Understanding National Remedies and the Principle of National Procedural Autonomy: A Constitutional Approach

Abstract This article provides a constitutionally grounded understanding of the vexing principle of ‘national procedural autonomy’ that haunts the vindication of EU law in national court. After identifying tensions and confusion in the debate surrounding this purported principle of ‘autonomy’, the Article turns to the foundational text and structure of Union law to reconstruct the proper constitutional basis for deploying or supplanting national procedures and remedies. It further argues that much of the case law of the Court of Justice of the European Union may be considered through the lens of ‘prudential avoidance’, ie the decision to avoid difficult constitutional questions surrounding the principle of conferral. As the last Part shows, a constitutional understanding of ‘national procedural authority’—not ‘autonomy’—helps clear up some persistent puzzles, and provides critical guidance for when deference to national procedures and remedies is appropriate, and when such deference is misplaced. Comparative references inform the argument along the way.


I. INTRODUCTION
For the most part, the application of EU law takes place in the courts of the Member States. This reflects the character of the European Union as a 'vertical' system, in which the union government generally relies on the resources and personnel of its component states. This contrasts with the more 'horizontal' federalism of, say, the United States, in which each level of government has the basic capacity to legislate, execute, and adjudicate its own policies and laws. 1 Just as the European Union mostly relies on national legislatures to pass implementing legislation, and on national executives to carry out its policies, it mostly relies on national judiciaries national courts. The constitutional approach pursued here establishes the normative foundations of the principle of national procedural autonomy (or 'national procedural authority', as we shall call it here), uncovers systematic features of what might otherwise appear as the confusing and inconsistent caselaw of the CJEU, and helps us understand better what work the principle currently does and properly ought to do.

II. THE DEBATE ABOUT NATIONAL PROCEDURAL AUTONOMY
There seems to be continued confusion over the principle of national procedural autonomy, and whether it simply suggests a kind of default in the absence of EU action or a stronger principle of national autonomy that EU law must heed. The original articulation of the principle, especially in light of foundational decisions of EU law, seems to support only a permissive default rule that automatically gives way to positive EU legislation. All the same, the principle continues to be invoked more stringently as prescribing temperance when it comes to the harmonisation of national procedures and remedies.
A. The Elusive Origin of National Procedural Autonomy -'Permissive' or 'Prescriptive'?
Traced back to the Rewe case of 1976, and repeated ever since, the principle of national procedural autonomy comes with a prefatory condition and two closing qualifications: In the absence of Community rules on this subject, it is for the domestic legal system of each Member State to designate the courts having jurisdiction and to determine the procedural conditions governing actions at law intended to ensure the protection of the rights which citizens have from the direct effect of community law, it being understood that such conditions cannot be less favourable than those relating to similar actions of a domestic nature. … The position would be different only if the [Member State's procedural] conditions . . . made it impossible in practice to exercise the rights which the national courts are obliged to protect. 7 The prefatory condition that national procedural rules govern '[i]n the absence of Community rules' may be a source of confusion, as it could point in opposite directions. First, it may narrowly suggest that Member States enjoy procedural authority only by default, ie only as long as the Community (now Union) has not made rules to harmonise the relevant procedural law. Put another way, on this view, Member States merely enjoy a 'presumption' of procedural authority. 8 Beyond that presumption, however, Member States have no real 'autonomy' at all, but only such national procedural discretion as the Union from time to time chooses to permit. Let us refer to this as 'presumptive' or, better, 'permissive' national procedural autonomy. Second, however, the prefatory clause may be taken to suggest more broadly that the Community (now Union) does not have the competence to make rules on procedural matters and that, therefore, primary Union law necessarily and constitutionally preserves Member States' freedom over procedure (unless and until the foundational Treaties were to be amended). This view suggests that Union rules on procedural matters would be ultra vires of the Treaties and hence unconstitutional under Union law. Put another way, on this second view, the 'absence' of Union rules on the subject is taken to flow not from any policy decision on the part of the Union legislator, but from the constitutional limitations prescribed by the Treaties. Member States' freedom over procedure is thus not merely permitted by Union legislation, but constitutionally prescribed by the primary law of the Treaties themselves. For the sake of clarity and to distinguish it from the first variety, we shall call this 'constitutional' or, better, 'prescriptive' national procedural autonomy. This latter, constitutionally prescribed national procedural discretion seems most compatible with the broad national procedural 'autonomy' that the principle's label seems to promise.
It does not take long to figure out, however, that the language in Rewe (and successor cases) does not support a broad version of constitutionally prescribed 'autonomy', if by that we mean complete Member State freedom over procedural rules. Following Constantinos Kakouris, we may, for instance, observe that Rewe's prefatory clause ('[i]n the absence of Community rules') differs significantly from the prefatory clause the Court consistently uses in other domains (eg 'as Community law currently stands' or 'at the present stage of development of Community law') to signal absence of constitutional authority, suggesting that the Court in Rewe was not signalling any Treaty limitation on Community competences, but merely the politically contingent absence of relevant harmonising rules. 9 Whenever the Court intends to signal hard, constitutional limitations on the Union's competence under the Treaties, it tends to use one of the latter formulations, implicitly including the Treaties in the phrase 'Community law'. In these latter constitutional formulations ('as Community law currently stands' or 'at the present stage of development of Community law'), the phrases 'currently stands' or 'present development' thus signal the current stage or development of the foundational treaties of the Community (and Union), not merely political choices of secondary legislative rules or administrative decisions. This reading of the case suggests that Rewe was merely stating what follows if secondary legislation happens to be silent on the matter of procedures. Rewe thus never intended to signal that the Union was constitutionally prohibited from making procedural rules for Member State courts.
Nor could it be any other way. Rewe's two closing qualifications (that Member State procedural rules not be 'less favourable than those relating to similar actions of a domestic nature', and that Member State procedural rules not render the enjoyment of Community rights 'practically impossible') place Member State procedures squarely within the scope of Community law. It would thus be 'giving with one hand while taking with the other', to read the prefatory clause as suggesting Member State procedures are beyond the Union's competence to regulate, when in the next breath Rewe adds two closing qualifications that national courts must nevertheless heed.
To be sure, the twin qualifications of equivalence and effectiveness, as they are called, do not necessarily imply free-standing legislative competence over national judicial procedure at the Union level. But they do-necessarily-suggest that Member State procedural law is subject to Union law, the interpretation of which is controlled by the Court of Justice. In that sense, Rewe cannot be understood as supporting any hard autonomy, at least not in the sense of complete freedom from Union control.
This conclusion should not come as a surprise. We can recall the early insight of Koen Lenaerts that there is 'no nucleus of sovereignty that the Member States can invoke, as such, against the Community'. 10 Similarly, we can say here: there is no black hole in the universe of Union law for Member State procedures. Member State procedural law, as such, is not immune from Union control. 11 B. Primacy Over Autonomythe 'Trinity of Legal Normativity' But there is more. Pushing aside national procedural autonomy in favour of EU primacy, one might note that Rewe's canonical formulation explicitly grounds the requirement of effectiveness in the basic obligation to safeguard 'the rights which the national courts are obliged to protect'. 12 Although this phrase in Rewe comes without quotation marks or citation, it invokes the primacy of Community law first formulated in Van Gend. Indeed, in that first landmark case the Court of Justice had already cut through national rules on the incorporation of treaties into national law, spoken directly to Member State courts, and told national judges in no uncertain terms that Community law contained 'rights which national courts must protect'. 13 As the trilogy of Costa, 14 Simmenthal, 15 and Factortame 16 later clarified, this primacy of Community law and the direct effect in national courts spans all aspects of Union law, or what we may understand as the trinity of legal normativity, ie substance, procedure, and remedy.
