An Administrative Crack in the EU’s Rule of Law: Composite Decision-making and Nonjusticiable National Law

Composite administrative procedures – Exclusive jurisdiction of Union courts to review non-binding national preparatory acts – No jurisdiction of Union courts to enforce national law – Autonomy and uniformity of EU law – No judicial control possible of violation of domestic law by national authorities – National rule of law gap – Judicial review, effective judicial protection, and principle of administrative legality

system is increasingly characterised by the EU legislature's creation of composite procedures, i.e. administrative procedures whereby decisional power is exercised jointly by, rather than being strictly divided between, national and EU authorities in a succession of national and EU procedural phases. 3 Herein lies the crux of the constitutional problem explored in the present paper. It is not that the dualistic design of the Union's system of administrative justice, whereby national measures may only be reviewed by national courts and EU measures only by EU courts, is inherently faulty. Rather, that system was designed for conditions of strict division of decisional power between national and EU authorities. The Treaties' framers could not have anticipated that this system would be outpaced by the rise of composite decision-making, whereby both levels of administration take decisions together.
The claim this paper makes is that, despite the best efforts of the Court of Justice and the General Court to mitigate the consequences of this mismatch, a serious judicial review gap remains. In composite administrative procedures, the violation of national law by national authorities is in most in most instances beyond judicial review. As the paper will further illustrate, this judicial review gap is inevitable given the current state of European Union law. It is an unintended and regrettable consequence of a peculiar alignment in the European Union's constitutional firmament between the jurisdictional boundaries of national and EU courts, loyal cooperation, separation of powers, and the principles of autonomy and uniformity of the EU legal order.
The second section clarifies what composite administrative procedures are, why they are at odds with the EU's judicial system, and how EU courts have sought to compensate the mismatch between the two. The section will also highlight the relevance of the recent landmark Berlusconi ruling, in which for the first time the Court explicitly confirmed something that had hitherto been only a matter of academic speculation. The Court held that whenever EU authorities have discretion in a composite procedurewhich is most often the casenational courts are forbidden from reviewing national preparatory acts.
The third section explains why that prohibition was the only plausible solution from an EU constitutional law point of view. However, if national courts cannot review national preparatory acts, the only other instance at which a violation of national law can be reviewed would be the EU courts, i.e. upon review of the final administrative decisions taken at the EU level. And yet, as the fourth section will argue, that is not constitutionally possible, either. The fifth section fleshes out the implications of the two that precede it. National courts enjoy jurisdiction to apply national law, but may not review national preparatory acts. EU courts may review national preparatory acts as part of the review of a final EU-level decision, but may not apply national law while doing so. The result is that, in most composite procedures, neither national nor EU courts have jurisdiction to review compliance with national law by the member states' authorities. Just to give a few examples, no judicial review would seem to be possible in cases involving a violation of national legal requirements with respect to quorum, advisory opinions, and conflicts of interest, or rules which offer a higher degree of protection for procedural rights than required by EU law. In general, no judicial review would appear to be possible for any aspect of national administrative law that is not simultaneously also enshrined in EU law.
Composite administrative procedures therefore generate, at the level of the member states, a serious gap from the point of view of three key requirements of the European Union's rule of law: the effective judicial protection of individuals, the judicial control of (national) public power, and the principle of administrative legality. While a few potential exceptions can be identified, as the sixth section will illustrate, there is little that the EU's existing system of judicial review can do to 'mend' this threefold gap.
The conclusion summarises the findings and suggests that the only definitive solution capable of mending the gap would be an unprecedented expansion of the jurisdiction of the European Court of Justice and the General Court so as to empower them to apply national law, and perhaps even the creation of 'reverse preliminary references'. Since similar Treaty reforms are unlikely to occur anytime soon, some ideas worthy of reflection are proposed for future research: is it time to consider non-judicial remedies to compensate for the limitations of the EU's judicial system?

I  
The EU's administrative system is increasingly characterised by cooperation between national and EU authorities. 4 Intensive cooperation, both horizontally between national authorities and vertically between national and EU administration, is often perceived as an approach that can secure effective and consistent 4 For a classic piece, see by H. Hofmann  implementation of EU law across the member states while respecting their reluctance to centralise administrative powers at the EU level. 5 From these two competing desiresuniform implementation of EU law throughout the EU and preservation of national administrative sovereigntya wide variety of forms of administrative cooperation have emerged in the EU's administrative system. 6 Those forms of cooperation range from informational cooperation, i.e. the exchange of information between different levels of administration, 7 to organisational cooperation, i.e. the creation of new administrative bodies where various levels of administration are represented, such as EU Agency boards. In addition, there is also procedural cooperation, which is embodied in the legislative creation of composite administrative procedures.
Composite decision-making has become a pervasive aspect of the EU's administrative system. It is the means by which measures of administrative implementation are adopted in countless different policy areas, ranging from the governance of genetically modified organisms to pharmaceutical regulation, from food safety to banking supervision, and from structural funds management to the environmental matters of aquaculture.
Like any administrative decision-making procedure, composite procedures follow a pre-defined sequence of procedural stages, with a succession of preparatory steps paving the way for the adoption of a final decision. 8 What characterises composite administrative procedures is that the final decisions adopted pursuant to them require a cumulative exercise of decisional competences at procedural stages at national and then EU levels. 9 The hallmark of composite decisionmaking is decisional interdependence, i.e. one level of administration may not exercise its competences in the procedure before the other level has concluded 5 T. Bach  the previous decisional stage. 10 National administration is generally involved in the earlier stages of a composite procedure, although its role can vary significantly. Depending on the relevant administrative procedure, a national authority may: (i) adopt binding preparatory acts which leave no discretion to the EU authority involved in a subsequent procedural phase, thereby predetermining the final outcome of the procedure; (ii) adopt provisional acts which will remain in force only until confirmed, amended, or rejected by an EU authority at a later procedural phase; 11 (iii) propose a draft decision which, while not binding on the EU administration, will serve as the basis for making that decision; 12 or (iv) serve as a 'local secretariat' with the merely formal or bureaucratic task of forwarding initial applications, documents, and information to the EU Agencies or the Commission, i.e. whichever will subsequently decide on the substance of the case. 13 Occasionally, legislation requires an additional concluding procedural stage at the national level after the EU administration has adopted a decision, at which the national authority involved enjoys no discretion and fulfils a merely formal 'rubber-stamping' role. 14 In such cases, we may well describe the EU-level decision as the procedure's final decision.
