Attention to creating a comprehensive system and framework on controls seems to have been lacking in EU legislative practice and research. The existing EU laws establishing shared tasks for EU and national authorities do not create watertight systems of controls. 1 The necessary connections between relevant concepts and types of controls to create a comprehensive framework of controls in the EU have not yet been found. Scholars from different disciplines, including law and public administration, have so far studied concepts, such as accountability and the principle of judicial protection, and the effects of Europeanisation upon them largely separately. The risk of this approach is in receiving only a limited pictureon one type of control and concept, not a "systemic" overviewand hence offering possibly "partial" solutions, rather than exploring potentially better options when, for instance, substituting different types or levels of controls with each other. This article argues that future research and legislative design of controls over public power in the EU need to be guided by the principle of connecting, aligning and facilitatiing interplay between institutions, procedures and scopes of different types of control belonging to the many jurisdictions, whose actors are involved in the executing of (shared) tasks in the EU. This is to prevent undesirable gaps and overlaps in controls and to design the necessary sophisticated system of controls in the EU. Connecting the disciplines that study these issues is a necessary prerequisite to this as well.
In recent decades, the proliferation of various (shared) tasks among EU and national authorities has been "squeezed" into the EU Treaty and constitutional arrangements that were created for a different reality. In that reality, built upon the classic international law aspirations, tasks and decision-making processes were strictly separated between the EU and national legal orders. The EU and national systems of controls over public powers could function separately. This is no longer the case, as we witness a rapid proliferation of various forms of shared structures such as joint supervisory and investigation teams and all kinds of mixed/hybrid decisions. 2 The proliferating shared structures in turn require the interconnection of the systems of controls belonging to different jurisdictions. In addition, no one formula for the sharing of tasks exists. Rather, these tasks have been shared vertically (EU-national) and horizontally or transnationally (nationalnational), including various configurations within these categories. Similarly, the interconnection of the systems and types of controls may need to be done in different ways. How should such a sophisticated system of controls be built in the EU?
This contribution is the first essential step in exploring this direction. Based on the analysis of the emerging research in this direction in several disciplines and a legal analysis of EU legislation and documents, it addresses two questions. What challenges does the multi-jurisdictional setting of the EU create for the system of control? What can be done to connect to study and design such a necessary, watertight and sophisticated system of controls in the EU? To address these questions, section II brings together the emerging landmark literature in the fields of accountability, the principle of effective judicial protection and the protection of fundamental rights in the EU. While the list of relevant concepts of controls is not exhaustive, this initial account shows where the connections can be sought. It is argued here that no typology seems to exist concerning the challenges to controls in the EU setting, though a number of challenges seem to have the same source: disconnection of the systems of controls between EU and national legal orders for shared tasks. A more comprehensive picture on the system of controls can be made by bringing together relevant concepts, types of controls and related analytical prisms used by scholars in different literature streams separately. This article distinguishes three prisms: institution-driven; decision-driven; and rights-driven. The added value of combining these three prisms for the sake of obtaining a more comprehensive picture on controls is illustrated by a concrete case of shared tasks in the EU, the Single Supervisory Mechanism (SSM) in section III. It is a representative case for the existing tasks in the population of all eight EU enforcement authorities. Section IV sums up and concludes. This article aims to raise awareness among academics and practitioners of the necessity of connecting the relevant concepts and aligning the systems of controls of different jurisdictions for (shared) tasks in the multijurisdictional EU to ensure control over the executive in the proliferating EU shared administration.

II. STARTING AN INVENTORY INTO THE EXISTING MULTIDISCIPLINARY LITERATURE
The EU complex system of governance is often considered to challenge the rule of law which ensures controls over executive public power and prevents arbitrary interferences with rights and freedoms of private actors. 3 The study of the rule of law by ensuring the control over the executive branch has been approached by scholars of different disciplines and sections of law via different concepts studied largely separately. 4 These include accountability, the protection of fundamental rights, the principle of judicial protection, liability and transparencythe list may not be exhaustive. 5 The concepts of judicial deference and independence should also be noted, as they influence the "quantity and quality" of controlling mechanisms, so that both the ideal of the rule of law and the need for effective operation of the executive can be maintained. In this section, I bring together the emerging landmark literature that I have dealt with in recent years, namely studies on the effects of Europeanisation on accountability, the principle of effective judicial protection and the protection of fundamental rights. The selection is thus "random" but it does allow a start to be made in the search for possible connections with a view to building the sophisticated system of controls in the EU.