To be sure, one might have imagined a system in which primacy was solely a substantive conflicts-of-law tiebreaker without direct effect. One might also have imagined a narrow kind of primacy and direct effect with a highly formalised distinction between substance (where primacy and direct effect rule) and procedure (where they do not), or even between substance and remedy, on the one hand, (where primacy and direct effect rule) and procedure, on the other, (where they do not). But the Court pursued a more comprehensive, practical, and constitutional vision, one in which primacy and direct effect extended broadly, including implied powers, 17 to cover all three normative aspects of Union law: substance, procedure, and remedy.
Costa drove home that Van Gend's primacy makes its way straight into Member States' legal systems in terms of the full substantive vindication of EU law, ie that EU law is 'an integral part of the legal systems of the Member States and which their courts are bound to apply'. 18 Simmenthal specifically reiterated that point as to Member State procedural law, even where the countervailing national procedure is grounded in a Member State's constitution. In particular, the Court in that case demanded that, where necessary to preserve the effectiveness of Community law, lower courts must ignore their national procedure that would have them refer a matter to their own constitutional court before setting aside nonconforming national law. Factortame, ultimately, did the same for remedies, declaring that Member State courts must go beyond their powers established under national (constitutional) law and issue an interim injunction suspending even the application of an act of Parliament to prevent irreparable harm. Having established primacy in substance, procedure, and remedy, the trinity of the Union's legal normativity would seem complete.
But what about national procedural autonomy? Some say Van Gend and its progeny indeed settled matters conclusively in favour of the primacy of Union law. For Constantinos Kakouris and Michal Bobek, for example, primacy and direct effect are enough to put any claim of national procedural autonomy to rest. To them, procedure and remedy are ancillary to substance, and that ends matters because substantive competence simply brings procedural and remedial competence as well (even if only as a matter of judicial competence for the Court of Justice to decide what procedures and remedies are necessary to make substantive Union law work). Kakouris takes it even a step further in that he sees all Member States' procedural law as Union law whenever national procedures apply to substantive Union law. The idea is that in the application of substantive Union law, national procedural rules are implicitly incorporated into Union law and hence operate as whenever Union law has not provided otherwise. 19 Bobek, in turn, agrees that Union law ultimately controls all procedural and remedial implementation of substantive Union norms, and concludes that Rewe's reference to equivalence should therefore be discarded as useless (and also hopeless). 20 In his view, Member State procedures and remedies should be accepted or set aside based exclusively on an assessment of effectiveness. 21 Put another way, Kakouris and Bobek both ultimately see 'national procedural autonomy', to the extent it exists, exclusively as a choice on the part of the Union legislature and Court of Justice to let Member State procedures stand. That is, they see 'national procedural autonomy', to the extent it exists at all, solely as what we earlier called permissive national authority. In their vision, it seems there is no place for any constitutionally mandated national procedural autonomy, i.e., prescriptive national autonomy does not exist.

C. Primacy and Autonomy -Balancing
Not so fast, argues Sasha Prechal. There is no contradiction, she submits, between Union law primacy and Member State procedural autonomy. 22 While primacy provides a rule of conflict resolution when two substantive rules conflict, it need not control the procedures that govern the application of substantive law in court. 23 This follows, says Prechal, from the incorporation of Union law into Member State legal systems after Costa: [O]nce it is accepted that Community law provisions are a part of the valid law in the Member States, they should in principle be handled in the same way as national law. Then it is difficult to understand why certain rules, such as [national] procedural provisions, should be dismissed with the sole argument that they hamper the full application or direct effect of Community law; after all, they hamper in the same way the application of national law, even, where appropriate, of the Constitution. 24 Rather than dismissing national procedural or remedial constraints whenever they curtail the effectiveness of Union law, she sees the Court of Justice moving (albeit inconsistently) toward a kind of 'rule of reason'. This analysis, properly in her 19 Kakouris, note 9 above. 20 Hopeless because the comparison for this non-discrimination prong is impossible to administer. See Bobek, note 6 above. One wonders, though, pace my good friend Michal, why comparisons are any more difficult here than they are in run-of-the-mill discrimination cases or, indeed, in indirect discrimination cases. For an excellent sorting out of difficult comparisons in the context of discrimination, see MB v Secretary of State for Work and Pensions, C-451/16, EU:C:2018:492 (Opinion of AG Bobek). 21 Bobek does concede that some restraint should govern the harmonization of Member State procedure, but he does not specify the normative basis for that restraint. 22 Prechal, note 6 above, p 685. 23 Prechal dutifully nods toward the fact that the distinction between substance on the one hand and procedures and remedies on the other is notoriously porous, and that in cases in which the procedure or remedy is deemed akin to substance the supremacy rule kicks in. Ibid, p 685 n 16. 24 Ibid, p 686.
view, conducts a 'balancing exercise between the interests which are served by the national rules at issue and the effectiveness of Community law'. 25 Koen Lenaerts more recently suggested much the same. 26 Lenaerts follows Michael Dougan's reading of the cases, which finds that an initial period of CJEU restraint led to a second period of more aggressively displacing national procedural rules, and finally to a third period of applying an 'objective justification model'. 27 Takis Tridimas agrees. 28 In particular, Dougan reads the Court as asking (1) whether the Union right is adversely affected by the national procedure, (2) whether the national procedure furthers a legitimate aim, and (3) whether the national procedure does so in a legitimate manner, consistent with the margin of discretion left to the Member States by Union law. Lenaerts puts the currently governing approach similarly, but with an equal focus on both sides of the equation-ie with an equal focus on both Union as well as Member State law: 'In cooperation with national courts, the European Court of Justice ('ECJ') endeavours to understand and protect the legitimate interests embodied in the national rules of procedure whilst seeking to ensure a sufficient level of judicial protection for EU rights'. 29 Whether we call the current approach a 'rule of reason', 30 an 'objective justification approach', 31 or 'selective deference', 32 some kind of balancing is presented as a sensible practical compromise between two equal and opposing normative principles -primacy of EU law, on the one hand, and Member State procedural autonomy, on the other. 33 High constitutional theory might interpret this along German lines of 'concordance', as Robert Alexy suggested in the case of fundamental rights, ie as a way of optimising two principles that clash. 34 For any such balancing to be performed, however, both principles must firmly exist in the first place. Even when performing such balancing merely to give effect to a presumption of national authority in the absence of EU law rules to the contrary, that presumption in favour of national authority must itself be normatively grounded. What, then, of the position that the only principle at stake is primacy, and that the principle of national procedural autonomy simply does not exist? Should we subscribe to that view? Or should we go along, instead, as Prechal suggests, with Member State procedural and remedial limitations that hamper the effectiveness of EU law-an autonomous source of law-just because Member State procedures similarly limit the enforcement of their own law? Do we not need more than an 25  equal amount of harm to national law before crediting the national side of the balance? Why should EU law care that national law has so limited itself? To be sure, we can always understand the balancing approach as a vague and diplomatically useful way for the CJEU to make friends with Member State courts. 35 But is there any normative foundation to national procedural autonomy? To put the question more precisely, is there a constitutional justification for claims to national procedural autonomy in EU law itself?

III. THE CONSTITUTIONAL FOUNDATIONS OF NATIONAL PROCEDURAL AUTHORITY
Returning to first principles, we can reconstruct a constitutional principle of deference to Member State procedures and remedies in the administration and vindication of EU law. We see traces of this principle in the institutional architecture of the Union, as well as in the positive text of the foundational Treaties. Both structure and text support a principle of deference to national procedural and remedial rules that goes well beyond a mere 'presumption' in favour of Member State rules. The principle thus treads beyond the 'permissive', and into the 'prescriptive' domain. Nonetheless, the principle of 'national procedural authority' outlined here falls short of any 'autonomy' in the strict sense of the latter term.