Composite procedures in which national administration has the role explained at (i) above are extremely rare. Practically all existing, composite procedures correspond to types (ii) to (iv). Put differently, most composite procedures are 10 See for example GC 30 April 2014, Case T-17/12, Hagenmeyer, para. 136 and ECJ 13 February 2014, Case C-31/13 P, Hungary v Commission, paras. 60-61. 11 See for example Arts. 53(1) and (2) and Art. 54 of Regulation (EC) No. 178/2002 of the European Parliament and of the Council of 28 January 2002. According to the administrative procedure established in the provisions, national authorities may adopt urgent measures such as the suspension of the sale of a potentially unsafe food item, which remain in force until confirmed or reappraised by the Commission. 12 In the sector of detergents, national authorities prepare an assessment report on their risk to human health or the environment and submit it to the Commission, which will adopt a final positive or negative decision that may totally diverge from the report (Arts. 5(3)-(4) and 6 of Regulation characterised by decisional dominance at the EU leveli.e. it is the EU administration that exercises the discretion to define the final decision's content. Only exceptionally does one encounter composite procedures that have national decisional dominancei.e. where a national authority adopts preparatory acts that are legally binding on the EU administration and predefine the terms of the final decision at the EU level, such as the registration of protected designations of origin 15 and the management of certain tariff quotas. 16 The variable relative weight of national decisional power is of critical importance to understanding the European Court of Justice's case law on the judicial review of composite procedures. The review of administrative decisionmaking is, in principle, strictly divided between national courts and EU courts. National courts have exclusive competence to review decisions taken by national authorities 17 and EU courts have exclusive competence to review decisions taken by EU authorities. 18 This dualistic separation of judicial review has been appropriately described by Marchetti as a principle of 'double exclusivity'. 19 Double exclusivity implicitly assumes that any given decision will be adopted either by national or by EU authorities. It presupposes that national and EU authorities exercise their respective decisional powers independently of one another. However, composite procedures are based on decisional interdependence. As they constitute forms of joint decision-making that involve both national and EU authorities, no decision can be individually imputed exclusively to only one of the two levels.
The European Court of Justice's case law on the judicial review of composite procedures aims to address the mismatch between the principle of double exclusivity and the decisional interdependence of composite decision-making. 15 See Art. 53(2)(a) of Regulation No. 1151/2012 of 21 November 2012. According to the relevant administrative procedure, national authorities receive applications from groups for the classification and registration of products as protected designations of origin or protected geographical indications. After they make a positive assessment of the products' compliance with the EU requirements for registration, and if no objection is raised by any legal or natural person, national authorities support the applications and forward them to the Commission. The Commission is then obliged to register the products for EU-wide protection, without enjoying any margin of discretion. 16 Commission Regulation No. 969/2006 of 29 June 2006, opening and providing for the administration of a Community tariff quota for imports of maize from third countries. National authorities receive applications for import licences and submit to the Commission a list of those which, according to their discretionary assessment, are found eligible. The Commission only checks whether traders have not defrauded the quota mechanism by submitting their applications to the authorities of more than one member state. That case law is based on the recognition that, although national and EU authorities exercise shared decision-making power in composite procedures, that power is never shared equally. The case law consists of a two-track approach, depending on the level of administration which, according to the relevant sectoral legislation, enjoys the power to define the content of the final decision. 20 The first approach concerns those rare composite procedures in which national decision-making predominates, as represented by the doctrine established in the European Court of Justice's Borelli ruling. 21 The Borelli case law holds that in composite procedures in which a national authority adopts a preparatory act that is binding on the EU authority to whom it is addressed, leaving it no discretion, and therefore 'predetermining the terms' of the decision it will take at the end of the administrative procedure, EU courts may not review the final EU decision on the grounds that it is based on an illegal national preparatory act. 22 Because the EU courts do not have jurisdiction to review national acts, they may not declare that final decision invalid. At the same time, to exclude the risk of a gap in effective judicial protection, 23 the Borelli case law requires national courts to review the binding national preparatory acts on 'the same terms on which they review any definitive measure adopted by the same national authority'. 24 The second approach applies to those composite procedures in which EU authorities are not obliged to decide in accordance with the preparatory measures of national authorities. This second approach, which developed in a rather fragmentary manner over the years, has recently been codified and clarified in the Berlusconi ruling. 25 The Berlusconi case concerned a decision taken by the European Central Bank pursuant to the composite procedure set out in Article 15 of the Single Supervisory Mechanism Regulation. 26 The administrative procedure in question governs authorisation for the acquisition of qualifying holdings, i.e. the ownership of shares in a credit institution that gives the holder thereof a quite significant degree of influence over its activities. It is a composite, two-step administrative procedure. A national supervisory banking authority first assesses whether the acquisition of a particular qualifying holding meets the substantive requirements for it to be authorised, such as the technical expertise and good reputation of the acquiror. Once it has made that assessment, the national supervisory banking authority forwards it to the European Central Bank, proposing that the latter adopt the final decision in the procedure by opposing or, as the case may be, not opposing the qualifying holding. The European Central Bank makes its own independent assessment of the case, and is legally empowered to decide differently from what was proposed by the national supervisory authority. In Berlusconi, the Bank of Italythe Italian supervisory banking authorityadopted a preparatory act proposing to the European Central Bank that Mr Berlusconi, due to a conviction for tax fraud, no longer fulfilled the good reputation requirements to hold a qualifying holding. The European Central Bank concurred, and adopted a final decision opposing Mr Berlusconi's qualifying holding. In a preliminary reference, the question was raised of which court, national or European, is competent to review the decisions taken in the composite procedure of Article 15 Single Supervisory Mechanism Regulation.