Accountability in a multi-level setting
The concept of accountability seems to have become the central perspective in studying the effects of the multi-level setting of the EU on controls in public administration literature. 6 Accountability has been said to be an ex post part of control; the latter has also "ex ante" and "ongoing elements". 7 The "ex ante" elements include, for instance, procedures for the appointment of top officials of various executive institutions and legal frameworks regulating the exercise of delegated powers. 8 Ongoing controls could be exercised during parliamentary hearings, to influence the activities and decisions of executive actors on ongoing events. 9 Bovens has narrowed the definition of accountability down to "a relationship between an actor and a forum, in which the actor has an obligation to explain and to justify his or her conduct, the forum can pose questions and pass judgement, and the actor may face consequences". 10 Accountability for any public task is essential to ensure "the check of the governors by the governed". The definition has served as an assessment framework of three stagesinformation, discussion, rectificationemployed in many research projects. The approach has been used largely in relation to specific actors, for instance EU agencies and the Commission and institutional arrangements that those actors may have to render account before various accountability forums, such as the European Parliament and the Council and the Court of Justice of the EU. 11 Whereas the three stages model has offered a useful evaluative framework for the analysis for political accountability, its framework to assess judicial accountability, for example, has been somewhat limited and can be improved if the elements of the principle of effective judicial protection are added. 12 Judicial accountability requires a concerned party to trigger it, which necessitates elaboration of the "three stages" approach of Bovens. In this article, both terms will be used: control is an overarching term, including accountability as its ex post element.
Brandsma has identified three complexities of accountability in the multi-level setting of the EU. 13 First, individual actors taking part in complex decision-making procedures cannot render account for the content of a collective decision. Second, the multitude of actors can hinder accountability, because it allows the individual actor to evade accountability by pointing fingers at other actors. Third, the multitude of accountability forums with different agendas can hinder accountability. In his seminal article on accountability in a multi-level governance system, Papadopoulos expressed a number of concerns regarding the consequences of the multiplication of control mechanisms, in relation to both the efficiency and democratic quality of a given regulatory regime, including the problems of "too many eyes", information overload, and reduced transparency. 14 The disadvantages of the multiplicity of actors and forums include free-riding, which may reduce the collective capacity to exert control over the "agent". 15 At the same time, Scott and Mulgan have stressed the advantages of having overlapping accountability mechanisms, which create a "redundancy safeguard" against the absence of any control. 16 Furthermore, Scholten et al have observed a potential correlation between types of division of competences between EU and national authoritieshierarchical, parallel and supportiveand accountability frameworks necessary in these different types. 17 They suggest that this could serve as an analytical prism for assessing and designing accountability in a multi-jurisdictional system. In a hierarchical type division, one could expect the institution in charge (EU or national) to render account for the operation of shared tasks. In the case of the parallel division of competences (ie where EU and national authorities enforce EU law at the same time) like in the area of EU competition law, one could expect two systems of control to be run in parallel (EU and national). Concerning the supportive type division, where different tasks are being done by authorities of different jurisdictions, one could expect accountability to be organised separately for each and every task.