A. The Normative Implications of Structure and the Paradox of Judicial Federalism Structure carries normative implications. Sometimes structure even carries multiple normative implications that can be in tension with one another. As we shall see, while the principle of national procedural autonomy is not spelled out in the Treaties, something like it does reasonably emerge from the structure of the Union. In particular, the dédoublement fonctionnel 36 that turns national courts into Union courts has an institutional and normative flipside. Even when they function as the Union's courts, Member State courts retain their basic organisational authority as courts in virtue of Member State-not Union-law. Member State law sets up and organises Member State courts based on the organic power derived from each Member State's constitution. This means the institutional infrastructure of a national court, even when applying Union law, remains a creature of that Member State, not of the Union. As the CJEU noted in an important case on preserving the rule of law in Poland: [ Put simply, Union law 'piggybacks' (to use a highly technical term) on the existing institutional infrastructure of the Member States. One might even say that this is true throughout the Union's functions and the national institutions that carry them outthus broadly reflecting the vertical structure of the system. In any event, it is surely true for the judiciary, where the Union merely borrows national courts for Union purposes.
In one sense, this structure resembles that of the United States, in which state courts are constitutionally bound to carry out federal law as their own. In the United States, the Supremacy Clause, which specifies both the supremacy and direct applicability of federal law in state courts, inexorably fuses federal and state law with supremacy for the former, similar to the way the CJEU's caselaw has done in the EU. Indeed, federal and state law are integrated even more in the US, where there is no space for the kind of constitutional pluralism that allows a component state to view union law through the lens of its own, component state constitution. At least since the Civil War and the Reconstruction Amendments, the US component state polities are constituted politically by the citizens of the United States, which-in contrast to the EU-gives the polity of the union a foundational status on which the polities of the several states are built. 38 Even in this more tightly formed union, the US Supreme Court has emphasised, much as its European counterpart has, that a state court applying federal law 'does not, while performing that duty, derive its authority as a court from the United States, but from the State'. 39 As a result, the touchstone in the United States for what federal law demands of state courts has long been that 'federal law takes state courts as it finds them'. 40 Thus, even in the United States, state courts applying federal law are seen as creatures of state law, and federal law does not simply remake state courts to its liking.
The limits of what US federal law can ultimately demand of its state courts are notoriously uncertain, and well beyond what we need to discuss here. 41 But one thing is clear: while the US Congress has no direct power to create or even reimagine state courts, the US Congress does have the express power-and has long exercised 37 Commission v Poland (Independence of Supreme Court), C-619/18, EU:C:2019:531, para 52. 38   the power-to create lower federal courts with broad jurisdiction to hear all claims of federal law whenever Congress 'deems' state courts will not suffice. Therefore, whenever state courts assert, say, a 'valid excuse' for refusing to hear a particular federal claim, 42 that federal law claim can always be brought in federal court. In the United States, then, the question of what federal law demands of state courts mostly turns into a question of deciding what should (and what Congress intended) to be decided in federal, as opposed to state, court. Not so in the European Union. Here, at least for the purposes relevant to our present inquiry, the judicial structure of the federal system is more radically decentralised. To be sure, the CJEU by express Treaty provision controls the interpretation of EU law, 43 and even has a formal legal monopoly over declaring acts of EU institutions to be in violation of the Treaty. 44 As far as real-world private plaintiffs are concerned, however, the gatekeepers of EU law are, for the most part, the Member State courts. When a Member State court declines to hear, or otherwise refuses to adjudicate, an EU law claim (and refuses to refer relevant questions to the CJEU), the plaintiff usually has no recourse other than an appeal or fresh action within that same Member State's system of courts. 45 Even after Köbler's extraordinary decision to allow an action in a lower Member State court to check the judgment of that same Member State's high court, 46 the claimant in such an action is still operating within that Member State's judiciary. To be sure, the Treaties allow for the creation of 'specialised courts' at the EU level where plaintiffs may access the EU's own courts directly. But the Treaties allow for these specialised courts only for 'certain' actions and, it would seem, only for 'specific areas' of the law. 47 We might yet witness a fundamental shift, and this authorisation could be interpreted more expansively in the future. But, so far, the Parliament and Council have availed themselves of this power only once, and only in the most traditional and limited way, ie by creating a Civil Service Tribunal for disputes involving the EU civil service (and then abolishing that court to move those disputes to the EU's General Court). 48 This seems to be close to a current constitutional convention or settlement. There is no sign that the EU legislative bodies are actively considering vastly expanding the jurisdiction of first instance courts at the EU level to match the jurisdiction of lower-level federal courts in the United States. There is not even a sense the EU legislature believes it has the power to do so. 42 Seinfeld, note 41 above. 43  Under these circumstances, then, the stakes in the EU for determining the duties of state courts are even higher than in the United States. For the most part, the quest in the European Union for private plaintiffs boils down to doing things properly in Member State courts, or doing them not at all. Individuals (and other private parties) have direct access to the EU's own courts in Luxembourg only when suing EU institutions and only under highly restrictive standing rules. 49 For all their other claims based on EU law, it is national courts all the way down. Put simply, private litigants suing anyone other than the EU itself will not get to an EU court without the assistance of some national court. To be sure, the Commission or another Member State may bring an action against the Member State for infringing the Treaties by failing to obey or enforce EU law, and the Commission and others can initiate an Article 7 proceeding 50 , or perhaps even withhold financial assistance. 51 But, again, unless the claim is against the Union itself, private litigants are limited to bringing their EU law claim in the courts of the Member States.
Returning to the normative implications of structure, and just briefly to the transatlantic comparison, this highlights the apparent paradox of judicial federalism as we compare the European Union with the United States. In both systems, component state courts are creatures of component state law, deriving their organisation and authority as courts from the component states themselves. But while the European Union is generally a less tightly consolidated federal union than the United States, the EU places more demands on its component state courts to make the overall system work.
Of course, one person's paradox is another's hydraulic connection. One might say that the EU places more demands on its component state courts than the U.S. precisely because the EU depends more heavily on Member State courts to maintain an overall complete system of judicial enforcement. Here as elsewhere, what the EU lacks in organisational infrastructure at the Union level, it must make up for by cajoling and conscripting the Member States to act on its behalf. This is, after all, the idea of a vertical system of federalism in Europe. Whereas in the United States, federal law can always, if necessary, be vindicated in federal court, the more radical decentralisation of judicial authority in the European Union means that a complete system of judicial enforcement can only be achieved in the EU by an unyielding obligation on Member State courts to implement Union law. The bottom line: as far as individual rights against anyone other than EU institutions themselves are concerned, in the European Union, the rule of law and the system of judicial enforcement are only as strong as Member State courts make it. 49 Ie, a suit against the civil service or a direct suit against the EU institutions under the highly restrictive standing rules of Article 263 TFEU. 50

B. The Positive Foundations of National Procedural Authority
Just as we can see both sides of the balance between EU primacy and national procedural authority in the deep structure of EU law, we see traces of both sides in the text of the Treaty. The tension between EU law primacy and effectiveness, on the one hand, and national procedural autonomy, on the other, is reflected here as well.
On the one hand, the EU treaties have come to gesture toward primacy and articulate quite strongly the obligation of Member State courts to bring about a complete system of judicial enforcement. For example, by now, the Treaties have codified-if only by way of non-binding declaration-the CJEU's holdings on the 'primacy' of EU law, including Costa's statement concerning primacy over all domestic legal provisions: [T]he law stemming from the [T]reaty, an independent source of law, could not, because of its special and original nature, be overridden by domestic legal provisions, however framed, without being deprived of its character as Community law and without the legal basis of the Community itself being called into question. 52 The Treaty Declaration incorporating Costa's insistence on primacy over all national rules 'however framed', thus suggests an all-encompassing primacy that would seem to extend to the entire trinity of normativity-from substance to procedure and remedy. Furthermore, since Lisbon, the Treaties make explicit-in Article 47 of the Charter of Fundamental Rights-that individuals have a right to an 'effective remedy' to vindicate their EU rights in court. The Treaty of Lisbon includes a corresponding provision-Article 19(1) of the Treaty on European Union-which provides that the CJEU's usual role 'to ensure that in the interpretation and application of the Treaties the law is observed' is expressly complemented by the Member States' duty to 'provide remedies sufficient to ensure effective legal protection in the fields covered by Union law'. 53 These provisions, in turn, expand on earlier, more general provisions, such as the duty of sincere cooperation, 54 as well as the judicially developed principle of effectiveness. 55 52 Declaration 17 Concerning Primacy, Treaty of Lisbon (quoting Costa, note 14 above) (emphasis added); cf Art I-6, Draft Constitutional Treaty. 53 Art 19(1) TEU. 54  Returning to Prechal just briefly, it would thus seem to follow, based on these provisions alone, that Member State rules of procedure should indeed give way to the effective enforcement of an EU right, regardless of whether that Member State enforces similar domestic rights less effectively. These provisions would therefore seem to support the Kakouris/Bobek view that the primacy of EU law is to be made effective and complete, no matter what.
