1.Introduction

Administrative appealsThe present note will use the term “administrative appeal” to describe the phenomenon. Terminology is inconsistent and commonly used alternative terms can be found throughout the text and footnotes. are in fashion. The idea that addressees of an administrative act (or other third parties meeting standing requirements) should first bring their grievances not before a court but before the authority that issued the act or another hierarchically superior authority, is attractive to judges and legislators alike.Both mandatory administrative appeals (whose prior conclusion is a precondition for subsequent court action) and facultative ones (where the applicant can choose between the administrative avenue and direct litigation) are within the purview of this paper; the distinction will become of relevance later in the text. It offers benefits for the court system in terms of reduced caseload and for the remedy-seeking public by avoiding some of the disadvantages of litigation, such as delays, high expenses, and publicity of proceedings. As a consequence, the concept can be found in many jurisdictions.Examples include the recours administratif in France, the Widerspruch in Germany, administrative review and pre-action protocols in the United Kingdom, and adjudication in the United States. The latter two, in particular, share similarities with the situation at the EU level in the sense that their systems for extrajudicial dispute resolution in administrative matters are fragmented among different areas of law, rather than providing for an overarching concept of administrative appeals across the entire legal system. See Philip J Harter, ‘Dispute Resolution and Administrative Law: The History, Needs, and Future of a Complex Relationship’ (1984) 29 VillanovaLR 1393 for the American historical development. For a comparative perspective, see Dacian Dragos and Bogdana Neamtu (eds), Alternative Dispute Resolution in European Administrative Law (Springer 2014) and Kieran Bradley, ‘Tribunals and Adjudication’ in P Cane et al. (eds), The Oxford Handbook of Comparative Administrative Law (OUP 2020). It has also taken root at the European level: EU law has no overarching concept of administrative review across the entirety of Union administration, but the founding acts of many agencies establish a tailor-made procedure, usually before a review body which is internal to the agency but composed of independent experts. The example with the greatest practical relevance is, quantitatively, the Board of Appeal of the Intellectual Property Office in Alicante,With routinely more than 2,000 review procedures per year; see Regulation (EU) 2017/1001 of the European Parliament and of the Council of 14 June 2017 on the European Union trade mark [2017] OJ L154/1, art 66-73 (‘Trade Mark Regulation’). Please note that the Regulation uses “Board of Appeal” in the singular whereas the Office’s publications tend to speak in the plural, distinguishing between the plenary Board of nine members and the three-member panels hearing individual cases. but similar structures can be observed in many agencies. This justifies the scholarly attention dedicated to the phenomenon.See, in lieu of many others: Luca Bolzonello, ‘Independent Administrative Review Within the Structure of Remedies under the Treaties’ (2016) 22 EPL 569; Paola Chirulli and Luca De Lucia, ‘Specialised adjudication in EU administrative law’ (2015) 40 ELR 832; and Dacian Dragos and David Marrani, ‘Administrative Appeals in Comparative European Administrative Law: What Effectiveness?’ in Dacian Dragos and Bogdana Neamtu (eds), Alternative Dispute Resolution in European Administrative Law (Springer 2014). Such review bodies are, despite their independence in the scrutiny they undertake, considered part of the agency whose acts they review, not an impartial tribunal;Case T-63/01 Procter & Gamble v OHIM [2002] EU:T:2002:317, paras 20-23. nonetheless, their reviews bear similarities to litigation before the Union courts, as applicants need to demonstrate standing against the decision they contestThe requirements for standing often borrow the “direct and individual concern” wording from Article 263(4) TFEU. Even where the statutory text governing the procedure does not use this primary law wording, the practice of the review bodies may well draw upon this case law, especially the seasoned Plaumann formula. See, for instance, the decision of the Second Board of Appeal of OHIM of 18 February 2011 in Case R 197/2010-2, para 32. and as the review bodies tend to follow the principle of stare decisis.Not in the sense of applying prior precedent as a formal source of law, but in the sense of tending to follow them as a matter of practice in most cases. For a doctrinal discussion see E Marchisio ‘Are EUIPO Trade Mark Examiners and the Boards of Appeal Bound by Precedent Decisions?’ (2020) 69 GRUR Int 893 (also emphasising the status of the Boards of Appeal as part of the Office).