In its Berlusconi judgment, the European Court of Justice held that in composite procedures in which the national preparatory act is not binding on the EU authority, the EU courts must have exclusive jurisdiction to review the final EU-level decision and any national preparatory acts that preceded the EU authority's final decision. Conversely, given such exclusive jurisdiction, the Court held that those national preparatory acts 'cannot be subject to review by the courts of the Member States'. 27 The Berlusconi ruling clarified that whenever the EU administration enjoys decisive powersas is overwhelmingly the caseit is left to the European Court of Justice to review all acts adopted in a composite procedure, including national preparatory acts. The question that the EU courts have yet to explicitly address is whether, in doing so, the European Court of Justice may review violations of national law.
The Union courts have yet to review national preparatory acts against standards of national law. They have so far only reviewed them against rules or principles comprised within EU law, such as the right to be heard. 28 However, EU legislation creating composite procedures often regulates only very specific aspects of the national phases. Whenever the relevant EU legislation is silent on the matter, it is 27 Berlusconi, supra n. 25, paras. 43 and 47. 28 Brito Bastos, supra n. 3, p. 126 ff.

70
Filipe Brito Bastos EuConst 16 (2020) national administrative law that governs the composite procedure at the national level. The issue here is that, while the European Court of Justice enjoys full jurisdiction to review violations of EU law by EU authorities, it does not have jurisdiction to apply national law or to review any violation thereof. Therefore, to address violations of national law, only two solutions are possible in abstracto.
The first would be to extend the reach of the Borelli case law to non-binding national preparatory acts, thereby making it possible for national courts to review them in light of national law. The second solution would be for the European Court of Justice itself to assume jurisdiction to apply national law when reviewing non-binding national preparatory acts in composite procedures. Nevertheless, as will be shown in the two next sections, neither solution is constitutionally tenable.

W         ()  
In instances of purely national decision-making, the European Court of Justice considers that national courts are not required to review preparatory acts adopted by national authorities. 29 According to the Berlusconi ruling, in contrast, national courts are forbidden from reviewing national non-binding preparatory acts adopted within composite procedures. That prohibition constitutes a significant innovation in terms of the Court of Justice's case law. Prior to Berlusconi, the European Court of Justice had only suggested this. In the 1990s, for instance, the Luxembourg Court was called upon to decide on a number of cases pertaining to composite decision-making with regard to the European Social Fund. According to the applicable administrative procedures, national authorities that detected irregularities in the use of European Social Fund funding had to draft and forward to the Commission a proposal for the reduction or withdrawal of financial assistance to the relevant beneficiary. In several cases the Commission argued that judicial review should be exercised by national courts, as national authorities had adopted the preparatory acts. However, the European Court of Justice rejected this argument, holding that it was 'the Commission which assumes, vis-à-vis the recipient of European Social Fund assistance, the legal responsibility for the decision by which its assistance is reduced, irrespective of whether that reduction was or was not proposed by the national authority concerned'. 30 By contrast, the Berlusconi case went much further. It did not merely state that judicial review should be conducted at the EU level, if that is where the power to make a decision in a composite procedure actually resided. The Advocate General and the Court of Justice held that any other kind of judicial review, i.e. at the 29 See ECJ 28 July 2011, Case C-69/10, Diouf, paras. 55-56. 30 See GC 12 January 1995, Case T-85/94, Branco, paras. 18, 24 and 28. national level, was forbidden under EU law. The two provided two distinct constitutional arguments in this regard.
In his Opinion, Advocate General Sánchez-Bordona indicated that the IBM case law could offer a useful analogy for the judicial review of composite decision-making. 31 The IBM case law concerns multi-step decision-making procedures which, unlike composite procedures, take place at the EU level only. Sánchez-Bordona highlighted that, according to settled case law, only binding legal acts of EU authorities can be challenged and that, consequently in the case of acts or decisions adopted by a procedure involving several stages, an act is open to review only if it is a measure definitively laying down the position (usually, of the Commission or the Council) on the conclusion of that procedure. Those purely preparatory acts may only be relied on indirectly on the review of the final decision that an EU authority will take. 33 The constitutional justification for this case law, although textually absent from the Advocate General's Opinion in Berlusconi, pertains to the separation of powers. The thrust of IBM is respect for the EU administration's own decision-making authority. A pre-emptive review of EU preparatory acts would result in judgments on issues 'on which the [EU administration] has not yet had an opportunity to state its position definitively'. 34 In other words, judicial review of purely preparatory acts would 'be incompatible with the system of the division of powers between the Commission and the Court ( : : : ) and the proper course of the administrative procedure to be followed in the Commission' (emphasis added). 35 In proposing application of the IBM line of reasoning to composite procedures, 36 Advocate General Sánchez-Bordona implicitly recognised that judicial review of national preparatory acts would have raised the same constitutional problem, albeit that, in that case, judicial review by a national court, and thus not by the European Court of Justice, would prematurely end the administrative procedure before the EU administration had even had the opportunity to exercise its decisional powers. Indeed, it had already been suggested in the literature that it would be 31 ECJ 11 November 1981, Case 60/81, IBM.