On the other hand, however, we also find constraints in the Treaties to suggest reservation on this score. We need not delve deeply into the important and pervasive feature of constitutional pluralism, 56 whereby the CJEU must always attend to what you might call 'the view from the other side'. On that pluralist structural account of the Union, Member States retain good cause to view the EU and its powers through the lens of each of their own constitutions. But even without actively considering constitutional pluralism, i.e., even purely from the point of view of EU law itself, there are limitations to any claim of unvarnished EU supremacy. Even purely from the point of view of positive EU law itself, it cannot be maintained that Union law pays no heed to Member State interests in their own procedures and remedies.
On the Member State side of the ledger, the treaties do not confer any direct, general legislative power on the Union to determine Member State procedural law. To be sure, under the doctrine of implied powers, the CJEU has held that the scope of EU law is to be understood in functional terms, where demands of procedure and remedies (civil or criminal), may flow from the enumeration of substantive competences. 57 And properly so. Procedures and remedies must be entailed by substance at least in part, or else the trinity of legal normativity-and hence the normativity of substantive law itself-would be incomplete. And yet, regulating the Member States' institutional organisation, such as judicial procedure, is nonetheless recognised as involving a notionally distinct exercise of power. This is evidenced, for example, by the express conferral on the Union of carefully circumscribed authority to regulate certain aspects of civil and criminal procedure for the purpose of enhancing judicial cooperation in matters with cross-border implications. 58 Procedural and remedial demands, therefore, must be considered an additional step of harmonisation beyond that of harmonising the underlying substantive law itself, and thus trigger review under the general background principles of Union law, including subsidiarity, proportionality, and respect for Member States' national constitutional identity. 59  procedural and remedial rules without concern for Member State interests. Indeed, as the Lisbon Treaty has made clear, subsidiarity applies not just to the substantive decision to regulate, but to the 'form of Union action' as well. 60 Neither 'shall … exceed what is necessary to achieve the objectives of the Treaties'. 61 To be sure, the drafters of the Treaty likely had the choice of regulation versus directive foremost in mind as they amended the statement of the principle of subsidiarity to include the 'form of Union action'. But we should not understand the applicability of subsidiarity to 'form' so narrowly. Subsidiarity should apply to both form and substance more broadly understood. After all, what is sauce for the goose is sauce for the gander: if the legal normativity of EU law encompasses substance, procedure, and remedy, the same must go for any restrictions on EU law based on subsidiarity, proportionality, and respect for national constitutional identities. All three aspects of EU law -substance, procedure, and remedy-must be subject to these constitutional constraints as well. Seen from the vantage point of conferral, especially after Lisbon, authority over procedures and remedies might thus be viewed as a kind of 'shared' competence, 62 subject to displacement by the Union under the usual constraints. Given the trinity of legal normativity, the direct conferral of substantive competences of the Union includes the implied conferral of the necessary attendant procedural and remedial competences as well. 63 Under ordinary rules of shared competences, the exercise of Union power does not have the effect of pre-empting all Member State activity in the entire field of such activity, but displaces national rules only 'to the extent' the Union has exercised its own powers. 64 As the Lisbon Treaty's Protocol on shared competences emphasises: 'the scope of this exercise of competence only covers those elements governed by the Union act in question and therefore does not cover the whole area'. 65 Exercising shared competence only regarding certain 'elements' thus anticipates Union regulation of substance while leaving national procedures and remedies in place. This alone precludes jumping from the regulation of substance to the conclusion that the Union is displacing Member State procedural and remedial rules as well. The exercise of the further step beyond substantive regulation into the regulation of procedure and remedies is thus revealed as a separate act of regulation, subject to a separate review for its legislative grounding and constitutional propriety.
Moreover, as the Union is rarely exercising any directly conferred procedural or remedial competences when harmonising national procedures and remedies, 66  instead mostly exercising implied powers based on the Union's substantive competences, extending substantive measures into harmonising procedures and remedies triggers a more careful review of the necessity and propriety of that extension than regarding those expressly authorised substantive measures alone. The Court has never policed the exercise of the EU's competences all that closely, 67 only rarely striking down an EU measure for having overstepped the scope of Union competences. In light of the general reliance on Member State's institutions, and the generally ancillary and hence implied nature of the Union's power over procedure and remedies, however, review for compliance with subsidiarity, proportionality, and respect for national constitutional identities is thus (properly) more stringent here than when it comes to the exercise of expressly conferred competences. This, then, is the normative flipside of structure as positively recognised in the Treaties. Because of the structural dependence of the Union on Member State courts, and because Member State courts receive their authority as courts from Member State law, 68 every procedural demand on Member State courts is itself an imposition on the Member States that must be compatible with the principles of subsidiarity, proportionality, and a proper respect for national constitutional identity. And given that (except for regulation under, say, Articles 81 or 82 TFEU on judicial cooperation in cross-border matters), we are dealing with the exercise of an implied power over procedures and remedies to effectuate a substantive measure, the review under those three constraints tends to be stricter than when considering the underlying substantive measure alone. This both explains and justifies the existence of a 'presumption' in favour of Member State procedures and, more important, the stringency with which it is pursued. It also provides the normative foundation for EU law interventions into those national judicial procedures. In sum, it drives the Court's initial presumption in favour of national procedures, and the so called 'balancing' exercise between EU primacy, on the one hand, and national procedural autonomy, on the other, to ascertain whether the presumption in favour of national procedures and remedies should be overcome.
Notice what we have just done. Having reconstructed the normative foundation of Member States' claim to procedural authority, that normative claim cannot simply vanish in the face of explicit EU law rules. We thus moved from permissive to prescriptive national procedural authority. Both exist.
Recall the opening line of the Rewe test ('In the absence of Community rules on this subject …'.). This says preciously little about what happens in the presence of Community law rules on the subject. It would be a plain logical fallacy to conclude that the inverse of Rewe's conditional statement were true. 69 The Rewe statement does not suggest, for example, that in the presence of Community law rules on the 67 Cf eg Tridimas, note 60 above. 68 Compare Kakouris, note 9 above (relying on Kelsen) with Halberstam, note 38 above (distinguishing Kelsen's unitary conception of federalism from more plural conceptions of Madison, Webster, and Schmitt). 69 Just as 'if x, then y', does not necessarily entail its logical inverse 'if not x, then not y', so, too, the statement 'in the absence of Community rules, national procedures govern' does not necessarily entail subject, any such Community rules inevitably trump all national procedural rules.
Having established the normative foundation of Member States' claim to national procedural authority, that national authority can only be displaced by a superior claim of authority on the part of the Union. Even where a given procedural rule is spelled out in positive EU law, one must therefore inquire-however quickly and with whatever deference to the EU legislature one might do so-whether such EU law displacement of national procedures is constitutionally justified. Therefore, having held substantive EU law valid, the CJEU must ask, separately, whether the further intrusion into Member State procedural and remedial law is justified according to subsidiarity, proportionality, and a proper respect for national constitutional identity.