Such a procedure was also introduced for the activity of the ECB within the Single Supervisory Mechanism (SSM). In this mechanism, established in November 2014, the ECB has assumed powers to adopt supervisory decisions addressed to credit institutions.Council Regulation (EU) No 1024/2013 of 15 October 2013 conferring specific tasks on the European Central Bank concerning policies relating to the prudential supervision of credit institutions [2014], OJ L 287/63 (‘SSM Regulation’). For the about 100 “significant” banking groups in the Banking Union, these direct ECB powers extend across the entire range of prudential banking supervision, whereas for the other, “less significant” banks, the ECB is directly competent only for a narrower range of tasks while the bulk of day-to-day prudential supervision is carried out by the national competent authorities (NCAs). The tasks which remain with the ECB for all credit institutions, also the less significant ones, include licensing: The NCAs’ pre-SSM power to autonomously withdraw banking licences was replaced by a power to propose such a withdrawal to the ECB (Article 14(5) of the SSM Regulation). The ECB is not bound by an NCA proposal but rather undertakes its own assessment of the case; if the ECB decides to withdraw, it uses, as a legal basis for such a decision, the transposition of Article 18 of Directive 2013/36/EU, where the grounds for withdrawals of banking licences are harmonised.From the perspective of EU law scholarship, the procedure, therefore, takes the form of a “composite procedure” whereby a national authority prepares a proposal but the ultimate decision is one of a Union institution; for the phenomenon more generally see Filipe Brito Bastos, ‘Judicial review of composite administrative procedures in the Single Supervisory Mechanism: Berlusconi’ (2019) 56 CMLRev 1355; Filipe Brito Bastos ‘An Administrative Crack in the EU’s Rule of Law: Composite Decision-making and Nonjusticiable National Law’ (2020) 16 ECLR 63; and Francisco H Fernández, ‘The Application of National Law and Composite Procedures in the Single Supervisory Mechanism’ (2021) 14 REALaw 5. Case law has emphasised the unitary nature of such composite procedures and excluded the separate judicial reviewability of the national stage before national administrative courts: Case C-219/17 Silvio Berlusconi and Fininvest SpA v Banca d’Italia and IVASS [2018] EU:C:2018:1023. These grounds also include serious breaches of the national provisions adopted pursuant to the Union’s Anti-Money Laundering Directives (Article 18(f) in conjunction with Article 67(1)(o) of Directive 2013/36/EU). The “bridge” which makes national statutes transposing Union directives available for application by the ECB is built by Article 4(3) of the SSM Regulation.

Article 24 of the SSM Regulation provides for a facultative administrative review of ECB decisions (including, in the absence of any indication to the contrary, licence withdrawal decisions) before an Administrative Board of Review (ABoR) that is internal to the ECB but composed of independent experts. In the recent judgment Versobank, the General Court had an opportunity to clarify a question concerning the interplay of such administrative reviews with the decision-making powers of the ECB, and in particular the legal fate of the initial decision that was under review after it has been replaced as a result of said review.Joined Cases T-351/18 and T-584/18 Ukrselhosprom PCF LLC and Versobank AS v ECB [2021] EU:T:2021:669 (rendered on 6 October 2021). The ruling is best referred to collectively as Versobank as this is the entity at the heart of the matter. An appeal against the first instance ruling to the Court of Justice is pending as Case C-803/21 P. – The same logic as in Versobank was subsequently applied by the General Court in the factually similar Cases T-247/16 RENV Trasta Komercbanka AS v ECB (order of 17 November 2021; ECLI not yet assigned at the time of writing) and T-321/17 Heikki Niemelä and Others v ECB [2021] EU:T:2021:942. Since these rulings followed the precedent set in Versobank (albeit without explicitly citing it), the present note will focus on the latter. Even though the case hinged, in part, on an institutional peculiarity of the SSM, the Court’s logic is of wider doctrinal interest for European administrative law.

2.Summary of the case and the ruling of the Court

The second applicant, Versobank AS, was a company incorporated under Estonian law and holder of a banking licence that had been granted to it by the Estonian authorities in 1999; the first applicant, Ukrselhosprom PCF LLC, was a Ukrainian company that held more than 85% of the shares in Versobank. As a less significant institution within the meaning of the SSM Regulation, Versobank was supervised by the Financial Services Authority (FSA, also known under its Estonian name Finantsinspektsioon), the national competent authority (NCA) of Estonia in the SSM. This included both prudential matters and compliance with obligations for anti-money laundering (AML) and combating the financing of terrorism (CFT); these fields are beyond the competences of the ECB but, in many Member States, combined with prudential supervision within a single NCA.Recital 28 of the SSM Regulation.