Interesting insights (for building an assessment-design framework) can also be gathered from studies focusing, roughly speaking, on the "quality", rather than "quantity" of accountability. Koppel distinguishes five conceptions of accountability: transparency, liability, controllability, responsibility and responsiveness. He has argued that "organizations are left to manage the many meanings of accountability", which have caused different evaluative standards and expectations. 18 This is the so-called "multiple accountabilities disorder" (MAD), ie situations where conflicting accountability expectations undermine the effectiveness of the organisation. Bundi sees a link between accountability and policy areas and argues that the need for accountability is higher in such policy areas, which are implemented by cooperative forms of governance. 19 The EU system of governance seems to be especially exposed 14 Papadopoulos, supra, note 6. 15  to such forms. This implies a stronger need for designing proper (and perhaps even more sophisticated) accountability mechanisms, because accountability forums from different legal orders may find such processes more difficult to oversee. 20 Blanc and Ottimofiore argue that the exercise of accountability is highly influenced by the determination of an agency's mandate and performance; "the best results in terms of effectiveness and accountability are found where mandates are more clearly linked to risks, risk management, specific outcomes in terms of public goods/welfare and more precisely and transparently defined, and the performance of the agency is designed and assessed on the basis of the goals/missions of the agency". 21 2. The principle of effective judicial protection and the protection of fundamental rights in a composite/shared/mixed/integrated administration Legal scholarship has studied controls in the EU setting, focusing predominantly on judicial controls, specifically via the principle of effective judicial protection and protection of fundamental rights 22 and the effects of the EU's "composite/shared/ mixed/integrated administration". 23 The concept of "accountability" seems to be used less often in this regard, and is used mostly in the context of political controls 24 or referred to as a principle. 25 The principle of effective judicial protection is about access to the court by concerned parties and the possibility for the court to review an act brought before it. The protection of fundamental rights focuses on affording legal protection to the parties who question a public authority's action. 20 See also Papadopoulos, supra, note 6. 21 F Blanc and G Ottimofiore, "The interplay of mandates and accountability in enforcement within the EU" in Scholten and Luchtman, supra, note 1, p 287. 22  One of the major conclusions of legal scholars has been that "judicial control of Europe's integrated administration faces : : : the dilution of responsibilities", which complicates the allocation of responsibility 26 of both actors and forums. In the mixed administration of the EU, a final legally-binding decision is often the result of composite decision-making, where (preliminary) drafts or actions could have been exercised by the actors from different legal orders. 27 This may create inconsistencies and gaps in judicial control. For example, in the Netherlands, the judicial review of an on-site inspection is excluded; the legality of inspection would be checked if a fine imposed on the inspected entity is challenged before the court. Under the new Single Supervisory Mechanism (SSM) regime, for instance, the fine is imposed by the European Central Bank (ECB) and hence may be challenged before the EU court. The arbitrariness of the inspection may thus stay unchecked, if the EU court does not do the check because the relevant legislation prescribes this task to the national court. 28 Furthermore, even if the draft or action have been checked, certain questions arise, such as whether it has been a relevant check (eg by a judicial rather than an executive authority), and whether the court reviewing the final decision can trust the review by an institution of another legal order and can recognise it; in other words the scope and standard of review may differ. 29 In addition, legal scholars have also touched upon the issue of how to control for inaction or omission. Here, the so far unanswered question has become about which type of controlpolitical or judicialwould be more appropriate. 30 Finally, EU criminal law scholarship has identified gaps in control over executive discretion in the case of transnational enforcement in the EU. 31 This concerns such issues as the choice of forum of where to investigate and sanction, and the interconnected negative effects stemming from the differences in national laws (eg differences in standards affording judicial protection and protection of fundamental rights). The consequence of this could be unbound, unchecked discretion of EU authorities that could choose to investigate in those jurisdictions that have the least checks, or that have lower standards of legal protection, with no obligation to justify the choice of jurisdiction.