32
See the Opinion of AG Sánchez-Bordona, para. 108, which reproduces the passage in para. 10 of the IBM ruling. incomprehensible for EU courts, due to separation-of-powers concerns, to refuse to review incomplete decision-making processes of the EU administration only then to enable national courts to do just that. 37 In Berlusconi, the Grand Chamber opted for a different, albeit equally plausible argument to denying the reviewability of non-binding national preparatory acts. The Court held that the division dictated by the Treaties between the jurisdictions of national and EU courts needs to be interpreted in light of the principle of loyal cooperation (Article 4(3) TEU). Where, as in the Single Supervisory Mechanism, the Union legislature opts for administrative procedures in which national authorities adopt acts that are preparatory in nature to the exercise of a decision-making power of the EU administration, it seeks to establish 'a specific cooperation mechanism' between national and EU authorities. 38 The implicit argument drawn from this characterisation seems to be that, if a decision-making procedure as such is defined by multilevel administrative cooperation, then the spirit of cooperation must inform how courts review that procedure. In this regard, the Court held that, 'in order for such a decision-making process to be effective', there may only be 'a single judicial review'. Where national authorities adopt non-binding preparatory acts, that review must be conducted 'by the EU Courts alone, only once the decision of the EU institution bringing the administrative procedure to an end has been adopted'. 39 If national courts were to accept the use of judicial remedies against the national preparatory acts, two risks would follow. 40 First, the national court and EU courts could come to 'divergent assessments in one and the same [administrative] procedure'. Second, by exercising judicial review of the national preparatory acts, member states' courts would anticipate the unitary judicial review that should occur only after the conclusion of the administrative procedure at the EU level and be conducted exclusively by EU courts. Accordingly, by virtue of the principle of loyal cooperation, national courts should defer to the exclusive jurisdiction of EU courts to review the entirety of the composite procedure, and refrain from reviewing national preparatory acts adopted in the course thereof. 41 The unreviewability of preparatory acts, in EU or national law, is generally not problematic, as such, because they can still be reviewed incidentally as part of 37 Brito Bastos, supra n. 3, p. 134. 38 Berlusconi, supra n. 25, para. 48. 39 Berlusconi, supra n. 25, para. 49. 40 Berlusconi, supra n. 25, para. 50. 41 Berlusconi, supra n. 25, para. 47. a judicial challenge to a final decision. At that stage, courts have the opportunity to check whether all relevant legal requirements were complied with in the course of the entire administrative procedure. However, the fact that national courts in composite procedures cannot review national preparatory acts generates a rule of law gap because the European Court of Justice cannot check whether national law has been observed when reviewing the final EU-level decision. There are at least three reasons for this. One is of a procedural nature and concerns the jurisdictional boundaries of EU courts. The other two concern formal attributes of EU law that are required by the EU's constitutional order: the related principles of the uniformity and autonomy of EU law.
First, the European Court of Justice cannot check whether national law has been observed in composite procedures since it has no competence to apply national law to begin with. Generally speaking, national law is a point of fact, not of law, in the European Court of Justice's judicial proceedings. 42 As early as the cautious ruling in Stork, the Court held it that is 'not normally required to rule on provisions of national law' when reviewing the High Authority's decisions. 43 Also, in preliminary rulings 'the jurisdiction of the Court is confined to considering provisions of EU law only' (emphasis added). 44 Similarly, in actions for annulment, the 'European Union judicature must, when adjudicating on an action for annulment ( : : : ) assess the lawfulness of the contested act in the light of the EC Treaty or of any rule of law relating to its application, and, thus, of European Union law', 45 with the consequence that 'a national provision cannot be usefully relied upon in support of an action for annulment directed against an EU measure' (emphasis added). 46 Over many decades, the case law has progressively derived from the Treaties an absolute prohibition for the European Court of Justice to apply national law in judicial proceedings of any kind. In his Opinion in the landmark ruling Association Greenpeace France, Advocate General Mischo made clear that the judicial review of composite decision-making is no exception. The case concerned a Commission decision adopted pursuant to the composite procedures for the 42 For a critical and more nuanced analysis, see M. Prek and S. Lefèvre, 'The EU Courts as "National" Courts: National Law in the EU Judicial Process', 54 Common Market Law Review (2017) p. 369. The authors admit, however, that there is only one instance in the Treaties that acknowledges that national law may serve as a point of law and not of fact, namely, Art. 272 TFEU. Prek and Lefèvre also consider that the renvoi to national law in EU legislation renders that national law applicable by EU courts. authorisation of genetically modified organisms under Directive 2001/18, which had been based on an irregular national preparatory act. While ultimately both the Court and the Advocate General accepted that that irregularity could be scrutinised as part of the review of the Commission's decision, the latter cautioned that it was 'impossible to see how the Community Court, which has sole jurisdiction to declare a Community act invalid, could form an opinion as to the existence of an irregularity with respect to national law'. 47 Some renowned scholars have suggested that, given the difficulties in ensuring the judicial review of composite procedures, the Court should reinterpret the limits of its jurisdiction. Gaja, for instance, has argued that the Court should be empowered 'to examine all the questions that are relevant for ascertaining the validity of the Community actwhether these questions relate to facts, EC law or national law' (emphasis added). 48 The European Court of Justice has significantly expanded its jurisdictional boundaries before. In rulings such as Les Verts and Sogelma, the Court claimed the power to review the acts of the European Parliament and EU Agencies, even though the wording of the Treaties does not envisage such a review. 49 If the European Court of Justice can broaden the scope of EU acts subject to judicial review, why should it not then be able to broaden the scope of the laws against which those acts may be reviewed? If the rule of law justifies claiming jurisdiction over acts of the European Parliament and EU Agencies despite the letter of the Treaties, should it also not justify claiming jurisdiction to apply national law in cases involving composite procedures?
In my view, the answer must be negative. The jurisdictional limits of the European Court of Justice are not the only reason it cannot review national preparatory against national law. The third and second reason concern the closely related principles of the uniformity and autonomy of EU law.
The case law of EU courts has long accorded paramount importance to the requirement of the uniformity of EU law. 50 EU law demands that measures of EU authorities 'must deploy their full effects, in a uniform manner in all member states, as from their entry into force and throughout the duration of their validity'. 51  in question'. 52 Even though Foto-Frost dealt with the specific risk to the uniformity of EU law resulting from 'divergences between courts in the member states as to the validity of Community acts', 53 composite procedures raise that same risk in the opposite scenario. They raise the same risk of the uniformity of EU law being compromised, not by differences in national judgments on the validity of EU acts, but by the European Court of Justice applying different national laws in the judicial review of EU acts adopted at the conclusion of composite procedures.
Consider the following example. Under EU law, 54 the establishment of fish farms requires administrative authorisation if they are intended to grow non-native species. Whenever the installation of those fish farms is likely to have cross-border environmental effects, the authorisation procedure is composite. A national authority receives the application for the permit and, in the event of a positive assessment, forwards a non-binding draft decision to the Commission for a final decision. Nothing is stated in the relevant EU Regulation about the right to be heard of affected parties within the member state where the aquaculture facility is to be installed.