Since 2015, the FSA has identified, via ongoing supervision as well as several on-site inspections, recurring breaches by Versobank in connection with the ineffectiveness of its AML/CFT regime and the inadequacy of its AML/CFT governance arrangements.Versobank (footnote 11) paras 4, 12, and 25. They led the FSA to adopt a “precept” in 2016, which is the Estonian law equivalent of a decision (within the meaning of Article 288(4) TFEU), i.e. a case-specific legal act imposing a binding and enforceable obligation on its addressee.“Ettekirjutus” in the original Estonian. The precept required Versobank to remedy its AML/CFT shortcomings and to provide evidence of this remediation to the FSA, but subsequent inspections found that this has not been complied with. In 2018, this motivated the FSA to submit a proposal to the ECB to adopt a decision withdrawing Versobank’s banking licence on several grounds, the most important of which were Versobank’s lack of necessary governance arrangements and an effective AML/CFT regime and non-compliance with the precept. The ECB initiated the European stage of this withdrawal procedure, assessed the proposal, and – following a hearing of Versobank – adopted a decision withdrawing its banking licence. This decision was contested through a notice of review submitted to ABoR, initiating the administrative review foreseen in the SSM Regulation.The identity of the applicants who brought this notice of review will be discussed in section 3.4 of the present case note; it constitutes a minor but still interesting side aspect of the judgment. ABoR concluded that the substantive and procedural infringements alleged by the notice of review were unfounded and that the ECB should adopt a second decision identical in content to the initial decision that had been contested; the ECB acted accordingly. The latter part – the replacement of the first decision with the second decision of identical content – might sound surprising at first, but it is in line with what Article 24(7) of the SSM Regulation mandates. It will become clear in the course of the present discussion that the adoption of a second decision, even where it is identical to the first, is of great importance to the establishment of a functioning system of remedies.

Since the review was therefore unsuccessful from the applicants’ perspective, an action for annulment against the two ECB decisions was subsequently brought before the General Court – necessarily on different dates, as the two-month deadline of Article 263(6) TFEU was counted separately for each decision as of its respective notification. The cases were logged as T-351/18 and T-584/18 on the Court’s docket and subsequently joined. Aside from this joinder and costs, the Court’s conclusions can be summarised in two main points: (i) T-351/18, the action for annulment against the first decision, was devoid of purpose since that decision had been withdrawn by virtue of the second one, and there was consequently no need to adjudicate on this action; (ii), T-584/18, the action for annulment against the second decision, was inadmissible as far as Ukrselhosprom was concerned since Ukrselhosprom was not the addressee of the contested decision but only a shareholder of the addressee, and this was insufficient to regard it as directly and individually concerned by the decision; as far as Versobank, the addressee itself, was concerned, the action was admissible but unfounded in its merits since the Court found the decision to be procedurally and substantively lawful.

The judgment offers interesting elaborations on a wide range of questions, such as the mutual delimitation of the FSA’s and the ECB’s competences especially concerning AML/CFT matters, their interaction in what doctrine has come to dub “composite procedures”,See, in particular, the literature cited in footnote 10. and the standards against which ECB decisions are scrutinised in terms of the statement of reasons and the analysis of proportionality. The present case note, however, will not attempt a comprehensive annotation of these diverse matters but take a more modest approach. It will focus on the Court’s elaborations on the subject matter of judicial review following a previous administrative review, i.e., point (i) of the two main points listed above.

3.Judicial review after prior administrative review by ABoR

3.1.Peculiarities of the ABoR procedure and ensuing problem

In comparison to other administrative appeals elsewhere, the ABoR procedure under the SSM Regulation displays noteworthy differences which are, ultimately, consequences of the ECB’s status as a Union institution established by the Treaties. One important difference lies in the facultative character of the procedure: Remedy-seeking applicants can choose to forego review by ABoR entirely and proceed directly to the Court of Justice.Article 24(11) SSM Regulation. This stands in contrast to the review bodies of many EU agencies whereby applicants must exhaust administrative review before an action for annulment before the Court of Justice would be admissible. The reason for this difference arises from Article 263(5) TFEU, according to which acts setting up bodies, offices and agencies of the Union can lay down additional conditions for actions against acts of these bodies. This provision does not apply to primary law institutions, which means it was not possible for the Council, as legislator of the SSM Regulation, to establish the exhaustion of administrative review as a precondition for litigation.

Another even more important difference arises from the primary law definition of the Executive Board and the Governing Council as the decision-making bodies of the ECB.Consolidated version of the Treaty on the Functioning of the European Union OJ C326/47 (‘TFEU’), Article 129; and Protocol (no 4) on the Statute of the European System of Central Banks and of the European Central Bank OJ C202/230, 9.3. The SSM Regulation could, therefore, not establish ABoR as such a body; instead, ABoR can merely issue a non-binding opinion, following which the Supervisory Board – an SSM body which prepares complete drafts of supervisory decisions for adoption by the Governing Council – proceeds to prepare a new decision to abrogate the initial decision or replace it with a decision of identical or amended content.Article 24(7) SSM Regulation. This stands in marked contrast to, for instance, the Intellectual Property Office, whose Board of Appeal can, if it finds an appeal to be well-founded, assume the powers of the Office and directly rule on the case by adopting a decision which is attributed to the Office.Article 71 Trade Mark Regulation (footnote 4).

