Unjust Enrichment in National
and European Law, and the
‘Remedies’ Jurisprudence of the ECJ
AT THE BEGINNING of the last chapter, three factors were identified which connect the EU to claims of unjust enrichment involving public bodies. The third, with which that chapter was concerned, was that the public body involved in the claim may itself be an institution of the EU. The first two, however, concerned the interrelation of EU law and national law in questions of unjust enrichment involving public bodies. In some of the cases considered in parts one and two,1 the ‘ultra vires’2 nature of the public body’s action arose from the application of European rather than national law. In other words, the source of the nullity was the EC, not the Member State. When this is the case, the ‘remedies’3 which must be provided by the Member State legal systems are governed in part by EU law as well as by national law. It is these two matters which will be the subject of this chapter.
However, closely related to these two issues are some of the fundamental themes of the EU legal order and its Member States. As a result these issues have ‘revealed the larger problem of the articulation of Community and national law’.4 In that sense, for an EU lawyer as much as for an unjust enrichment lawyer, the interplay between Member State and EU law in unjust enrichment claims can provide an important insight into the ECJ’s so-called ‘remedies jurisprudence’. Comprehensive accounts of the essential principles of EU law can of course be found elsewhere,5 but it is worth briefly listing those that are particularly important in this context.
The first point to be made is that, of the ‘three pillar’ structure of the EU prior to the entry into force of the Lisbon Treaty in December 2009,6 it was the first of the three, the so-called ‘Community pillar’ which was relevant to the current investigation. As is well-known, following the failure of early attempts at political union, the origins of the current Union are economic,7 and only after the Treaty on European Union (TEU) was the term ‘Economic’ dropped from the title of the European Community which formed this first pillar (this is a matter of particular importance given that the law of unjust enrichment is specifically concerned with transfers of value).8 Indeed, in the political sphere, the transfer of power to the Union from the Member States, either substantively or structurally, has not always been smooth and on many occasions the resulting division of competence or responsibility between the Member States and the Union has required arbitration, and in this the role of the European Court of Justice (ECJ) has been vital in further developing EU law.9 For example, it was against the background of the so-called ‘Luxembourg sclerosis’ on the political front10 that the ECJ developed the concept of direct effect and supremacy of what was then EC law, according to which EU Treaty provisions and other forms of legislation (regulations, directives and decisions) can be directly relied upon in national courts by individuals,11 whether or not they have been implemented by national legislation, and whether or not there is conflicting national law, 12 regardless of whether that national law pre- or postdates the EU law in question. Craig and de Búrca refer to the ECJ’s ‘bold theory’13 in this context, and Arnull speaks of its ‘central role’ and ‘creative’ or ‘surprising’ use of its powers,14 and it is these principles which provide the first way in which EU law has an impact on cases involving unjust enrichment and public bodies in the Member States. The principles of direct effect and supremacy mean that EU law imposes an additional set of limits on the powers of any public body in a Member State, so that a tax, contract, payment or other transfer of value can be invalidated, not just by domestic law, but also by EU law.
Most of this process has taken place through the dialogue between national courts and the ECJ provided for by Article 267 TFEU (ex Art 234 EC). However, as well as providing for preliminary rulings on interpretation of EU law, which can then have an impact at national level on the relevant domestic rules, Article 267 also provides for rulings on the validity of acts of the institutions of the Union (the Council, Commission and European Parliament) and of the European Central Bank. Challenges to validity can also arise under Article 263 TFEU (ex 230 EC) EC, although the relationship between the two is tightly controlled.15 In particular, there is an interesting parallel between the presumption of exclusivity in favour of Article 263 over Article 267 on the one hand, and on the other, the earlier presumption of exclusivity in O’Reilly v Mackman16 in favour of the public law procedure rather than the private law procedure in English law (where challenges were made to the validity of administrative actions). This is particularly so given the comparable protections available under Article 263 and Ord 53; short time limits and narrower standing requirements. Nevertheless, it should be noted, as is apparent from the case of ex p Accrington Beef,17 that the two ‘presumptions’ operate separately. Thus ex p Accrington Beef was brought through the national public law procedure, but allowed to proceed at European level outside the procedural protections of Article 263. It would, conversely, be possible for a national issue to contain a private law cause of action (as in a European version of the British Steel18 case), which would take it outside the public law procedure. However, in dealing with the incidental question of validity of a European administrative action, the applicant could be denied the Article 267 challenge to validity on the basis that he/she/it ought to have brought an Article 265 action.