If an application for an aquaculture permit in Portuguese sea waters were to be submitted, its assessment would fall under the competence of that country's Ministry of the Sea. However, under Portuguese administrative law, local government authorities are entitled to be heard in any administrative procedure to defend certain public interests, including public health and the environment, even if the outcome of the procedure does not change their legal position. 55 Accordingly, if the Ministry failed to give an adequate hearing to a municipality whose interests could be affected by the fish farm, the draft decision sent to the Commission would be illegal from the perspective of Portuguese law. If the European Court of Justice could and then did annul the final decision on grounds of national law, that would mean, in this particular example, that the Commission decision authorising a fish farm was invalid. However, not all member states give their local authorities the same broad rights of participation in administrative procedures. Had the composite procedure for the issue of the aquaculture permit been initiated in a different member state, the Commission's decision would have been valid even if all other factual circumstances remained the same. If the legality of final EU decisions depends on the conformity of their preparatory acts with national law, then those final EU decisions will be legal or illegal depending on the variations between the administrative laws of the member states in which the preparatory acts were adopted. In brief, the uniform application of the legislative instruments establishing composite procedures would be compromised if the validity of final EU administrative decisions depended upon the proper observance of the varying requirements of national administrative law. In this light, the European Court of Justice cannot be expected to apply national law in the review of composite procedures. Historically, it was the demand of uniformity of EU law that justified the European Court of Justice's very existence, 56 a link that is perhaps most evident in the preliminary reference mechanism under Article 267 TFEU. 57 It would be entirely inconsistent with the very constitutional function of the European Court of Justice if it were called upon to be the guardian of national administrative law idiosyncrasies in the judicial review of composite procedures.
Lastly, an expansion of the European Court of Justice's jurisdiction to the application of national law is discouraged by the principle of the autonomy of the EU legal order itself. The remote origins of that principle lie in the Bosch ruling, in which the Court of Justice first acknowledged that the 'law of any member state ( : : : ) and Community law constitute two separate and distinct legal orders'. 58 Shortly thereafter, the Court delivered the Van Gend en Loos and Costa/ENEL rulings, 59 in which it famously anointed the EU as a 'a new legal order' with 'its own legal system'.
By claiming its own normative autonomy, the EU legal order asserts itself as a self-referential legal system, in the sense that the 'creation, validity, application and interpretation of a legal rule depend exclusively on the order of which that rule constitutes a part'. 60 Among those 'legal rules' of the EU legal order, one must logically count those created by decisions taken by EU authoritiesincluding those taken by EU authorities at the conclusion of a composite administrative procedure.
The autonomy of EU law, as is well-known, has been defended by the European Court of Justice in respect of international law; 61 but crucially, it also means that the legality of EU acts can only be assessed against criteria established by EU law, to the exclusion of national law. 62   of the EU legal order, 'the law stemming from the EU Treaty and the FEU Treaty, an independent source of law, cannot because of its very nature be overridden by rules of national law, however framed, without the legal basis of the European Union itself being called into question'. 63 In this light, it becomes apparent why the principle of autonomy militates against the European Court of Justice reviewing the conformity of national preparatory acts with national law as part of the judicial review of the final decision taken by an EU authority. Like a sort of administrative Trojan horse, those national preparatory acts would carry the violation of national law into the EU level, leading to the final EU decision being invalid on grounds of national law, and therefore breaking the autonomy of the EU legal order from within.
In the Berlusconi case, the Advocate General made it explicitly clear for the first time that this had been a real concern in previous European Court of Justice rulings on the judicial review of composite decision-making. 64 To accept the notion that national preparatory acts that are illegal on grounds of purely national law could contaminate the final decision of an EU authority would be in essence the same as judging said decision against standards of national law.
If the European Court of Justice were to claim jurisdiction to apply national law in cases involving composite administrative procedures, as some of the literature suggests, it would have to depart from its traditional creed that autonomy constitutes an existential principle of the EU legal order. 65 Indeed, that extension of its jurisdiction would lead to the review of final EU-level decisions against national legal requirements that had been violated at the preparatory stages of the composite procedure. It would be entirely inconsistent for the European Court of Justice to have forbidden national courts from reviewing EU acts against national rules for decades, only then to claim jurisdiction to conduct that review itself.

I ()  ,           
Given that national courts must refrain from reviewing national preparatory acts that are not binding on the EU administration, a violation of national law cannot be reviewed at the national level. However, for the reasons considered in the previous section, the violation of national law cannot be checked by Union courts, either. As there is no court that can review a violation of national law in composite procedures, national authorities that breach their own national administrative laws are immune to judicial control in substantial swaths of the EU's administrative system. This poses obvious concerns from the point of view of the rule of law. The rule of law constitutes a foundational principle of the European Union. 66 Listed today in Article 2 TEU as one of the values on which the Union is based, the notion of the rule of law was first introduced into EU constitutional law by the Court of Justice. In Les Verts, the Court famously declared that the European Economic Community is a community based on the rule of law, inasmuch as neither its member states nor its Institutions can avoid a review of the question whether the measures adopted by them are in conformity with the basic constitutional charter, the Treaty. 67 It has been suggested that the above passage embodies the Court's espousal of the notion of the Community as a Community of law, or Rechtsgemeinschaft, which was at the centre of Commission President Hallstein's constitutional and political thought. 68 'The majesty of the law', Hallstein believed, was to 'accomplish that which centuries of blood and iron could never have'a united Europe. 69 In the place of the contest of 'military power and political pressure', 'of power and its manipulation', or of the 'striving for hegemony', law's rule was to become the defining feature of Europe's new political order. 'The Community is a community of law', Hallstein wrote, 'because it serves to realize the idea of law'. 70 It is striking how this account resonates with the wording of Les Verts. Indeed, the ruling makes plainly evident that, as participants in the European Rechtsgemeinschaft, neither the institutions nor the member states can evade law's rule. The EU's version of the rule of law is therefore an overarching ideal that must inform and limit the exercise of public power not just at the supranational, but at the national level as well. It is beyond the scope of this paper to propose solutions to such perennial theoretical disagreements as whether the rule of law has universal features, 71 or whether its requirements are only formal, or also substantive, in nature. 72 Even if one conceives of it in its most elementary version, the gaps of judicial review caused by composite decision-making are clearly irreconcilable with the rule of law.