This setup means that ABoR acts not as a decision-maker or a (quasi-)judge who can strike down ECB decisions, but merely as an adviser to the decision-making bodies of the ECB in the adoption of a new decision on the same matter.Nonetheless, it has been held that the opinion of ABoR (which is notified to the applicant) constitutes part of the statement of reasons for the new ECB decision: Case C-450/17 P Landeskreditbank Baden-Württemberg – Förderbank v ECB [2019] EU:C:2019:372, para 95. One would have to assume that this is limited to cases where the ECB decision follows the ABoR opinion rather than deviating from it, see Andreas Witte, ‘Die Architektur des einheitlichen Bankenaufsichtsmechanismus und die Bedeutung administrativer Widerspruchsverfahren im europäischen Prozessrecht’ (2017) 52 EuR 648, at 657. Such a decision has to be adopted irrespective of the outcome of the administrative review. As a consequence, in cases where the ABoR review results in a second decision with identical content as the first or in a second decision that has the same operative part as the second (the withdrawal of Versobank’s banking licence, in the case at hand) but amended reasoning, there will be not one but two decisions of the same legal nature and effects. This raises the question that lies at the core of the present note, namely how these two decisions interact in case the applicant now wishes to seek judicial review: Is it necessary, or even possible, to contest both decisions or only one of them, and if so, which?

3.2.The problem which the General Court had to address

In Versobank, both decisions were contested separately. As an outside observer, one gets the impression that this is hardly in the applicants’ interest: Article 263(6) TFEU stipulates a strict two-month deadline; failure to bring an action within this period will lead to the finality and incontestability of the act in question.See Case C-310/97 P Commission v AssiDomän Kraft Products AB and Others [1999] EU:C:1999:407, paras 57-71 for an example of the strictness with which case law treats finality. This meant, effectively, that the applicants had to build their case against the first decision – retain legal counsel, collect evidence, formulate a reasoned submission to the Court, etc. – at a time when the ABoR review was still pending. The deadline for an action against the second decision, on the other hand, expired more than three and a half months later.The first decision was adopted on 26 March 2018, the notice of review initiating the ABoR proceedings was filed on 26 April, the action for annulment against the first decision was lodged on 5 June (within two months plus the ten days extension on account of distance from 26 March), the second decision (post-ABoR) was adopted on 17 July, and the action for annulment against that decision was lodged on 27 September 2018. In cases where – as it was here – both decisions are identical and based on the same legal and factual circumstances, this additional time is of no use to an applicant who needs to have the entire case built for the first submission already, but it could be fruitful in case only the second decision is contested. Why did the applicants choose to contest (also) the replaced first decision, then? One may only speculate as to their rationale. Possibly, the reason lies in the aim of avoiding what has been called a “Pyrrhic victory”:Andreas Witte, ‘Avoiding Lacunae for Judicial Review in the Interplay between National and Union Law in the Banking Union’ (2020) 45 ELR 569, 582. Since the second decision contains a clause repealing the first, there is room to argue that an action (only) against the second decision would, if successful, lead to the annulment also of that clause and thereby to the resurrection of the first decision. The applicants would then have been successful in their case but not obtained any advantage since the same operative part stands by virtue of the revived first decision. To avoid such an outcome, one might either rely on Article 264(2) TFEU to carve out the repealing clause from the annulment of the second decision or contest both decisions. Possibly the applicants considered the latter option safer, given the highly discretionary nature of Article 264(2) TFEU.Case 112/83 Société des produits de maïs SA [1985] EU:C:1985:86, para 18.

Likewise, a combination of actions against both decisions is not in the interest of the judicial system. It would lead to an unnecessary proliferation of cases to initiate two largely identical proceedings rather than one. What is more, since the administrative review would typically last longer than the Article 263(6) TFEU deadline (as it did here), an action contesting the first decision would, in most cases, have to be lodged at a time when it is still unknown whether the administrative review gives the applicants sufficient redress of their grievances. The procedure would then fail to achieve its core purpose, namely, the reduction of judicial caseload by means of resolving at least part of the disputes between applicants and the ECB extrajudicially.Of course, applicants can discontinue their proceedings at any time (Article 125 of the Rules of Procedure of the General Court). However, it is hardly a satisfactory solution to force applicants to initiate litigation in all cases in which they also go to ABoR in parallel, only to then discontinue a fair share of these actions subsequently if they win in the administrative review. To make things worse, discontinuance leaves applicants with the risk of remaining saddled with their costs (Article 136 of the Rules of Procedure).