The Principles of National Procedural Autonomy, Equivalence and Effectiveness, and the Three Phases of the ECJ’s Case Law
Once the concepts of direct effect and supremacy had been established, a related question arose. There was no point in EC law, as it then was, being supreme unless it was also to be uniform and effective across the Member States.19
But how should the rights thus established be protected? How should jurisdiction be divided between the national courts and the ECJ? Three distinct phases can be identified in the Court’s approach to answering these questions,20 and it is notable that such issues have often arisen in cases where charges have been imposed by Member States in breach of the Treaty, so that it is here that European law has had its second big impact on unjust enrichment cases involving public bodies in the Member States.
The First Phase
The first phase of the ECJ’s jurisprudence is typified by Rewe-Zentralfinanz GmbH v Landwirtschaftskammer für das Saarland21 and Comet BV v Produktschap voor Siergewassen,22 both of which contained such issues. In Rewe-Zentralfinanz, a claim had been brought before the German courts for the repayment (with interest) of charges which had, in an earlier case before the ECJ, been held to have an effect equivalent to customs duties within the meaning of Article 13(2) EEC.23 Comet concerned levies imposed in breach of Community law, which the applicant could no longer claim back under national law, the national time limit for doing so having expired. The applicant thus wanted instead to set these amounts off against other amounts it owed to the defendant. The ECJ in Rewe-Zentralfinanz held that
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.24
It ruled in almost identical terms in Comet.25 This principle of national procedural autonomy was applied in several cases to permit the continued operation of various national laws. For example, in Hans Just I/S v Danish Ministry for Fiscal Affairs,26 it permitted the application of a Danish rule relating to the passing on of the expense as a defence to a claim for the return of money unlawfully levied.27 In Société Roquette Frères v Commission,28 and Express Dairy Foods v Intervention Board for Agricultural Produce,29 the question of interest was, on this basis, left to be determined by the national rules. In Rewe-Handelsgesellschaft Nord GmbH v Hauptzollamt Kiel, the ECJ affirmed that the Treaty ‘was not intended to create new remedies in the national courts to ensure the observance of Community law other than those already laid down by national law’.30 However, as Arnull notes, the significance of this remark should not be exaggerated since the court was not being asked to create a new remedy in that case.31 Crucially, the ECJ in Rewe-Zentralfinanz had imposed two important limitations on the principle of national procedural autonomy:
• The principle of equivalence: the procedural conditions laid down by national law ‘cannot be less favourable than those relating to similar actions of a domestic nature.32
• The principle of effectiveness: procedural conditions laid down by national law must not make it ‘impossible in practice to exercise the rights which the national courts are obliged to protect.33
It was through these limitations that the law later began to develop.
The second phase of the ECJ’s case law in this context was also heralded by another issue concerning the levy of illegal charges, this time health inspection charges, on the importation (by San Giorgio)34 into Italy of dairy products from other Member States. Again, the question related to passing on, and in particular whether a particular charge could be presumed to have been passed on, in the absence of written evidence to the contrary, when the goods to which it related were transferred. The answer was no, the ECJ emphasising that ‘any requirement of proof which has the effect of making it virtually impossible or excessively difficult to secure the repayment of charges levied contrary to Community law would be incompatible with Community law’.35 The shift in the principle of effectiveness from ‘not impossible in practice’ to ‘not virtually impossible or excessively difficult’ (emphasis added) shows the increased judicial interventionism exercised by the ECJ in this second phase.