IV. HOW THE COURT MADE
For starters, of course, 'national procedural autonomy' is an unfortunate misnomer. A better term would be 'national procedural authority'. The latter more accurately describes the existence of such prima facie authority without the confusing overtones of a special kind of untouchability that 'autonomy' implies. We shall therefore use 'national procedural authority' instead of 'national procedural autonomy' here, realising full well that seeking to change the moniker of any grand general principle of EU law is an uphill battle. In any event, speaking about national procedural 'authority' more properly situates the inquiry in the domain of competences and the principle of conferral. 70 But the trouble does not end there. Let us return to the twin ideas of permissive versus prescriptive national procedural authority (or autonomy, if you insist) to see how, in failing to distinguish between the two, the CJEU further contributed to confusion. We can easily distinguish between the two in theory, as we have done here: permissive Member State procedural authority exists whenever the Union legislature in its legislative wisdom willingly permits Member States to apply their existing procedures and remedies; prescriptive Member State procedural authority, by contrast, exists where the EU Treaties, as matter of EU constitutional law, prescribe limits on Union interference with those Member State rules.
The difficulty is that in practice, the permissive and prescriptive often run together. This is not just true of this court and this particular area of adjudication. It's a pervasive feature of all constitutional adjudication around the world in all domains. Instead of determining that a legislature (or executive) overstepped its constitutional limits, judges in all areas of constitutional adjudication will often rather decide that the legislature (or executive) never intended the potentially unconstitutional act in the first place. To be sure, at times the legislative (or executive) purpose may be reasonably clear, such as when a directive mandates shifting the burden of (F'note continued) that 'in the presence of Community law, national procedures do not govern'. In each case, a separate enquiry is necessary to determine the consequences of negating the hypothesis. 70 Cf Kakouris, note 9 above (making a similar argument, albeit on Kelsenian grounds).
proof. 71 But when it is not clear what the political actor chose, a judge can hold in favour of the party complaining of a purportedly unconstitutional or ultra vires act without declaring the act itself to be unconstitutional, simply by giving a narrow interpretation of the act. The judge thereby prudently interprets the act in question by construing its purpose in such a way as to avoid the more difficult constitutional standoff. This leaves the potential constitutional conflict between a more broadly construed act and any hard constitutional constraints for another day. By postponing that conflict to another day, a judge need not decide in the case at hand, and may even studiously avoid deciding in the future, whether a more broadly construed act would actually violate the constitution. This is simple stuff. It is ubiquitous in constitutional adjudication the world over. Americans call it the Ashwander principle after the eponymous US Supreme Court case that most prominently invoked the idea, 72 or the more descriptive 'constitutional avoidance'. 73 Canadians, the British, and much of the Commonwealth call it 'reading down' a statute. 74 But the CJEU does not yet seem to have a name for it-let alone acknowledge the practice.
'Prudential avoidance', as we shall call it here, is generally a welcome judicial practice. The term 'prudential avoidance', which we shall use here, is especially suitable for the setting of the EU Treaties, which still not everyone likes to call a constitution. Above all, 'prudential avoidance' emphasises that the avoidance of the larger Treaty (ie constitutional) question is not arbitrary, mistaken, or lazy, but a deliberate and often wise exercise of judicial caution. Consistent with judicial minimalism and conducive to productive dialogue between the courts and the political branches, prudential avoidance keeps courts from drawing unnecessary lines in the sand, especially when those lines are, themselves, difficult to ascertain and defend as matter of constitutional interpretation. Prudential avoidance also allows the political branches to reconsider whether they really intend the constitutionally questionable act or whether they wish to recraft their act in response to a narrowing judicial construction, all without the embarrassment of having their work product struck down. In response to a narrowing construction motivated by prudential avoidance, the political branches can either let the narrow construction stand, or take new action that is more explicit, better tailored or broader, and coupled with a more full-throated defence and stronger supporting record. And if the political branches do insist on re-enacting the broader, more constitutionally troubling version of the act, prudential 71 See eg Art 8 Racial Equality Directive. Our discussion focuses on legislative acts here, even though there may be other acts, such as administrative or executive-type decisions that are subject to the same calculus we discuss in the text. 72 Ashwander avoidance on the earlier iteration allows the court, in turn, on the second time around, to reconsider its constitutional assumptions and preliminary considerations without the embarrassment of having to overturn its own precedent in doing so. Overall, then, opting for prudential avoidance is often a wise judicial choice. But prudential avoidance can be obfuscatory and unhelpful when courts do not indicate whether they are doing it or whether, instead, they are actually, positively, and definitively adjudicating the constitutional limits of the contested action. To be sure, every decision of prudential avoidance implicitly has taken a long look at the constitutional question it seeks to avoid. After all, the court must determine whether the constitutional question it is avoiding is a difficult one or whether the question is readily resolved. But a court usually ought to signal whether it means to decide that constitutional question definitively, or only to suggest that in light of a difficult constitutional question, which the court leaves undecided, the legislation (or other act) should be read more narrowly to avoid deciding that harder question. To be sure, signalling that a court is engaging in prudential avoidance does not always require identifying the precise constitutional question it seeks to avoid. That would sometimes defeat the purpose of prudential avoidance. But when engaging in prudential avoidance, it is generally useful for a court to make clear that it is ultimately disposing of the case at hand on the basis of a positive interpretation of the secondary legislation (or executive act) involved. This, then, seems to be the main difficulty with the CJEU's current jurisprudence on national procedural autonomy: the Court is often entirely unclear whether it means to rule in favour of Member State procedural autonomy as a matter of the (prudential) construction of the EU legislature's purpose, or as a matter of some hard prescriptive constitutional judgment. For instance, Michael Dougan has described the caselaw as giving priority to Member State procedural law (properly, in his view) where the general harmonisation in a particular area of the law is slight, and more easily displacing Member State procedural law where harmonisation is strong. 75 But it is not clear-either in Dougan's account or the decisions of the Court he is analysing-whether this is due to an interpretation of legislative purpose (ie that the Union legislature intends to harmonise Member State procedures if and only if the substantive law is strongly harmonised as well), or whether this is due to hard constitutional requirements, such as the principle of subsidiarity or proportionality (ie that a justifiable need to harmonise procedure can only exist where substantive Union law is harmonised as well). Dougan is not at fault in failing to specify these options. The Court itself does not consistently link its deference to Member State procedures, whenever it gives such deference, to the Court's construction of the Union's legislative purpose. Nor does the Court link such deference to any (even implicit) constitutional judgments regarding subsidiarity, proportionality, or respect for Member State constitutional identity.
Returning to our opening distinction, what Prechal, Dougan, and Lenaerts describe as a balancing test (ie weighing the objective reasons to defer to existing Member State procedures against the importance of vindicating Union law), should properly 75 See eg Dougan, note 27 above, pp 201-02, 310 ff. be seen as two balancing tests, or, put more precisely, as a balancing test performed in two different registers: the first permissive and the second, prescriptive. The first determines which procedures and remedies the EU legislature intended to permit Member States to control, and the second investigates which procedures the EU legislature might be constitutionally required to leave under Member State control.
As part of the first investigation into legislative purpose, the CJEU should generally presume that Union law does not intend to displace Member State procedural and remedial rules. Especially in the case of secondary legislation, the EU legislature must be presumed to operate with a good understanding that Member State courts remain creatures of the Member States, that the Union's competence over national procedural and remedial rules outside Articles 83 and 84 TFEU is based on a functional extension of the underlying substantive competence, and that the Union in this separate regulation of national procedures and remedies must comply separately with the principles of subsidiary, proportionality, and a proper respect for Member State constitutional identity. The Court therefore can and should, on balance, defer to Member State procedures unless there is good evidence (either by express stipulation or as might emerge from a particular context) that the Union's legislature indeed intended to displace those Member State procedures.
Only where the Union legislature may be fairly said to intend that Union procedures and remedies operate in Member State courts, must the Court squarely confront the second, prescriptive constitutional question. This, too, involves a balance, and it looks much like the first, but this time it occurs in the hard register of constitutional adjudication as opposed to legislative interpretation coupled with prudential avoidance. 76 Now the Court must decide definitively whether, due to the principles of subsidiary, proportionality, and a proper respect for Member State constitutional identity, the Union is constitutionally prohibited by the Treaties from imposing certain procedures on Member State courts.