Challenges in the multi-jurisdictional setting of the EU
The reviewed literature above has identified various challenges to controls (in the EU). These include pointing fingers to each other to avoid consequences, free riding, escaping of control due to too many eyes and hands, different standards of review or unbound executive discretion, reduced transparency and gaps in holding actors to account for collective decisions, "MAD", information overloads or shortages. What is important to note is that there are no classifications of these challenges. Yet, it seems that we deal with at least two partially overlapping phenomena: multi-levelness and multiplication of controlling mechanisms, actors and forums and the effects thereof. The latter may not necessarily be a distinctive feature of the former, as multiple and possibly redundant and/or overlapping controlling mechanisms can also exist in a "one-level system". Further research needs to bring conceptual clarity, in order to analyse which challenges can be interrelated with the "multi-levelness" and complex governance systems (and perhaps upon what conditions) and which cannot. Interestingly, those challenges that can be interrelated with the "multi-levelness" of "shared administration", such as different standards of review for the same/shared tasks or actions, seem to have the same cause. It is disconnection between the systems of control belonging to different jurisdictions (legal orders). It is important to establish this, in order to guide the focus in studies on controls in the EU setting and when assessing and designing them: knowing the cause of a problem facilitates finding an appropriate solution.
As a side remark, a multi-jurisdictional setting of the EU may be a better way of describing the EU's complex governance system, in order to focus on the challenges of control there, ie the lack of connection between the systems of control belonging to different jurisdictions. It allows for reflection on both the vertical (EU-national) and the horizontal (national-national or transnational) dimensions of these disconnections. The terms used in the legal scholarship (mixed or integrated administration) are more nuanced than the term "multi-level". Yet, from the perspective of identifying the source of controls challenges, they seem to be not as nuanced as the term "multi-jurisdictional". It is not the mixing or integration that causes the challenges in controls, but the presence of multiple jurisdictions (legal orders) whose systems of control are not connected or not aligned to control the "mixed" or "integrated" tasks. Further debate is certainly invited on this point ("is it all in the term and which term is more suitable?").

Useful connections
Researching and subsequently designing controls over public power in the EU via individual concepts has limits, which can be (at least in part) overcome by making connections between: (i) relevant concepts; (ii) the types of controls they may represent; and (iii) the analytical prisms that they create. One example mentioned above concerned making Boven's accountability framework more suitable for an investigation of judicial accountability, by adding the elements of the principle of effective judicial protection. The specifics of the judicial type of accountability make this essential for the Boven's three stages to be operationalised effectively. Furthermore, different decisions, actions and inactions may require different types of controls and different EU jurisdictions may opt for different solutions, which could be a hurdle when aligning the systems of control and creating relevant cooperation. For instance, while some Member States use administrative sanctions for violation of EU environmental or financial laws, others introduce criminal sanctions. Any cooperation between relevant authorities on such issues as exchange of information among them will have different implications for controls, especially legal protection standards that need to be ensured in EU legislation. 32 I will focus on the third issue, ie the connection of analytical prisms, in more detail. Three prisms can be tentatively distinguished based on the investigated three concepts and literature streams. To obtain a more complete picture on the issue of controls, bringing these perspectives together is essential. These are: • an institution-driven prism: to what extent have relevant institutions been accountable and via what accountability mechanisms?
• a decision-driven prism: to what extent are there political, judicial or other checks upon a specific (preliminary) decision/action? and • a rights-driven prism: to what extent is legal protection of the affected parties ensured?
Similarly to the observations of Koppel on different conceptions of accountability, here too these three prisms taken in relation to the same case could bring different results in terms of having controls or not, and the extent to which an answer to the question has been comprehensive. Roughly speaking, an agency can be generally held to account via reporting and other mechanisms, and each of its (preliminary) decisions can be available to be checked, including by a forum from another jurisdiction if necessary. In addition, while the judicial review of its decision can be possible in theory, the effectiveness of this mechanism of control may be limited due to, for instance, strict standing requirements at a court, and negative implications from the differences between laws governing legal protection in different jurisdictions. Further research is essential to establish conceptual connections, types of controls and the relevant analytical frameworks comprising different prisms. Which exact concepts and their elements should an overarching framework consist of and what material needs to be collected to apply it? For now, I will illustrate the usefulness of looking through the three prisms and different concepts in the following section, using some of the existing operationalising elements. Political accountability (accountability before representative organs), as defined by Bovens, involves the three stages: information (eg reporting to parliaments), discussion (eg parliamentary hearing) and sanctioning (eg lessening of a budget). Judicial accountability (accountability before the courts) is interconnected with the elements and rights of the principle of effective judicial protection, ie the right to have access to the court by the parties concerned and the possibility for the court to review an act brought before it. 33 Depending on the specifics of a regulated area, the "rights-driven" prism can include the right to have effective judicial protection and/or rights depending on the actions performed and decisions taken, eg participation and consultation for rule-making procedures and defence rights for enforcement actions.