At a bare minimum, the rule of law must reflect the common concern that public power should be limited by legal standards that are enforceable by independent courts, 73 suitable for preventing the arbitrary or excessive use of public power, 74 and foreseeable enough for individuals to adapt their behaviour accordingly. 75 In EU law, the ideal of the rule of law translates into a cluster of legal principles, thus addressing various aspects of these requirements. These subprinciples of the rule of law have been fleshed out by the European Court of Justice over time. 76 They include the principle of proportionality, the principle of protection of legitimate expectations, the principle of legal certainty, as well as principles relating to procedural guarantees such as the right to a reasoned decision, the right to be heard by the administration, and the latter's duty of care. Nevertheless, the unreviewability of violations of national law in most composite procedures proves particularly problematic from the point of view of three other EU rule of law requirements. The first two pertain to effective judicial protection and judicial reviewtwo distinct, 77 albeit interrelated requirements. The third pertains to the principle of legality.
The principle of the fundamental right to effective judicial protection was first explicitly incorporated into EU constitutional law in Heylens. 78  Rights. In recent years, the Court of Justice has delivered a number of salient rulings concerning the interpretation of that right. It is considered to be 'the essence of the rule of law' in EU law. 79 ASJP and Commission v Poland clarified that the right to effective judicial protection requires that all national courts which may hear cases within fields covered by EU law must fulfil the criteria of impartiality and independence from external (e.g. political) pressure. 80 Nevertheless, the right to effective judicial protection not only requires courts to have certain institutional characteristics that shield them from bias or external influence: Article 52(3) of the EU's Charter of Fundamental Rights requires that the rights enshrined in the Charter that correspond to rights enshrined in the European Convention on Human Rights must be interpreted as having the same meaning and scope as under the Convention. The right to effective judicial protection is one such right. 81 If the right to effective judicial protection as enshrined in Article 47 of the Charter must be given the same meaning as the right to an effective remedy as enshrined in Articles 6(1) and 13 ECHR, then the European Court of Justice will only be in full compliance with that fundamental right if it has the 'jurisdiction to examine all questions of fact and law relevant to the dispute before it'. 82 Given that the European Court of Justice has exclusive jurisdiction to review a final decision taken by the EU administration, as well as the national preparatory acts leading to it, a violation of national law governing the national procedural phases of composite procedures is undoubtedly a question of law that is relevant to a dispute on the review of the final decision. And yet, as Advocate General Bot has put it, 'the adage juria novit curia does not extend to national law, of which the EU judicature is not deemed to be aware'. 83 The lack of jurisdiction of the European Court of Justice to conduct an incidental review of national preparatory measures in light of both EU and national law compromises the ability of affected parties to challenge the entirety of the decision-making process, which would otherwise be indispensable to ensuring full observance of their right to effective judicial protection. Effective judicial protection is not, however, the sole purpose, or even the dominant purpose, of judicial control by EU courts. The protection of individual rights is only one of the purposes, i.e. its subjective purpose, of judicial control, in general. An additional objective purpose can, however, be identifieda judicial check on whether the exercise of power is in conformity with the applicable legal requirements. 85 As Barents explains, EU procedural law 'is featured by a rather strong orientation towards objective legality review'. 86 And yet, the unreviewability of violations of national law also fails on this account. Even if an individual's rights are not actually affected as such by a breach of national law in a composite procedure, the fact that no judiciary can review that breach constitutes a failure from the point of view of the judicial control of objective legalitythe enforcement of the binding legal standards to which public authorities are bound. The European Court of Justice's solution to the review of composite decision-making leaves little room for the essential role of courts in suppressing violation of the law by national authorities and repairing the law's integrity.
For this same reason, the unreviewability of national law violations in composite procedures is irreconcilable with any meaningful understanding of the principle of legality in public administration, another requirement of the rule of law. There is little point in having binding legal requirements that limit the power of the national administration if those requirements cannot be judicially enforced. While the principle of legality has traditionally required that the public administration be bound to the sovereign will of the people, as expressed through parliamentary legislation, 87 a common historical trend across European countries has seen an expansion of the legal limitations on administrative power to include other legal sources, such as constitutions, international law sources, and, indeed, regulations adopted by the administration itself. 88 Given that under Article 19(1) TEU the Court of Justice must respect the 'law' as a whole in the interpretation and application of the Treaties, and not just a specific type of legal source, EU law cannot but adopt a similarly demanding and holistic understanding of the principle of legality. 89 Even though the principle of legality in EU administrative law has been mainly developed by the European Court of Justice for the EU's own administration, 90 the non-justiciability of national law does not seem to be any less problematic from the point of view of that principle. Indeed, as the Court of Justice concluded in Hoechst, the principle of legality must be seen as a general principle of EU law because it is recognised 'in all the legal systems of the member states' that 'any intervention by the public authorities in the sphere of private activities of any person, whether natural or legal, must have a legal basis and be justified on the grounds laid down by law'. 91 The fact that no courtnational or Europeancan review compliance with national law in most composite decision-making means that the age-old demand of subordination of public administration to the law cannot be enforced.

A ,   ?
It is generally up to the member states, in the exercise of their national procedural autonomy, to define the procedural rules that are required to implement EU law. 92 This means they can also add procedural requirements to the national phases of composite procedures beyond those foreseen in EU legislation. 93 Indeed, as di Pretis has pointed out, composite administrative procedures are by definition incomplete procedures, in that the EU legislation that establishes them never does so exhaustively. 94 Wherever that EU legislation is silent on particular aspects of the national phase of the procedure, it is governed by the general administrative law regime of the member state or by national legislation specifically passed for the relevant composite procedure.