An action only against the first decision, on the other hand, is not a solution; it is incapable of preventing the finality of the operative part of the second decision, which still stands. All considerations, therefore, point to the conclusion that in such scenarios, the second, and only the second decision, should be the subject matter of judicial review.

The challenge for the General Court in Versobank was to find a line of reasoning that could lead to this reasonable result in a doctrinally and theoretically sound manner. Some jurisdictions achieve this (or an equivalent, in practical terms) outcome by direct legislative fiat. German law, for instance, stipulates that where an administrative appeal (Widerspruch) has been conducted and the applicant subsequently seeks judicial remedies, the subject matter of this litigation will be the initial decision “in the shape” (in der Gestalt) which it received via the second decision.Section 79(1)(1) of the Code of Administrative Court Procedure (Verwaltungsgerichtsordnung). The second decision can only be contested in isolation if it contains a grievance not yet present in the initial one (section 79(1)(2)). – In the light of this statutory wording, German administrative law scholarship is unanimous in emphasising the “unity” of the two decisions, but some detailed aspects of this unity remain controversial: J Pietzcker, Annotations to section 79 VwGO in Friedrich Schoch and Jens-Peter Schneider (eds), Verwaltungsgerichtsordnung (Beck, 17^th supplement 2008), para 3. Writing from a German perspective but for a French audience, Ulrich Stelkens ‘L’influence actuelle du droit administrative français sur le droit administratif allemand’ in Claude Witz et al. (eds), Soixante ans d'influences juridiques réciproques franco-allemandes (Société de législation comparée 2016) 197 at 203 speaks of an ambiguous nature of the German Widerspruch. In other words, the second decision is conceptually treated not as a decision in itself but a modification of the first, with the first (as modified accordingly) being the subject matter of litigation and deadlines counted as of the second decision.Section 74(1) of the German of the Code of Administrative Court Procedure (footnote 27). In French law, both decisions will often be reviewed by the tribunal administratif, but an explicit legal provision provides for a prorogation of the judicial deadline against the first decision in case a recours administratif is lodged.Article L411-2 of the Code des relations entre le public et l'administration. The same principle had already been developed by case law in Conseil d’Etat, Section, 10 July 1964, Centre médico-pédagogique de Beaulieu, n° 60408, 399. O Le Bot, ‘La sécurisation des MARL par le juge administratif. Les modes alternatifs de règlement des litiges en droit administratif’ 2017 hal-01699440 <https://hal.archives-ouvertes.fr/hal-01699440/document> accessed 2 February 2022, at 14-15 regards this as a case of judicial intervention in an alternative dispute resolution process. This characterisation is understandable from the French point of view, where the first decision is the subject matter of litigation before the administrative court; it can hardly be conceptualised as such from the German point of view (see footnotes 27 and 28) nor from the perspective of the doctrine developed by the General Court in the present case: In these regimes, litigation follows the administrative appeal procedure but does not intervene in it.

EU law, on the other hand, offers no such solution by means of an explicit statutory provision. The Treaties know and define decisions, but not administrative appeals or reviews, and consequently the SSM Regulation could only construct an edifice of administrative review using decisions as building blocks. Owing to the primary law constraints described above, the only way the second decision can be adopted is after a non-binding opinion of ABoR that leads to the Supervisory Board proposing a new draft decision for adoption by the Governing Council. How could the outcome described above be arrived at under such conditions, where two decisions exist whose finality needs to be prorogued whilst a review is pending?

3.3.The solution found by the General Court and remaining problems

A factually similar previous case, Trasta Komercbanka, had already raised the same issue. There, the General Court had concluded that even after the repeal of the first decision by the second, an action for annulment remains admissible against the former. This result was based on the consideration that the applicant may retain an interest in the annulment of a repealed act for non-contractual damage claims.Case T-247/16 Trasta Komercbanka AS v ECB [2017] EU:T:2017:623, paras 17-23. This reasoning was in line with prior case law (cited by the General Court) that had held the same. It is, however, worth pointing out that in the cases cited, that repeal had not been the result of a formalised pre-judicial administrative review. The General Court’s order in Trasta Komercbanka was subsequently set aside by the Court of Justice upon appeal.Joined Cases C-663/17 P, C-665/17 P and C-669/17 P ECB and Others v Trasta Komercbanka and Others [2019] EU:C:2019:923. The elaborations on the continued admissibility of the repealed first decision had, however, not been appealed, and consequently, the Court of Justice did not analyse this question nor give guidance on it.