III. CONTROLS IN THE SINGLE SUPERVISORY MECHANISMS
The added value of bringing the discussed concepts and prisms to an analysis of controls in a system of shared tasks is the possibility of obtaining a more comprehensive picture of this issue; a picture one would not have if using only one of the concepts or prisms.
To start with the illustration selection, the rise of shared administration has been witnessed in all kinds of tasks that can be roughly distinguished as regulatory and enforcement. I take an example of Single Supervisory Mechanism (SSM) here, which is a representative case for the whole new population of EU authorities with direct enforcement powers. 34 The EU direct shared enforcement has been an emerging, yet rapidly proliferating field of truly shared tasks and decision-making procedures between EU and national actors (and among the latter). It has spread through eight policy areas: micro prudential supervision, supervision of financial markets, competition law, aviation safety, pharmaceuticals, fisheries policy, food law, and fight against EU financial fraud. Eight EU Enforcement Authorities (EEAs) exist already in these fields: the European Central Bank (ECB), the European Securities and Markets Authority, DG Competition of the European Commission, the European Aviation Safety Agency, the European Medicines Agency, the European Fisheries Control Agency, Directorate-F of the European Commission, and the Anti-Fraud Office of the European Union. The European Public Prosecutor's Office has recently been established, but is not yet operational. These EEAs enjoy all or some powers of direct enforcement, ie monitoring compliance with EU law by private actors, investigating alleged breaches of law and sanctioning for non-compliance. 35 None of the EEAs replaces relevant national counterparts. On the contrary, they work together, which creates a system of shared tasks in a multi-jurisdictional cooperation. Three types of sharing direct enforcement tasks have been distinguished: hierarchical; parallel; and supportive. 36 The case of micro-prudential supervision by the ECB and NCA is representative in this light and in terms of challenges in control that can arise in this area of shared tasks. 37 The ECB is, in accordance with Article 4 of the so-called SSM Regulation, exclusively competent to carry out, for prudential supervisory purposes, the tasks laid down in that article; NCAs are responsible for assisting the ECB. 39 The tasks include authorisation of banks and withdrawal of authorisations, and supervision of the compliance by banks with the relevant EU law. The ECB can issue (policy) guidelines, recommendations, instructions to NCAs, decisions to launch an inspection and investigation of a suspicious violation of law and (sanctioning) decisions against banks. Sanctions against natural persons can only be imposed by the NCAs, perhaps also upon the instructions of the ECB. 40 The supervision and enforcement imply that the ECB's joint supervisory teams (JSTs), comprised of EU and national officials, have the power to monitor compliance with relevant laws, to investigate suspected cases of violation of laws by ECB teams (which can comprise national officials), and impose sanctions by the ECB, which can be measures (imposed only during the monitoring stage) and administrative pecuniary penalties and fines. These powers also include the ability to issue formal decisions, request information from banks, ask questions and make on-site inspections. The ECB can request that NCAs undertake actions, such as making on-site visits and other activities, when the ECB lacks the power to do so, but the NCA does enjoy that power in accordance with the national law.