The approach taken by the Court in this phase also led to the development, in Francovich and Bonifaci v Italy36 of the principle of state liability in damages for breach of EC law, the breach in question in that case being the failure to implement a directive. The relevant conditions, held the Court, were that the directive in question must have envisaged the grant of rights to individuals; that the provisions of the directive made it possible to identify the contents of those rights; and that there be a causal link between the state’s failure and the loss suffered by the claimant.37
The Third Phase
However, subsequently the ECJ returned to a more conservative approach, promoting once more the principle of national procedural autonomy. Again, this phase contains a case on allegedly invalid taxes, Fantask A/S v Industriministeriet,38 in which several companies sought to rely on an indirect taxes directive in order to reclaim certain taxes levied by the Danish authorities. One of the questions sent by the national court to the ECJ under Article 267 TFEU (ex 234 EC) was whether the time limit for reclaiming could begin to run before the Directive had been properly implemented. In reply, AG Jacobs distinguished the earlier complex case law on the matter,39 holding that it could. This increased judicial restraint in the appreciation of national procedural rules was accompanied by a narrow approach to the further definition of the conditions in which state liability in damages arises. Thus Francovich was succeeded by the well-known joined cases of Brasserie du Pêcheur SA v Germany and R v Secretary of State for Transport, ex p Factortame Ltd (Factortame III),40 which established that liability would only exist for a ‘sufficiently serious breach’ which would arise when the Member State had ‘manifestly and gravely disregarded’ the limits on its discretion.41
However, it should be noted that phase three, which currently informs the decisions made by the ECJ, is not therefore a return to phase one, but rather a midpoint between phases one and two. In phase three, some of the developments towards increased effectiveness have been retained, but these requirements are now balanced more subtly against the justifications for the national procedural rules.42 As before, cases on sums paid but not due have been central to these developments. For example, it is possible to introduce time limits which reduce the period within which repayment of sums collected illegally may be sought, as long as this period complies with various requirements. In Marks and Spencer v Commissioners of Customs and Excise,43 M&S had overpaid VAT on gift vouchers since 1992 and when this was revealed to be the case by the decision in Argos Distributors v Commissioners of Customs and Excise on 24 October 1996,44 M&S submitted a claim six days later. However, the Finance Act 1997 retrospectively ‘capped’ the recovery open to M&S, restricting repayment of VAT to three years before the making of the claim (that is from 1993 onwards only). The ECJ held that:
Whilst national legislation reducing the period within which repayment of sums collected in breach of Community law may be sought is not incompatible with the principle of effectiveness, it is subject to the condition not only that the new limitation period is reasonable but also that the new legislation includes transitional arrangements allowing an adequate period after the enactment of the legislation for lodging the claims for repayment which persons were entitled to submit under the original legislation. Such transitional arrangements are necessary where the immediate application to those claims of a limitation period shorter than that which was previously in force would have the effect of retroactively depriving some individuals of their right to repayment, or of allowing them too short a period for asserting that right.45
On the other hand, retroactivity is not absolutely fatal to the provisions as long as there are sufficient transitional periods.46 In addition, the time limit must be reasonable and must not be specifically targeted at the effects of a particular decision of the ECJ.47
In coming to their conclusion in Marks and Spencer, the ECJ specifically recognised
the argument of the United Kingdom Government to the effect that the enactment of the legislation at issue in the main proceedings was motivated by the legitimate purpose of striking a due balance between the individual and the collective interest and of enabling the State to plan income and expenditure without the disruption caused by major unforeseen liabilities.48
It then held that these arguments could justify reasonable limitation periods as long as they also ensured that rights conferred on individuals by Community law were safeguarded.