Nonetheless, there are passages in Versobank (para 93 of the judgment) that seem to indicate that the General Court considered itself, as a result of the overturning of its order in Trasta by the Court of Justice, unable to maintain this position. Therefore, in Versobank, the General Court’s conclusion was the opposite: It argues that by the time of the adoption of the second decision, the licence had already been withdrawn by the first decision and could therefore not be withdrawn a second time. It followed that the second decision inevitably had to have a retroactive effect as of the time the first decision had taken effect (para 82). This retroactive effect comprises both components of the second decision – the repeal of the first decision and the withdrawal of Versobank’s licence as pronounced in the second decision – and is, in the Court’s view, not limited to cases where the two decisions are identical. Also, where the second decision is identical in its operative part, but divergent in its reasoning or even where the second decision has a different operative part (for instance, that the licence is not withdrawn), the Court considers the retroactive ex tunc application of the second decision necessary (paras 83-84), an effect which the Court dubs the “definitive disappearance of the original decision from the legal order” (para 85). The language used in these passages of the judgment is so strong that it appears that the Court regards this outcome as inevitable, irrespective of the exact wording of the clause in the second decision that repeals the first.Indeed, shortly after Versobank, the Court used the same logic to arrive at the same conclusion when it ruled again on the remanded Trasta Komercbanka case (see footnote 11) even though the repeal clause in the second decision appears to have been different in that case than in Versobank: Compare paras 53-54 of Trasta Komercbanka with para 86 of Versobank. In Niemelä (also see footnote 11), the wording of the repeal clause appeared to have been more ambiguous (see para 46 of the Niemelä order).

This conclusion is, at least at first glance, curious on several grounds. First, it seems to disregard that, as a matter of fact, the first decision did exist during the time between its adoptionOr, more precisely, its notification, since that is the time as of which a decision with addressees takes effect: Article 297(2) TFEU. and that of the second decision; it was this first decision, which remained applicable and enforceable during that time,Unless suspensory effect under Article 278 TFEU is obtained, the threshold for which is high. which prohibited Versobank from continuing its business in that period. Second, the Court, in Versobank, hardly even mentioned the key argument that had led it to reason differently in Trasta (and the case law on which the conclusions in that case were based), namely the relevance of a pronouncement of the illegality of the first decision for non-contractual liability under Article 340(3) TFEU. Third, the Court’s conclusions appear to stand in contrast with well-established case law that has held that legal acts which merely confirm a previous act are not contestableJoined cases 166 and 220/86 Irish Cement Ltd v Commission [1988] EU:C:1988:549. – a conclusion which implies that in such scenarios judicial review takes place via a challenge to the initial act. Lastly, the Court presupposes that not only the repeal of the first decision, but also the remaining parts of the operative provisions of the second decision have effects as of the time of adoption of the first decision. This appears to go against long-standing case law according to which, in Union law, legally detrimental acts cannot be applied retroactively.In the terminology used by Roemer AG in his opinion in Case 1/73 Westzucker GmbH [1973] EU:C:1973:61, this is even a case of “real retrospective effect” (application of an adverse legal act to a time chronologically prior to its adoption) and not merely of “material or quasi retrospective effect” (application of a legal act to a factual situation that has commenced before the adoption of the act, but with the act’s effects limited to the future). See also Case C-60/98 Butterfly Music [1999] EU:C:1999:333, para 25.

The first two of these four problems are not insurmountable. It is true that during the period between the two decisions, the first decision was what prevented Versobank from continuing its banking business, but even if Versobank were to win an action against the first decision, its inactivity during this time could not be undone after the fact. This means that a claim under Article 340(3) TFEU is not merely one, but the only possible aspect under which the applicant could conceivably retain a legal interest in the annulment of that decision; there is no other aspect under which the applicant could obtain any benefit from it. But even under this heading, an action for annulment against this decision is not indispensable: The second decision is indubitably contestable since this decision forms the basis for Versobank being prohibited, in the present, from continuing business; as far as pecuniary damages which can be attributed only to the first but not to the second decision are concerned,It requires a little creativity to think of damages which would fall within this category; the best candidate would be foregone profits which Versobank would have made in the period between the two decisions if it had remained in business. a pronouncement of the illegality of the first decision under Article 264 TFEU would be a sufficient but not necessary condition for the establishment of unlawful conduct by the ECB that would be required for a non-contractual claim.Case C-445/06 Danske Slagterier [2009] EU:C:2009:178, paras 38-39. This is a consequence of the nature of the non-contractual liability action as an independent remedy.Case 4/69 Alfons Lütticke GmbH v Commission [1971] EU:C:1971:40, para 6. The legality of the first decision can therefore be addressed just as well as an incidental question to the non-contractual liability action.

As to the third problem, the Court’s logic rests on the notion that the second decision after ABoR did not merely confirm the initial decision but rather replaced it after a re-examination of the circumstances of the case – first by ABoR, then by the Supervisory Board and the Governing Council after receipt of the opinion of ABoR. Even under pre-existing case law, such a re-examination precludes the second decision from being relegated to a merely confirmatory (and hence incontestable) status.Case T-235/95 Anthony Goldstein v Commission [1998] EU:T:1998:56, para 42.