To be sure, reading down legislation as a matter of prudential avoidance is itself a significant judicial intervention, 77 with all the legal and political consequences of forcing the legislature to pass new legislation if the legislature wishes to insist on something the CJEU thought not explicit enough in the original legislation. Hence the insistence here on identifying the normative foundation even for the mere interpretive 'presumption' that creates a speedbump (but not a roadblock) for the political branches. But as long as the Court itself is open about what it is doing, deferring to the Member States based on the narrow prudential interpretation of Union law may 76 Hard is relative. Constitutional adjudication, too, can be set aside later without a court confessing error. A court, for example, could strike down a measure as not justified only to uphold a redesigned measure later in the context of quite similar legislation on the same matter that has been more carefully enhance political engagement, increase transparency, and contribute to the Court's legitimacy. Especially on the latter, different degrees of deference across time or across different areas of the law to Member State procedures as a result of prudential interpretation of the varying legislative purposes seems far more plausible than hard constitutional prescriptions that seem to vary inexplicably from case to case.

V. SOME APPLICATIONS
With a constitutional approach based on competences, a distinction between permissive and prescriptive national procedural authority, plus a recognition of the role prudential avoidance may play, we can sort some decisions more sensibly. For instance, respect for, say, the passive role of a judge in civil proceedings 78 or for the preclusion of loss as a result of unilateral appeal (reformatio in pejus) in the Netherlands, 79 are likely prescribed by the Treaties, at least in certain settings, 80 and therefore may represent constitutional prescriptive judgments regarding the Treaty-based limits of EU law. Such respect for Member State procedures should therefore not vary much across time or different areas of the law. By contrast, other decisions on deference to Member State procedures, such as deference to Member State rules limiting the retroactivity of damages, 81 are better seen as the result of the Court's prudential interpretation that the Union's legislature intended to permit national rules to govern national proceedings. In the latter case, the question whether the Member State procedural rules can be displaced consistent with the primary law of the Union can easily be (re)considered, if the Union legislature were to insist more clearly on pre-empting such Member State procedures, or if a different context might make displacing the national rules a more pressing concern. 82 This constitutional approach also allows us to shed systematic light on two perennial concerns about the Court's decisions in this area. First, with regard to primary law, it helps us clear up a hitherto persistent puzzle in the supposed conflict between national procedural authority and supremacy. Second, with regard to secondary law, it may help explain the Court's varying deference to national procedures and remedies in different areas of substantive law.  82 To be sure, a prescriptive constitutional judgment can be revisited as well, whether to overturn the earlier judgment or to reassess the constitutional balance of subsidiarity and proportionality in light of new circumstances. But, again, the permissive calculus is performed in the softer register of construing legislative purpose, whereas the prescriptive calculus is performed in the more fraught register of constitutional construction. I shall refrain from elaborating on this any further here.

A. Primary Law: Resolving the Simmenthal/Rewe Contradiction
Unpacking the judicial calculus, as we have done here, helps clear up the 'puzzle' or 'contradiction' that Prechal and others identified between the Simmenthal/ Factortame line of judgments, on the one hand, and the Rewe/Comet line of judgments, on the other. 83 The charge is generally one of inconsistency and confusion in deciding Simmenthal and Factortame on primacy grounds, as opposed to the equivalence and effectiveness grounds of Rewe and Comet.
A straightforward response might be to charge the Court with conflating procedures and remedies. Recall Koen Lenaerts's opening bid allocating not just 'procedural rules' but also 'remedies' to the courts of the Member States. He did so in the course of discussing the balancing act in accommodating national procedural autonomy. Some might say President Lenaerts and the Court are confusing by not distinguishing carefully enough between procedures and remedies. If only we were to distinguish carefully between the two-and extend national procedural autonomy only strictly to procedural rules but not to remedies-matters would be clear. National procedural rules would thereby be saved from EU intrusion, while national remedial rules would be stamped out by the hard primacy of substantive EU law. After all, so this argument goes, it is called the principle of national 'procedural' autonomy, not the principle of national 'remedial' autonomy.
Indeed, as any decent remedies textbook will tell you, substance is your right, remedies are what you get for that right, and procedures are what take you from one to the other. 84 Remedies (such as whether you may bring a cause of action for damages, whether you get damages or equitable relief, or whether you can appeal an adverse judgment to a higher court) are therefore more closely tied to substance than those petty pesky procedures (such as time limitations, filing requirements, evidence rules, and so on). Given that remedies are more closely related to substance, they tend to tap into the primacy of substantive law. After all, for every right there is a remedy, as Chief Justice Marshall was not the first to say, with nary a mention of procedure. 85 Ubi ius, ibi remedium. 86 Procedures, on this view, are mostly the technicalities that do not really relate to substance, and hence do not tap into the primacy of substantive law.
Unless, of course, they do. As that same textbook should tell you, at least if it really is decent, the distinction between procedures and remedies, and indeed the distinction between procedure and substance, more broadly, is not so crisp after all. Even if you 83 See Prechal, note 6 above, p 687. do not subscribe to Justice Holmes's 'bad man' theory of the law, 87 or to a process theory of rights, 88 if there is one thing sociological jurisprudence and legal realism have taught us, it is to be sceptical of leaning too heavily on such formal distinctions. 89 In light of these critical movements, it should appear quite obvious that any substantive right is controlled, limited, and, in practice, significantly defined by, the procedures that govern access to the remedies for any breach of that right.
As we suggested at the outset, 90 the Court's judgments therefore properly recognise the functional interconnectedness among substance, procedure, and remedy. The point of considering this as a kind of trinity of legal normativity, as we did earlier, was to understand the three elements of substance, procedure, and remedy as different aspects of a legal norm that cannot be sharply distinguished from one another. Plus, each of the three is at least partially constitutive of the other two. A quick second look for purposes of the more specific question we are considering here confirms that no simple distinction between (pre-empted) national remedies and (preserved) national procedures can explain the Court's decisions. It would seem a rather pointless word game, for example, to classify Simmenthal as a 'remedial' case to explain its jumping straight to primacy without first considering national procedural autonomy in the balance. The judgment concerned Member State rules on the sequencing or timing of (1) sending a conflict between EU law and Italian legislation to the Italian Constitutional Court and (2) referring a question regarding that conflict without delay to the CJEU (then-ECJ). To be sure, we can classify the case as involving a question of 'remedy', as it involved the ability of a national lower court to set aside conflicting national law and the national remedy of an appeal or reference to the Italian Constitutional Court. At the same time, however, the judgment may just as fairly be classified as involving 'procedure', as the question did not concern the national remedy as such, but the national procedural rule that demanded that an appeal or reference to the Italian Constitutional Court must precede any potential reference to the CJEU or setting aside of national law. Thus, if there is a principle of national procedural authority (even short of the 'autonomy' the traditional label seems to promise), it should apply to Member State 'remedies' in addition to more narrowly conceived 'procedures'.
A better way to resolve the supposed contradiction is to begin by acknowledging, as suggested here, that primacy is always operating as one factor in the decision whether to leave in place Member State procedures and remedies. In that sense, the two lines of cases, ie Simmenthal/Factortame representing primacy, on the one hand, and Rewe/Comet representing Member State procedural authority (or 'autonomy', in the traditional formulation), on the other, are always present. 91 This is true even if only one line of cases or the other is mentioned in any particular judgment. Once we recognise that both principles-primacy and national procedural authority (or 'autonomy', if you insist) -are always present, we can chalk up the express mention or more elaborate discussion of one or the other to judicial rhetoric, strategy, or economy.