Member states could, for instance, determine that the national competent authority in a composite procedure is required to consult an advisory body before finalising the preparatory acts it will forward to the EU administration. In fact, Regulation No. 708/2007 of 11 June 2007, which, as noted above, governs the composite procedure for the authorisation of aquaculture facilities, explicitly accords member states the discretion to designate an advisory committee to assist the national authority involved in the decision-making process. 95 It is therefore possible for member states to require that the national competent authority request an opinion from the advisory committeeor indeed oblige it to adopt a preparatory act that is in conformity with that opinion. Yet creating requirements of this kind is a purely domestic legislative choice, and violation thereof would accordingly be a violation of a procedural requirement of purely national law. For the reasons set out previously, such a violation would not be reviewable. Neither would the violation of national provisions that establish any other legal requirements beyond what is demanded by EU law be reviewablee.g. rules defining the necessary quorum for national administrative bodies to take decisions, legal impediments requiring its members to refrain from partaking in the deliberations due to conflicts of interest, mechanisms for the participation of affected parties and civil society, in general.
Even so, a few potential exceptions to the unreviewability of national law violations in composite procedures can be identified. One such exception can be found in the Single Supervisory Mechanism, from which the Berlusconi judgment originated, as a result of the specificities of the legislative framework by which it is governed. One of the peculiarities of the Single Supervisory Mechanism is that it confers on the European Central Bank an unprecedented mandate not only to apply all relevant EU law but 'where this Union law is composed of Directives, the national legislation transposing those Directives'. 96 It has been suggested that the European Central Bank's mandate to apply national law extends to national rules on administrative procedure to which national supervisory banking authorities are generally bound, 97 and perhaps even to those found in general administrative law codes. While the Advocate General in Berlusconi considered that the European Central Bank's new mandate implies an 'extension of jurisdiction' of the Court of Justice to apply and interpret national law transposing the EU's directives in banking law, 98 and although there have been a few judgments suggesting that the General Court supports that view, 99 three issues arise.
First, it is not entirely obvious that just any provision of general administrative law, i.e. one typically intended to be followed by national administrations in whatever area of public law they might be involvedbe it environmental protection, banking supervisions, social security, or food safetyfalls neatly into the category of 'national law transposing Directives' in the banking supervision sector. Second, even if it did, it is by no means evident that the fact that EU secondary legislation grants an EU authority the competence to apply national law automatically means that the boundaries of the European Court of Justice's jurisdiction as defined in primary law should be extended to the application of national law as well. Third, and crucially, even if one were to accept that the European Court of Justice's lack of jurisdiction to apply national law could be expanded by secondary legislation in specific circumstances, i.e. in the event that national law transposes directives, that would still not resolve the issue of how to review a violation of national law that does not transpose directives. Accordingly, in all scenarios, the European Central Bank's mandate to apply national law cannot entirely compensate for the deficiencies in the EU's system of judicial review for ensuring compliance with national law in composite decision-making. In any event, Article 4(3) SSMR is quite unique and has no equivalent in any other EU legislative instrument establishing composite procedures in other policy areas.
A second exception to the unreviewability of national law concerns those principles of administrative procedure that are common to both national and EU administrative law. 100 For instance, violation of the right to be heard by national authorities, if not compensated by the EU administration at any subsequent procedural stage, has been often recognised by the European Court of Justice as a reason to annul the final decision in a composite procedure. 101 Procedural rights of this kind exist both in national administrative law and in EU administrative law, which is why, if they are violated at national procedural phases, EU courts will strike down a final decision taken at the EU level. 102 Nevertheless, even though certain procedural rights principles are enshrined in both national and EU administrative law, a member state's laws could conceivably offer a more generous standard of protection for those rights than EU law does. It is difficult to see how national standards could then be enforced by the European Court of Justice. Indeed, the Court of Justice insists that even when national law offers a higher level of protection of fundamental rights than the Chartersuch as the administrative procedural rights enshrined in Article 41 of the Charterthe standards of national law may not be applied if that would undermine the primacy, effectiveness, and unity of EU law. 103 The demand that the unity of EU law must not be compromised is, in effect, the same as the requirement of uniformity of EU law. That requirement militates against the European Court of Justice reviewing compliance with national administrative law requirements, which inevitably vary across the member states. The right to be heard, for instance, is recognised in EU law; certain member states might impose more demanding legal requirements on the administration than others, in that respect. There might be significant variation in terms of the parties entitled to be heard, whether said parties have a right to an oral hearing, or whether they are entitled to comment on both the facts upon which the administration will base its decision as well as the draft of the final decision. If such varying standards of procedural protection had to be enforced by the Union courts, then identical composite administrative procedures would potentially yield EU decisions that are either valid or invalid, depending on the law of the member state in which the procedure was initiated. If the same European administrative procedure yields valid or invalid decisions depending on the member state concerned, said procedure is unsuitable for ensuring the uniform implementation of EU law across the European Union.
There is a third and final category of potentially reviewable violations of national law, i.e. gross violations of national legal provisions that define which specific administrative body has competence during the national phase of a composite procedure. Some case law suggests that the European Court of Justice could declare a final decision taken at the EU level illegal if it was based on a preparatory act taken by an authority that manifestly lacked competence, in accordance with national law, to adopt it. The case law in question emerged from the EU's legislation on blacklisting, one of its common instruments in the fight against terrorism. Common Position 2001/931/CFSP and Regulation (EC) No. 2580/2001 create a composite procedure aimed at the blacklisting of individuals and groups suspected of supporting terrorist activities. If there is serious evidence of such support, national competent authorities can decide to freeze any funds owned by the targeted persons or groups. At the following stage, the Council, based on the proposal of the national competent authorities, adopts a regulation blacklisting those persons and groups and determining that their assets will be frozen throughout the EU. Some targeted organisations have on occasion pleaded before EU courts that the national body involved in the national stage of the blacklisting procedure was not the competent national authority. In one such case, the General Court did not hesitate to rule that a decision taken by the United Kingdom's Home Secretary constituted, 'in light of the relevant national legislation', a decision of a 'competent national authority'. 104 If the Union courts are able to assess whether the authority involved in the national phase of a composite procedure is the competent authority, then perhaps they can also conclude the opposite, thus ruling the entire national phase illegal.