As to the fourth problem, it has to be kept in mind that in Union law, the prohibition on retroactivity emanates doctrinally from the principles of legal certainty and the protection of legitimate expectation. For this reason, the Court of Justice accepts exceptions to this prohibition where the legitimate expectations of those concerned are duly respected.Case C-376/02 Stichting “Goed Wonen” [2005] EU:C:2005:251, para 33.Versobank does not discuss this issue,There is a discussion of legitimate expectations in paras 357-366 of the judgment, but this addresses an argument made by the applicants that was different from the retroactivity of the second decision. but one may assume that in the Court’s view, the substantive identity of the two decisions is the critical factor justifying this retroactivity: Versobank received, by virtue of the second decision, the same licence withdrawal that it had already received previously by virtue of the first. This is doctrinally interesting, as it bridges and overcomes, on the substantive level, the rigidity on the formal perspectives: On the formal level, the legal architecture of the SSM has to employ two separate ECB decisions, both adopted by the Governing Council upon proposal by the Supervisory Board, because this is – in the absence of a statutorily provided administrative appeal procedure in the Treaties – the only way such a procedure can be manufactured with the available toolkit. On the substantive level, however, the Court refuses to overemphasise the distinctness of these two decisions; it recognises that the two are, in substance and in particular as far as their effects for the addressee are concerned, the same.

3.4.Requirements for standing before ABoR and the Court

A minor procedural point of the judgment lies in the fact that the administrative review that led to the adoption of the second decision had not been initiated by Versobank but instead by its main shareholder Ukrselhosprom. ABoR held this shareholder to be directly and individually concerned by the licence withdrawal decision addressed to Versobank (para 19 of the judgment). The background that led the members of ABoR to arrive at this conclusion was the order of the General Court of 12 September 2017 in Trasta Komercbanka case that has already been referred to. There, the General Court held that shareholders of a credit institution are directly and individually concerned by a decision withdrawing the authorisation of the bank in which they own shares.Case T-247/16 Trasta Komercbanka and Others v ECB [2017] EU:T:2017:623, paras 59-72. As mentioned above, this order was subsequently struck down by the Court of Justice, who concluded that shareholders are not directly and individually concerned by such a decision, but this appeal ruling was rendered only after the conclusion of the ABoR review. In other words, the reasoning on which ABoR relied for holding the notice of review admissible was established in case law at the time but subsequently overruled. Consequently, the General Court held the action for annulment against the licence withdrawal decision inadmissible in so far as Ukrselhosprom was concerned. One may expect that ABoR will, in the light of the guidance given by the Court of Justice, adjust its interpretation of direct and individual concern for administrative review purposes, thereby realigning the locus standi requirements with those for judicial review.