Illustrating the point, we can see both principles at work side-by-side in the Court's Unibet decision, 92 which answered three distinct questions spanning remedies and procedures. The first question concerned the availability of a free-standing declaratory action to challenge a national law's compatibility with Union law; the second concerned the availability of an injunction as interim relief in a pending action for damages; and the third concerned the procedural rules that govern the issuance of that interim relief. The Court expressly cites both the Rewe/Comet line of cases on national procedural autonomy as well as the Simmenthal/Factortame line of cases on supremacy in one breath. 93 And even though it does so expressly in answer to the first question, the same principles inform the entire judgment throughout. Moreover, the way the two lines of cases play out in answering the three questions differs not based on whether a given question formally concerns either 'remedies' or 'procedures'. Instead, both lines of cases, which the Court cites in answer to the first question, undergird the Court's answers to all three questions. On the first, the Court essentially holds that as long as equivalence and effectiveness are maintained, Union law does not require a free-standing declaratory action to determine whether a given national law is incompatible with Union law. On the second, the Court essentially harkens back to Factortame to hold that a national tribunal seized of a matter must be able to provide interim relief where necessary to ensure the full effectiveness of Union law. And on the third, the Court holds that the procedures governing the issuance of such interim relief are to be determined by national law, as long as equivalence and effectiveness are preserved. All three questions, then, are decided on equivalence and effectiveness with supremacy kicking in whenever the national court would otherwise fail on one or the other.
Second, when seen in this light, we might notice that the principal cases failing to mention the Rewe/Comet line altogether tend to concern Member States implementing primary, as opposed to secondary, Union law. For instance, even though the dispute in Simmenthal itself was nominally about the enforcement of secondary law, ie Regulation (EEC) No 805/68 on the common organisation of the market in beef and veal, the real dispute was about effectuating a core constitutional provision of the 91  Union's primary legal architecture, ie the reference procedure of Article 177 EC (now 267 TFEU) itself. Factortame, in turn, involved primary law on both counts: first, the underlying substantive question of whether the UK's fishing vessel rules violated the Treaty's non-discrimination provisions and, second, the integrity of the Treaty's basic reference procedure. In both Simmenthal and Factortame, the Court jumped straight to primacy as the framework for analysis. We see the same in cases regarding general principles and fundamental rights. 94 The fact that the dispute in such cases concerns primary-not secondary-Union law may help explain that quick jump to the primacy of EU law. When the demand on Member State procedures and remedies is made directly by primary Union law, considering prudential avoidance (ie asking whether the EU demand could be 'read down' just to avoid the harder constitutional question) makes little sense. In the case of interpreting primary law, any 'prudential' narrowing of the Union's demand on the Member States could generally be overcome only by amending the Treaties themselves, ie by constitutional change. Whenever primary law demands on Member State procedures are involved, then, there is no such thing as 'prudential avoidance', at least not in the sense we have used the term here. 95 All there is, in such cases, is constitutional adjudication. Any interpretation deferring to national procedures and remedies would have to result from hard constitutional interpretation, as opposed to more temporary and fluid prudential construction of Union legislation. 96 Third, especially in cases that involve restrictions on such central elements of primary law as the reference procedure, any limitation on Union law as a matter of constitutional adjudication would spell a limitation on the effectiveness of Union law everywhere. The Court has been consistent (and rightly so) in its fierce protection of the reference procedure, given that it lies at the very core of the EU's constitutional architecture. 97 As far as the reference procedure is concerned, only full effectiveness will do or else all of Union law is compromised. No surprise, then, that in primary law cases, especially those involving the effectiveness of the reference procedure itself, the Court is rather quick to jump emphatically to EU law primacy. 94 Eg Marguerite Johnston v Chief Constable of the Royal Ulster Constabulary, 222/84, EU: C:1986:206. 95 We set aside for purposes of this article that the pluralist constitutional practice of 'mutual accommodation' bears some resemblance to the prudential reading down that we discuss here, in the sense that constitutional pluralism often involves mutual accommodation in the interpretation of the various constitutional claims as a way to avoid a grand conflict, while allowing that a more explicit and deliberate constitutional demand might not be read in a similarly accommodating manner. 96  Finally, as for the absence of open balancing in these cases, the Court has never admitted that the principles of subsidiarity, proportionality, or any other form of balancing have any place in interpreting primary Union law. 98 No surprise, then, that balancing is not much discussed in primary law cases when it comes to procedures and remedies. The only remaining principle for the Court to consider in such cases from the perspective of EU law would seem to be Article 4(2) TEU's respect for Member State constitutional identity, which is the Treaty's way of formally incorporating constitutional pluralism into EU law itself. And here we have arrived at the highest level of judicial diplomacy and self-preservation, which often makes for bad, or rather opaque, judicial reasoning in actual judgments. 99 Investigating the dance of mutual accommodation that occurs at this level is beyond the scope of this Article. Suffice it to say here that if the CJEU ultimately gives in to Member State constitutional demands as part of such pluralist accommodation of national constitutional identity, it will usually do so not by acknowledging any real balancing or any real diminution of the effectiveness of EU law on account of national constitutional constraints. Instead, the Court will rather upload and integrate the Member States' constraints into EU law, and thereby adopt them as the EU's own. We have seen this play out time and again from Solange to Taricco and beyond. 100 In short, the centrality of core primary law to the case at issue, and the corresponding absence of prudential avoidance, may often explain the neglect of the Rewe/ Comet line of cases in favour of more ready displacement of Member State procedures on the basis of the primacy of EU law. 101

B. Secondary Law: The Shifting Role of Prudential Avoidance
In contrast to primary Union law, secondary legislation may indeed be interpreted narrowly to preserve restrictive Member State procedures, whether simply as a matter of ordinary legislative interpretation or as a matter of prudential avoidance to head off 98 But cf Halberstam, note 1 above (arguing that judgments, such as Germany v Parliament & Council, C-376/98 on tobacco advertising, may be understood as implicitly applying principles of subsidiarity in the judicial construction of the limits of conferred powers). 99  a larger constitutional question or clash. After all, if the legislature created the right to be enforced in court, the legislature might have created a more limited right, ie one restricted by procedural and remedial limitations in national courts, just to avoid any uncomfortable or difficult constitutional questions a broader right might bring. In the course of determining the scope of such permissive national authority, ie the national discretion EU legislation chooses to allow, balancing would seem quite natural and, therefore, front and centre. Hence, in cases concerning the effectiveness of simple secondary legislation, the Court will often prominently consider whether it might not be adequate or sufficient to let Member State procedures stand and whether Member State procedures serve legitimate functions the Union ought to respect (or incorporate into EU law itself). 102 This essentially is the run-of-the-mill reasonableness check and the balancing the Court routinely performs in the absence of EU law specifying any particular procedural rules. 103 And this is also the core of prudential avoidance and restraint. As discussed previously, a deferential form of balancing, where the tie goes to the Member States, avoids more difficult constitutional questions, and may initiate further dialogue with the EU legislature to determine whether certain procedures and remedies are really needed (or precluded).
But prudential avoidance does not always make sense. First, and rather obviously, prudential avoidance makes little sense where the procedural demand of secondary EU legislation is clear. In part this simply explains the basic Rewe test, ie where EU law rules on procedures and remedies are indeed present, the remainder of the Rewe formulation falls away. The Rewe presumption, after all, was conditioned on the 'absence of Community rules on the subject'. 104 But even if you further agree, as argued here, that some solicitude for Member State procedures is nevertheless constitutionally prescribed even where Community law rules do exist, prudential avoidance in such cases still makes little sense. Where secondary legislation expressly specifies, say, the statute of limitations, or the shifting of the burden of proof, any limit on EU law's procedural and remedial demands can be determined only by considering the constitutional constraints of the foundational Treaties. To be sure, clarity is ultimately in the eye of the beholder, but, in principle, you do not read down unambiguous legislation. The same goes for situations where EU law does not expressly spell out its procedural and remedial demands in the text of legislation, but such demands are nonetheless quite clear from the context of the legislation. In such situations, too, only prescriptive constitutional considerations should keep EU law from displacing national rules.