A certain parallel can be drawn here with the conditions for the admissibility of preliminary references under Article 267 TFEU. Under that Treaty provision, the Court of Justice may only accept references for a preliminary ruling made by a body qualifying as a 'court or tribunal of a member state'. What precisely counts as a 'court or tribunal' depends on a number of criteria, 'such as whether the body is established by law, whether it is permanent, whether its jurisdiction is compulsory, whether its procedure is inter partes, whether it applies rules of law and whether it is independent'. 105 Even though the Court of Justice insists that this is a matter 'governed by EU law alone', it is evidently incapable of determining whether a court or tribunal is established by law, or indeed whether any of the other criteria are fulfilled, without interpreting the national provisions governing that court or tribunal. 106 Despite the three (merely potential) exceptions set out in the present section, violations of national law in composite procedures remain largely unreviewable. None of the potential exceptions identified are capable of preserving the rule of law at the national level in any meaningful way. The first exception, because it only applies to a few of possibly hundreds of composite administrative procedures. The second exception, because it in fact only allows the review of violations of national procedural standards to the extent that they offer the same degree of protectionnot more, not lessas the equivalent procedural standard in EU administrative law. The third exception, because its scope is likely limited to quite rare instances of manifest incompetence on the part of the national authority that authored the preparatory acts in a composite procedure. The three possible exceptions might attenuate the rule of law gap at the national level caused by composite decision-making, but they certainly do not solve it.

C
In EU law, the rule of law can be understood as a cluster of principles. Effective judicial protection, the subordination of public power to judicial control, and legality are certainly part of it. Principles are, by their nature, optimisation requirementsi.e. norms that require a certain ideal or value to be protected to the greatest degree factually and legally possible. 107 The boundaries of that legal possibility are contingent upon the specificities of each particular legal order. In the EU legal order, the legal possibilities for ensuring the constitutional principle of the rule of law in composite decision-making are inevitably constrained by other constitutional requirements. Those requirements are brought about by the strict constitutional divide between the jurisdiction of national and EU courts, as well as by EU constitutional principles of an existential and definitional naturethe requirements of the uniformity and autonomy of the EU legal order.
Even though the rule of law has been the bedrock of administrative law for as long as it has existed, 108 it is nonetheless hard to see how it can be fully observed in the law of composite procedures without making adjustments for other crucial EU constitutional principles. The sacrifice of the rule of law at national level that results from such adjustments is certainly dysfunctional and regrettable; it is also an unavoidable result of the fact that the Treaties are simply not designed to accommodate such forms of mixed administrative decision-making. Opinions naturally differ as to just how serious this rule of law gap is, depending on whether one believes that, under EU law, legal constraints on the power of the national administration must be taken seriously, regardless of whether they originate in national or in EU law. The view taken here is that, as a matter of principle, they must. One cannot simultaneously hold the rule of law to be a foundational principle of the European Union while maintaining that EU law is indifferent to whether national law is actually observed.
It is not that EU courts have failed to offer solutions for the judicial review of composite decision-making that are consistent with EU constitutional law. Indeed, as Vermeule has recently suggested in regard to American administrative law, the correct working out of constitutional doctrines and principles can sometimes be unfavourable to the rule of law's grip on public administration. 109 It is the institutional and administrative reality of the EU that no longer fits in with its own constitutional order.
Unlike most national systems of administrative justice, which are detailed in ordinary legislation, that of the European Union is constitutionally entrenched in the Treaties. The implication is that it cannot be easily reformed. Instead, every change requires the approval of all EU member states according to their own constitutional procedures for the ratification of international treaties. Some have suggested, for example, that issues of judicial protection in composite decisionmaking could be addressed by the introduction of 'reverse preliminary references', presumably through Treaty reform, which would enable the Court of Justice to  seek guidance from national courts as to how to interpret and apply the national law provisions that are relevant for the review of composite procedures. 110 Yet the reconstruction of primary law required to completely fill the rule of law gap examined in this article would have to reach far deeper. Without reforms that somehow create exceptions to the autonomy of EU acts vis-à-vis national law and give EU courts jurisdiction to apply national law, it is difficult to imagine an outcome implying anything other than the unreviewability of national law violations. It is nonetheless also difficult to imagine that the member states would agree to any of these reforms anytime soon. We may in the end have to simply adjust our expectations with respect to the degree to which the rule of law can be realised in European administration under EU constitutional law as it now stands. There may be exceptions, as this paper has sought to explain. Certain strategies could perhaps even mitigate the rule of law gap without Treaty change. One could be for the EU legislation that establishes a composite procedure to regulate national procedural stages much finer detail, so that those stages are governed, to the greatest extent possible, by legal provisions enforceable by EU courts. It is, however, difficult to predict whether it would be politically feasible to have member states systematically replace their own administrative law with EU legislation. Indeed, the creation of such extensive administrative law regimes by the EU legislature would come at the risk of constituting a harmonisation of national administrative law, something for which the EU lacks a legal basis. 111 One other strategy might entail the reappraisal of whether the shortcomings of the EU's judicial system could be partly mitigated with non-judicial remedies, 112 e.g. the creation of joint review boards composed of both national and EU officials drawn from the administrative authorities involved in a given composite procedure. If the European Central Bank can be given a mandate to apply national law, then perhaps the aforementioned review boards could be given a similar mandate, thus allowing them to apply the national administrative law that the European Court of Justice is unable to enforce. That would nevertheless beg the question of how to judicially review decisions taken by said review boards.
In any event, until significant reforms are carried out either at the Treaty level or by means of secondary legislation, the European Court of Justice will inevitably be faced with a stark choice. It can either take the unprecedentedly bold step of claiming jurisdiction to enforce national law in order to preserve the rule of law in the member states' administrations, or it can respect the constitutional boundaries of its jurisdiction, the principle of uniformity, and EU law's existential claim to autonomy. It cannot do both.