4.Conclusions and outlook

In the present reviewer’s opinion, the General Court’s logic – that after ABoR review, the second, and only the second, ECB decision can form the subject matter of an action for annulment – constitutes an attempt to achieve a result very similar to that in jurisdictions where administrative appeals exist as a concept in positive law. This result was partially already built into the wording of Article 24 of the SSM Regulation: It stipulates that an administrative review by ABoR must in any case lead to a second decision, even where that decision is identical to the first. This requirement, which was inserted into the Regulation at a relatively late stage of the legislative procedure, ensures the start of a new two-month deadline.The initial Commission proposal for the SSM Regulation (COM/2012/0511) did not foresee any administrative review. The Parliament’s Committee on Economic and Monetary Affairs proposed it in the form of a “Board of Appeal” with the power to directly rule on the dispute (A7/2012/392), but for the reasons explained in section 3.1 this was not feasible. It was the resolution of the plenary Parliament (CELEX 52013AP0213) that first included the amendment to establish ABoR in its present form. This goes, however, only halfway to achieving the result sketched out in section 3.2. For a complete solution, it is necessary to add a second rule: The obsoleteness, and hence incontestability, of the repealed first decision.Going forward, it appears that even a third rule might become necessary to complete the picture: That, in the case of a successful action for annulment against the second decision, this does not lead to the resurrection of the repealed first one. Versobank did not cover this aspect since the applicants lost their case, but the result appears inevitable if one wishes to avoid the absurdities of the “Pyrrhic victory” described in section 3.2. How can this result be achieved? Following the principle of ne ultra petita, careful applicants might explicitly request the Court, in their form of order sought, to keep the repeal clause in the second decision intact. But even if applicants forget to request this, one would assume that judicial fairness requires the Court to do so out of its own motion, if it strikes down the second decision. In the present reviewer’s opinion, Article 264(2) TFEU is the preferable vehicle to achieve this. In the absence of an explicit provision establishing this rule, it was far from clear whether the first decision was, indeed, incontestable – as evidenced by the fact that the Court itself had previously decided differently in Trasta. The aim of the entire exercise was, thus, to arrive at a conclusion that does pragmatic justice to two different, but not mutually conflicting, objectives: The interest of the judicial system at large not to unnecessarily inflate the workload of the Union courts, and the interests of remedy-seeking applicants to obtain the effective judicial review to which they are entitled under the Charter of Fundamental Rights. This objective was achieved, and even though the Court’s conclusions meant that the applicants in casu lost their application against the first decision as devoid of purpose, the result should be welcomed by future applicants in particular as it works in their favour with respect to the calculation of deadlines.There is, however, leeway for the Court to alleviate the impact of this consequence when deciding ex aequo et bono on costs. And indeed, in Versobank (as well as the other cases cited in footnote 11 that follow it), the Court found it fair to order each party to bear its costs for the action against the first decision. What is more, the General Court was – as mentioned above – vocal in emphasising that its conclusions are not limited to cases where the two decisions are substantively identical; they supposedly apply also where, after administrative review, the ECB concludes that the shortcomings identified can be remedied by less onerous measures (para 83). This leaves a wide range of possibilities since a licence withdrawal is the ultima ratio of banking supervision, relegating virtually any other measure to the level of a “less onerous” one. Does this mean that the ECB could, after the administrative review of a licence withdrawal, replace that withdrawal with, for instance, a governance rearrangement requirement and apply the obligations of this requirement retroactively? The wording of the judgment indicates that the answer is affirmative, but where the new requirement is substantively very different from what has previously been on the table, such retroactivity could leave the reader with a sense of unease, especially where the retroactive requirement would be underpinned by sanctionability.To illustrate this, let us develop the scenario from the text further: A licence withdrawal decision is contested before ABoR, and following this review, the licence withdrawal is replaced with a newly formulated requirement to make governance changes. In the time between the two decisions, the bank in question has not acted in line with this new requirement (because neither side had even thought of it at that time). Can the ECB now penalise the bank for this non-compliance, citing the retroactivity of the second decision as established by Versobank? Even though Article 18(7) of the SSM Regulation in conjunction with Regulation (EC) No 2532/98 does not require mens rea, it seems that here one would have to disregard the principle of ignorantia legis non excusat: The bank could not reasonably have been expected to know, during the time in question, that such a requirement would be formulated and imposed in the future. More generally, what about the reverse scenario, where the second decision is more onerous than the first?Union law contains a prohibition of such a reformatio in pejus where a ruling of a lower court is appealed to a higher court (Opinion of AG Trstenjak in Case C-16/06 P Les Éditions Albert René SARL v EUIPO [2007] EU:C:2007:728, paras 35-36). It is doubtful whether the same principle applies in administrative appeals (see the scepticism in Case T-817/16 Vans v EUIPO [2018] EU:T:2018:880, para 157), which are conceptualised as administrative rather than quasi-judicial in nature. Here, it appears obvious that what matters is contesting the second decision, and the first is moot (and hence incontestable). What if the two decisions are identical in their operative parts but differ in their statements of reason? The Court’s elaborations state that no action is available against the first decision in such a scenario either. This means that, where the reasoning of (only) the first decision contains statements that the applicant considers detrimental to its reputation, no rehabilitation can be sought by means of having that decision annulled judicially to clear the applicant’s name.This consideration was a part of the case law cited by the General Court in Trasta in support of the view that even a retroactive repeal of a decision does not eliminate the legal interest in an action against it: Case C-239/12 P Abdulbasit Abdulrahim v Council and Commission [2013] EU:C:2013:331, para 83.

On a wider level, it has been pointed out above that the background to Versobank was heavily influenced by peculiarities of the SSM, especially the ECB’s institutional status under the Treaties that forced secondary law to limit ABoR to a purely advisory role. This casts some doubt on the transferability of the Court’s logic to other areas of law – in particular to administrative appeal set-ups where the review body does have its own decision-making powers to rule on the case directly, as, for instance, in the practically very relevant field of intellectual property. Nonetheless, the Court’s line of argument in the relevant passages sounds categorical. This would indicate that the Court intended to develop an aspect of a general theory of administrative appeals, not merely to decide on a very narrow question that is applicable only to the specific institutional architecture of the SSM. This should be welcomed, since the doctrinal analysis of administrative appeals as a generally observable and widespread phenomenon is, in Union law,And in particular in case law; scholarship has – as the sources cited in this paper illustrate – taken up the challenge of analysing administrative appeals in EU administration. lagging behind that in national jurisdictions where a cross-sectoral explicit basis for them exists in general administrative law.