As was noted in chapter five, the Court will also examine the date from which the time begins to run in order to ensure that this too complies with the requirement of effectiveness. Accordingly, in Preston v Wolverhampton Healthcare NHS Trust,49 the ECJ held that where the claimants were employed under temporary contracts, a time limit which ran for six months from the end of each contract rather than from the point at which that series of contracts (and thus the ongoing stable employment relationship) was interrupted, rendered the exercise of the Treaty right excessively difficult and was therefore precluded by EC law. Similarly, in Manfredi v Lloyd Adriatico Assicurazioni SpA,50 the Court held that time could not begin to run from the date on which an anticompetitive agreement was concluded because that would mean the limitation period could expire before the infringement of competition law was brought to an end. This in turn could result in a situation whereby an individual could potentially suffer harm after the limitation period for challenging the source of that harm (namely the anticompetitive agreement) had already expired. However, in Recheio-Cash and Carry SA v Fazenda Publica/Registo Nacional de Pessoas Colectivas,51 the limitation period was permitted to run ‘from the end of the period allowed for voluntary payment of the charges’, and in Edis v Ministero delle Finanze,52 the permissible period ran from the date of payment itself.
This balancing technique has also been applied to the requirement of equivalence which, having originally taken a back seat during the development of the principle of effectiveness, has now been developed more fully. In Edis,53 another case concerning the reclaiming of money paid but not due, the ECJ held that as long as the same rule was applied to recovery of charges in breach of EC law as was applied to charges in breach of national law, the applicable national rule did not have to be the most generous rule available in any circumstance in national law. It was therefore perfectly acceptable for the Member States in both Edis,54 and Ministero delle Finanze v Spac,55 to lay down shorter limitation periods for cases involving public bodies than would be applicable to private claims, as long as those limits applied to all public claims and not just those involving EC law. This is not the only area in which EU law has recognised a distinction between public and private law at European level. For example, the Court has developed case law on the definition of the state both for the purposes of distinguishing horizontal from vertical direct effect (such as Marshall v Southampton and South-West AHA,56 Costanzo (Fratelli) SpA v Commune di Milano57 and Foster (A) v British Gas plc58) and for the purpose of establishing state liability in damages.59
All these principles, then, provide the basic constitutional structure for the development of EU law and its relationship with the legal systems of its constituent Member States. Indeed, it is these principles with which rules such as those contained in the Finance Act 2007 and Article L190 of the French book of fiscal procedures must be compatible.
Unjust Enrichment and Ultra Vires in National and
EU Law; Two Different Models for Analysing
the Division of Labour between the
Member States and the EU
One of the problems with this process of dividing labour between the national and EU systems is that it is usually referred to as the ECJ’s ‘remedies jurisprudence’, and yet the principle of leaving most of the responsibility for those ‘remedies’ to the Member States is often referred to as ‘national procedural autonomy’. This lack of clarity in the terminology used is unhelpful,60 and various academics or practitioners have studied the area in an attempt to analyse the division of labour in more precise terms.61 One example of this is the work of van Gerven, who interestingly also chose to use a mapping process.62 His aim was to try to resolve the tension between the ECJ and the national courts over the division of labour between them in enforcing European law.
The key to van Gerven’s work, as his title indicates, is a distinction between ‘rights’, ‘remedies’ and ‘procedures’. Rights to him are legal positions that a legal person may have, and which can be enforced by that person against some or all others before a court. This can be done by means of one or more remedies, which are classes of action intended to make good infringements of the rights concerned, in accordance with procedures which govern the exercise of such classes of action and are intended to make the remedy concerned operational. The concept of remedy then further subdivides into constituent and executive elements.63 ‘Rights’ and ‘constituent elements’ of the remedies, he argues, are for decision by the ECJ. On the other hand, ‘executive elements’ of the remedy are for the Member States.64
van Gerven’s own use of the word ‘remedy’ is interesting. In examining public law, it was seen that the division between public and private law has focused on the availability of public law ‘remedies’,65 and in that context it was argued that at least some of the public law ‘remedies’ could be seen as ‘responses’ to ultra vires ‘events’ (for example, a quashing order for a biased decision). Here again, the word ‘remedy’ is at the heart of a legal division, but it is already clear that this time van Gerven does not just use the word ‘remedy’ as a synonym for ‘response’.