Second, prudential avoidance may make less sense in areas of dense, substantive harmonisation. As noted earlier, Michael Dougan and others have observed, and even advocated for, a sectoral approach, in which the Court gives less deference to Member State procedural rules in sectors that are, in substance, intensely harmonised. 105 Distinguishing carefully, as we have done here, between the interpretation of legislative purpose, hard constitutional review, and prudential avoidance helps explain and support that argument. Even without an express intervention into Member State procedures in the text of EU legislation, there may be good reason, on the understanding put forth here, for the Court to consider the density of Union regulation in the course of adjudicating whether EU procedures should displace Member State procedures. Even if the Union's secondary legislation does not make a particular procedural demand explicit, the dense regulation or harmonisation of a particular area of the law may nonetheless suggests an underlying legislative purpose to have Union law displace 'domestic legal provisions, however framed' 106 (ie including Member State procedural law).
Put in simple (almost simplistic) terms: the more the Union legislature regulates in a given area, the more the Union legislature is willing to displace Member State rules, and the less qualms the judiciary should accordingly have to read EU legislation as intending to displace not only national substantive law but national procedural and remedial rules as well. Intense harmonisation in a given area thus generally argues against prudential avoidance, and, instead, in favour of interpreting legislation in that area more broadly, while preserving restrictive national rules only when required by hard constitutional constraints on EU power. Furthermore, if the intense harmonisation is indeed justified for purposes of substantive law, the density of substantive regulation would seem to provide a good foundation on which to argue that the additional harmonisation of national procedures should satisfy subsidiarity and proportionality as well. 107 Third, and finally, where secondary legislation, such as, say, the equality directives, gives effect to a general principle of Union law, prudential avoidance should also have little purchase. Such legislation straddles the divide between primary and secondary law. And as noted earlier, prudential avoidance makes less sense in the case of primary law, given that such primary law cannot simply be amended in response to narrow judicial construction. Moreover, despite the EU legislature having a legitimate voice in the interpretive construction of primary law, one thing is clear: the legislature is neither the source nor the sole keeper of primary law. Accordingly, secondary legislation implementing general principles by providing for individual justiciable rights to vindicate such primary law, should generally not be subject to the narrowing construction of prudential avoidance. The presumptively broad construction of the substance should carry with it the broad construction of procedural demands to vindicate that substance broadly as well. 108 Returning to Lisbon just briefly, this last conclusion is underscored by the codification in Article 47(1) of the Charter of Fundamental Rights of 'the right to an effective remedy' for '[e]veryone whose rights and freedoms guaranteed by the law of the Union are violated'. Plus, as Article 19(1) now expressly confirms, the obligation to provide an effective remedy for such rights is shared among the various courts of the Union. The patchwork judicial system of the Union we described earlier, ie the complex judicial system comprising the courts the Union itself establishes plus the courts of the Member States, must work seamlessly to provide a coherent and complete system of judicial remedies for all the rights and obligations under Union law. 109 This becomes of special importance in cases of secondary legislation that gives effect to general principles of primary law, such as equal treatment, as the rights contained in that secondary legislation are not exclusively of the legislature's own making.
It is inappropriate, therefore, in cases of secondary legislation that expressly gives effect to general principles in substance, to read the accompanying legislative purpose regarding procedures and remedies narrowly. Prudential avoidance should have little purchase here. Put another way, in such cases it must be presumed that the EU legislature in enacting substantive obligations, intends those obligations to include broad access to effective relief in court. Thus, for example, even where the Working Time Directive, which gives unconditional expression to the substantive right to paid annual leave contained in Article 31(2) CFR, failed to specify that Member States provide effective access to the judiciary, this no longer matters: It is true that Directive 2003/88 contains no provisions on judicial remedies available to the worker, in the case of a dispute with his employer, to enforce his right to paid annual leave under that directive. However, it is not disputed that the Member States must, in such a context, ensure compliance with the right to an effective remedy, as enshrined in Article 47 of the Charter. 110 Any narrowing construction of such secondary legislation to avoid a difficult constitutional question is thus unwarranted.

VI. CONCLUSION
Calling the principle of 'National Procedural Autonomy' a complex oxymoron would be unfair. 111 But it is indeed a complexly unfortunate term. As an initial (F'note continued) Gender Equality Law in the Light of European Requirements' (1993) 13(3) Oxford Journal of Legal Studies 320, p 347. The approach discussed in the text still allows the Union legislature to fine tune any deference to national procedural law, as long as it chooses to do so explicitly (and as long as such deference to national procedures does not itself entail a fundamental rights violation). It thus still tethers the CJEU to the EU legislature in the interpretation of such secondary law. 109  matter, the adjective 'procedural' is problematic, as the arena of Member State authority to which the principle extends, includes not only national procedures, strictly understood, but also national remedies (including what some might further divide into remedies and relief). It thus governs not only matters most narrowly considered 'procedure' (eg time limitations and rules of evidence), 112 but also allowable forms of action (eg free-standing constitutional complaints), 113 as well as judicial relief (eg money damages, publicity, reinstatement, and injunctions). 114 The noun in national procedural 'autonomy' is also deeply misleading-and the reason why this article prefers calling it the principle of national procedural 'authority'. National procedures and remedies are not immune from EU law. They are not even uniquely shielded from EU law intrusion. What we have instead, is an acknowledgment that EU law mostly piggybacks on the institutional organisation of national courts without any general authority over their organisation, and thus mostly with only implied power to regulate national procedures and remedies insofar as necessary for the effectuation of the Union's substantive law. As such, EU law preserves and respects national judiciaries' organisational rules, interfering only as needed to protect the substance of EU law, and doing so only as consistent with subsidiarity, proportionality, and a proper respect for national constitutional identity. This helps explain and, to a certain extent, justify the Court's varying deference to Member State procedures and remedies, and it provides the proper normative contours for the inquiry the Court must perform. The Court must first consider the permissive aspect of national procedural authority, ie what national procedures and remedies EU legislation positively permits. Only after determining the extent of positive permission, and concluding that existing national procedures and remedies do not match the legislative purpose, must the Court squarely confront the more difficult, prescriptive question of any further respect for national procedures and remedies the EU treaties may mandate.
In a practice of prudential avoidance, however, the Court already has that more difficult constitutional question in mind-and seeks to avoid it-when determining the initial question regarding the positive purpose of EU law. Thus, construing the positive legislative purpose, the Court will read secondary EU law as not demanding new or altered national procedures and remedies, unless such EU law demand for certain procedures and remedies is clear from either the text or context of the substantive provision at issue (including a consideration of what the effectiveness of EU law may require). Where EU law overcomes this presumption, the Union demand for certain procedures and remedies must then actually be checked against the prescriptions of subsidiary, proportionality, and a proper respect for national constitutional identity.
In conducting the constitutional review under subsidiarity, proportionality, and a proper respect for national constitutional identity, moreover, these three constraints tend to be pursued more stringently than in the usual case of considering EU substantive measures for such compliance. In part, this heightened concern is due to the usual resort to implied powers in the case of regulating procedures and remedies, as opposed to the use of expressly enumerated competences over substantive issues. And in part, the heightened concern for constitutional constraints in the area of procedures and remedies is due to the fact that certain aspects of the judiciary's institutional organisation, such as the passive or active role of judges, can cut very close to the national constitutional identities the Union ought to respect. This, then, in a nutshell, is what the principle of national procedural authority is all about. The constitutional approach pursued here ultimately allows us to make better sense of the caselaw of the Court. It helps us understand which cases likely reflect only contingent deference to national procedures and remedies, and which cases should be read as a more permanent respect for national procedures and remedies across time and different areas of substantive law. It clears up the longstanding puzzle over why the Court in some cases decides the question of deference to national procedures and remedies by resort to a balancing test, and in other cases displaces those procedures more easily by jumping straight to the primacy of EU law. And, finally, it helps provide guidance for when secondary legislation should be read cautiously to preserve national procedures, and when it should be read more boldly as constrained by national procedures and remedies only insofar as hard constitutional limits demand.