The ECB and NCAs, roughly speaking, divide the task of supervision based on the significance of the bank: significant institutions (SIs) are to be supervised by the ECB; less significant institutions (LSIs) by the NCAs. The criteria to determine significant and less significant banks are defined by the ECB in accordance with the criteria laid down in the SSM Regulation, and the ECB can take over the supervision of an LSI if necessary. As the first judgment in the field of SSM by the General Court (T-122/15) shows (at the moment of writing, the judgment is under appeal), the ECB has discretion in interpreting those conditions. 41

Analysing controls within the SSM through the three prisms and different concepts
The ECB is the leading authority for the smooth functioning of the SSM system; NCAs assist the ECB. From an institution-driven perspective, these institutions are under respective accountability obligations in their jurisdictions. Problems can arise when it comes to holding them to account for shared tasks. More specifically, politically, the ECB can be limited in rendering account for the functioning of the whole system if it cannot control fully its "assistants". The SSM Regulation sets up a framework concerning political accountability; it is supplemented by an Interinstitutional Agreement (OJ 2013 L 320/2) with the European Parliament and a Memorandum of Understanding with the European Council (MoU) for political accountability. The ECB is accountable largely before the European Parliament and the Council for the implementation of the SSM Regulation (Council Regulation (EU) No 1024/2013) via annual reporting, hearings and (confidential) discussions specially designed for the supervisory mechanism (Article 20), though national parliaments also have some "controlling" powers. National parliaments may address their reasoned observations on the annual report to the ECB, may request the ECB to reply in writing to any observations or questions submitted by them to the ECB in respect of the tasks of the ECB under the regulation, and may invite the Chair or a member of the Supervisory Board to participate in an exchange of views in relation to the supervision of credit institutions in that Member State together with a representative of the national competent authority (Article 21). Furthermore, personal accountability of the Chair and Vice-Chair of the ECB Supervisory Board is envisaged via the possibility to approve the candidate, after the public hearing, by a relevant committee of the European Parliament. The Council then takes the formal decision to appoint as well as to remove these officials, preceded by a proposal from the ECB and approval by the Parliament. The grounds for removal include poor performance and serious misconduct; in terms of the US administrative law, this is a "removal for cause" clause, to ensure independence from politics.
NCAs are accountable to national parliaments in accordance with national law for the performance of tasks not conferred on the ECB by the SSM Regulation and for the performance of activities carried out by them in accordance with Article 6. Duijkersloot et al have analysed the Dutch and Greek systems in this respect. In the Netherlands, for instance, the accountability of the Dutch Central Bank is organised via the Minister of Finance, who has though a limited responsibility for the independent DNB (the so-called "responsibility at a distance"). The personal accountability is organised via removal by the Royal Decree (government) but also "for cause". In Greece, the Bank of Greece has a more direct accountability relationship with the national parliament via annual reporting. The top-level officials are appointed by the executive branch. 42 Whereas respective institutions have relevant accountability mechanisms in their own jurisdictions, it is more challenging to apportion accountability for their shared tasks, because of the habit of blaming other parties and pointing fingers at each other. The division of competences concerning the establishing of the "supervisory policy" between the ECB and NCAs has been designed in a hierarchical way. The ECB is in charge and can render account for that before the relevant forums. Yet, an internal mechanism of control for the ECB against the NCAs is weak, if existent, 43 which makes it difficult for the ECB to take responsibility for the operation of the whole system. Here, it is useful to add a decision-driven perspective by addressing the question for what the political accountability is to be rendered. The ECB may issue guidelines and arguably individual instructions vis-à-vis NCAs. However, what if an NCA does not comply? The ECB has no sanction vis-à-vis the non-complying NCA. This could result in a situation where the ECB could point fingers at NCAs for a 42 Ibid; Ter Kuile et al, supra, note 24. 43 Karagianni and Scholten, supra, note 24. failure before the EU political accountability forums and an NCA could do the same before its national parliament. The accountability design of the system does not connect the systems of controls belonging to different jurisdictions, which could prevent pointing fingers. Here, the recent development in the case of Europol, ie the creation of a joint parliamentary scrutiny committee of representatives of the EU and national parliaments, is worth mentioning and is interesting to consider in future research. Furthermore, it is unclear what legal status an ECB's instruction vis-à-vis an NCA has and hence which type of controljudicial and/or political -ought to be established for it and at what level. For instance, a policy instruction seems to require political accountability, whereas a legal decision-type instruction would require judicial control. Neither of these are prescribed at this moment.