Birks has argued that there are five different meanings of the word ‘remedy’.66 It is submitted that it is his first definition that best captures van Gerven’s meaning: an ‘actionable story’ typified by the old English forms of action. It was Birks’ argument that for many years the use of such forms of action obscured the true nature of unjust enrichment actions, and prevented the development of a clear event/response taxonomy. It is interesting, then, to see that the far newer and therefore less developed system of EU law should be, as van Gerven’s work perhaps shows, at this same level of development, using the word remedy in sense one, as a general ‘actionable story’, rather than containing a more sophisticated division.
It is also apparent that van Gerven’s and Birks’ concepts of the word ‘right’ seem to align with each other,67 but it is not so easy to establish the particular substantive rights to which van Gerven is referring in his article. It is arguable that there are in fact two separate categories of right at issue here, and as a result, two possible models for the precise division of labour between the national courts and the ECJ.
The EU68 has gradually extended its jurisdiction over various matters, and has passed legislation of various sorts in addition to the treaties themselves, all with the purpose of changing the behaviour of the Member States and private entities and individuals in those states in various ways.69 Also, as noted above, the supremacy of EU law means that where national law conflicts with the relevant EU law, the latter will prevail, and must be upheld by the national courts. As the ECJ held in van Gend en Loos (NV Algemene Transporten Expeditie Onderneming) v Netherlands Inland Revenue Administration ‘the Community constitutes a new legal order of international law for the benefit of which the states have limited their sovereign rights, albeit within limited fields’.70 The existence of those European limits thus renders unlawful any actions of the Member States which go beyond them, for example by creating an impediment to the four freedoms.71 Conversely, in Francovich the ECJ held that
the EEC Treaty has created its own legal system which is integrated into the legal systems of the Member States and which their courts are bound to apply. The subjects of that legal system are not only the Member States but also their nationals. Just as it imposes burdens on individuals, Community law is also intended to give rise to rights which become part of their legal patrimony.72
Thus the finding of EU law unlawfulness, as seen above, gives the individual concerned the ‘right’ to an effective/adequate ‘remedy’ which is equivalent to the national ‘remedy’ applicable in the same circumstances. If ‘remedy’ in this context does indeed mean ‘an actionable story’, then the EU law right of adequacy/ effectiveness is the right to such an actionable story at national level, or as we would now understand it, the right to recognition that a particular event has occurred which in turn gives the claimant a right to certain responses.
It can thus be seen that there are two types of right being discussed here, the right to an actionable story of whatever kind, and the right within that actionable story which connects a particular legal event (to use the Birksian terminology) to a particular response. It is not clear to which of these two meanings van Gerven is referring when he says that definition of the ‘right’ requires a uniform content across the Member States, so that it should take place at the European level.
On the one hand, he could simply be referring to the EU ‘rights’ to the conceptualisation of an unlawfulness, leaving the whole of the remedy, in the sense of ‘actionable story’ to the jurisdiction of the Member States subject to the requirements of equivalence and effectiveness. Under this model, the fact that the ‘substantive conditions underlying the right’ and the ‘constituent element of the remedy’ are equal is simply a statement that the unlawfulness must be recognised by both the ECJ and the national court; a restatement of the supremacy of EU law. Under this model, the EU is only a general source of unlawfulness and as long as the Member State recognises that unlawfulness and gives some cause of action based on it, the Member State’s EU obligations will be satisfied.