Given the importance of financial accountability in the field of political control, it is important to mention that the ECB's financial accountability is organised via external audit. The SSM is funded by supervisory fees levied from the supervised banks. Still, the courts of auditors can have powers to check the independent supervisors, though the Bank of Greece is exempted from accountability before the Greek Court of Auditors. The European Court of Auditors can examine the operational efficiency of the ECB management. 44 The Dutch Chamber of Auditors holds the DNB to financial account, which includes the right to obtain access to confidential information from the DNB. And these differences in mandates between the courts of auditors between different jurisdictions hint at another shortcoming of the system of financial and political accountability caused by the absence of connection between the systems of control belonging to different jurisdictions. Should the European Court of Auditors want to cooperate with its national counterparts (or the latter among themselves) in the respective field of shared enforcement, this could be difficult to arrange given the differences in their powers and remits. Boers et al show in this respect that 7 of 13 national courts of auditors that they investigated have no mandate to check the financial supervisors in their jurisdictions, and those who do have that mandate may lack the powers to request all the necessary information. 45 In this light, when designing respective controlling systems for shared (enforcement) tasks in the EU, it would be advisable to connect, align and ensure interplay between relevant controlling institutions. This could be done, for instance, by making relevant EU legislation prescribing similar tasks and scope of review for accountability forums, including EU and national court of auditors.
With respect to judicial accountability, from an institutional perspective, decisions of the ECB are generally judicially reviewable by the EU courts. Decisions taken by the NCAs are generally reviewable at the national level. As stated above, however, judicial accountability requires a concerned party to trigger it; the courts do not review enforcement decisions in abstracto. Therefore, it is important to consider the elements of the principle of effective judicial protection: access to the EU judicial accountability forums is possible only if the tight Plaumann criteria are met. Furthermore, national courts can also be involved in checking the arbitrariness of an inspection by the ECB if national law so requires. This system works when it comes to a parallel division of competences, eg the division of supervision of significant and less significant banks between the ECB and NCAs. The relevant supervisory decisions can be thus challenged at the court of the respective jurisdiction. The situation becomes complicated when it comes to a more "supportive" type of division of competences, such as when NCAs make preparatory acts for the final decision to be taken by the ECB. How should accountability for such support be arranged, when the ECB takes final decisions and renders account for those decisions at the EU level? Can the EU court obtain access to preliminary decisions or the actors involved when it reviews the final decisions? These questions require both decision-and rights-prisms for analysis.
Prechal et al show that this may not always be possible, which implies a gap in control. 46 They describe a bank's authorisation application procedure within the SSM. The NCA prepares a draft decision and sends it to the ECB; the latter can support it or object to it. As the authorisation decision of the NCA can include conditions and restrictions, also based on national law, if the ECB objects, then the EU court may be confronted by the necessity to interpret national (procedural) rules. Because the preliminary reference procedure is only organised in a one-way direction (national courts can ask the Court of Justice about the interpretation of EU laws, but the Court of Justice cannot ask national courts about the interpretation of national laws), the legitimacy of the Court of Justice's interpretation of national laws can be questioned. Also, the question arises of whether the appropriate institution is held to account and at the appropriate level: the NCA prepares an act but the ECB is rendering the account for it (and the process by which the preliminary act was made) before the EU court if the bank challenges the objection decision of the ECB. This situation can be avoided by connecting the judicial systems of control to make it possible for the Court of Justice to ask "preliminary" questions on interpretation at relevant national courts. 47 The possibility of employing such an institutional solution is unclear. At this moment, the Court of Justice has been trying to divide the jurisdiction between the EU and national courts to review different types of acts in composite procedures. 48 Questions about judicial accountability for shared tasks within the SSM arise also when it comes to inspections of bank premises. This concerns questions of legal protection, and is also connected with the rights-driven prism. The JSTs, comprised of ECB and national officials, as well as enforcement and investigation teams from the ECB, can inspect banks. As banks can have premises in different countries, supervisors are able to select a jurisdiction in which to make the inspection. The choice of which jurisdiction should investigate a certain case can have implications for the level of legal protection, appeal possibilities and procedural guarantees that companies may have in one or another jurisdiction. 49 Furthermore, some jurisdictions like the UK prescribe the necessity of a national judicial check before an inspection of a private actor can take place, to afford legal protection to private actors against arbitrary interferences by the executive, while other jurisdictions do not request this. As in the area of EU competition law, 50 there seems to be no justification for making the choice of jurisdiction in the area of SSM. Private actors have no possibility to question this choice by the supervisor, to ask whether the choice has been made for "good reasons", and whether the supervisors have "abused" their discretion of choice by selecting a jurisdiction which offers less protection than others. For instance, in the Netherlands, one can bring an action to the court to question an inspection by the national supervisor only when the fine is imposed upon the private actor, 51 which may lead to the choice by the supervisor to inspect in the Netherlands, rather than in other jurisdictions, like the UK, with stricter rules. The problem here is twofold. On the one hand, there is no alignment between national systems of controls. This leads to the possibility of "playing with" existing diversity in legal protection, which affects the position and rights of the private actor who may not be able to check the exercise of the public power and can thus be deprived of some rights due to the choice made by the supervisor. On the other hand, as legal diversity in the EU is largely allowed upon the principle of national institutional autonomy, clear legislative language needs to specify how the discretion needs to be accounted for, including via political and/or judicial routes; where and how, including possible necessary connections between EU and national and among national systems of controls.

The added value of connecting
Combining the prisms used by scholars from different disciplines and using several concepts enriches the analysis of the system of controls over public power. This method of analysis could facilitate the addressing of specific challenges by seeking out links and connections with other concepts, types and systems of controls. The question of the extent to which control over the executive for shared supervisory and enforcement tasks within the SSM is ensured can clearly be answered more completely if there are institutional and process-related accounts. The concept of accountability offers an institutional perspective on controls. One looks at the mechanisms that the ECB, as an organisational entity, has towards relevant forums. The legal concepts of the principle of effective judicial protection and protection of fundamental rights focus more on the legality check of actions, decisions and relevant processes. Furthermore, combining concepts can improve the existing analytical prisms they employ. The analysis of judicial accountability according to Boven's stages is better off by adding the necessary procedural (such as access to the court) and content (legal protection against arbitrary interferences by the executive) elements. Finally, combining types of controls in one studypolitical and judicialprompts the questions of which type of control is better for which actions and inactions. The connection between political and judicial (and subsequently other) types of controls is certainly worth further exploration in, for instance, such cases as the limits of judicial review due to judicial deference and the validity review shortcomings under Article 267 TFEU 52 and the limits of political accountability due to possible independence clauses.

IV. SUMMING UP AND CONCLUDING
Complex governance structures like the EU and their "shared tasks" require sophisticated systems of controls over public power to safeguard the rule of law: this seems to have been ignored in the EU's legislative practice and has hardly been investigated. This article has called for the need to adjust the "legislative formula" when designing controls for shared (enforcement) tasks, and to boost multi-disciplinary research to address this "omission", leading to the challenges of safeguarding the rule of law in the EU. How should the necessary sophisticated system of controls for the EU be created?
This article suggests that the answer must be found through the creation of a framework in which a number of elements need to be connected. These include relevant concepts of control, types of controls that these concepts may represent, analytical prisms that they may create, and systems of controls belonging to different jurisdictions in the EU (EU-national and national-national). Further research should explore the questions of which concepts of control need to be included, how they should be effectively connected, and asking what could strengthen their individual assessment frameworks, and help to build one comprehensive framework for research and design. Future research and the EU legislative practice specifically should pay attention to connecting controlling institutions, the scope of their powers, procedural aspects and relevant standards of protection and review between and among EU and national jurisdictions for the (shared) tasks created by EU laws. Since the treaty revisions do not seem politically feasible, one could think about other "horizontal" solutions, for example mechanisms and laws offering possibilities to create ad hoc tribunals and parliamentary meetings to deal with transnational cases and tasks and the novel challenges they bring. That a public authority in the EU may have various accountability obligations towards relevant controlling institutions of its jurisdiction does not necessarily mean that its specific decisions or actions resulting from a shared task exercised with other EU and/or national authorities are under control.