On the other hand, it is possible that the jurisprudence of the ECJ has developed beyond this, and that van Gerven could be referring to the second kind of right, namely the right to the law’s conceptualisation of certain events entailing certain responses. On this basis, in terms of van Gerven’s division of jurisdiction, definition of the event would require a uniform content across the Member States, so it should take place at the European level. If this were the case, then the second element of van Gerven’s ‘remedy’, namely‘executive conditions’73—would be the ‘response’, and it would be this which should, in van Gerven’s words, be ‘assessed for consistency within the framework of adequacy’, through a proportionalitystyle balancing test as in van Schijndel and van Veen v Stichting Pensioenfonds voor Fysiotherapeuten74 and Peterbroeck v Belgian State,75 while remaining within the jurisdiction of national courts. If this were the case, then the event/response division would fit with the EU/Member State substantive division of jurisdiction, with the EU providing a uniform definition of the event to which the Member States would be free to respond in their own way, subject to the requirements of equivalence and effectiveness.76
The distinction between these two models is not unlike the distinction between the two options identified in chapters two and six as justifications for recovery in a case such as Woolwich Equitable Building Society v IRC,77 where it was pointed out that while English courts have regarded restitution following an ultra vires charge as being based on the private law principle of unjust enrichment, the courts of Canada have regarded it as a purely public law ‘remedy’ for the ultra vires nature of the charge itself, rather than a freestanding cause of action in itself.78 The French courts have followed each of the two approaches, depending on the particular context.79 The difference is, of course, that here the distinction is simply between the right to an actionable story in general (Model 1) and the right to a particular actionable story (Model 2, whether that specific story is in public or private law), whereas in chapters two and six the distinction was between an actionable story wholly in public law and an actionable story wholly in private law. Nevertheless, to the extent that the Canadian approach and the Model 1 approach described here do share a tendency to regard any restitutionary actions as part of a simple ‘mopping-up process’, it is not impossible that conclusions drawn here about the stability of the Model 1 approach could potentially have implications for the Canadian approach.
Obviously, in the EU law context, it is important to determine which of these two models is operating in a particular area, because the more an area is governed by the reasoning of Model 2, the greater the influence of EU law on that area, even at the domestic level, and the greater the chance that the rules affecting the availability of the cause of action and its defences will derive, not from national law, but from EU law. Also, as is evident from the reasoning of Lord Goff in Woolwich noted at the beginning of chapter seven,80 whatever the impact of EU law turns out to be, it is unlikely to be confined solely to cases raising an EU law element, but rather it has the potential to extend even into purely domestic situations, by analogy.
It is clear that in general there is evidence of both models operating at the European level. In the course of his investigation, van Gerven examines various different areas of EU law, and in the first of these, ‘setting aside national measures’, it is plain that only the first of the two alternative models is operating. EU law merely establishes an ‘unlawfulness’ that national law must supplement with its own ‘actionable story’ including both event and response. The claimant has a right to this actionable story, but the particular event contained in that story is to be defined by national, not EU law. Conversely, in cases involving ‘compensation’, the event is a wrong, and it seems reasonable to suggest that the elements decided by EU law81 all relate to this wrong event, while the ‘content and extent of the remedy’, re-analysed as the ‘response’, of reparation is left to the national jurisdiction. This set of cases is thus the opposite of ‘setting aside national measures’ and they cannot be interpreted simply as ‘illegalities’ which trigger the national events. Of the two models, only Model 2 (EU event) can thus be operating in relation to claims for compensation. In a sense, this is not surprising, since it was noted at the end of chapter seven that in actions brought against the institutions of the EU there is a similar conception of an event of wrongdoing, and indeed the ECJ now deals with both sets of wrongs as sharing a single line of jurisprudence.82
Cases of unjust enrichment,83 however, have generally tended to follow Model 1, and again, this may well be regarded as unsurprising given that in actions against the EU institutions, ‘unjust enrichment’ is a general principle only, not a freestanding cause of action like that relating to wrongs.84 The Advocates General, on the other hand, tend more often to follow Model 2 reasoning, and more recently there has even been some evidence of this route being taken by the ECJ. Cases involving unjust enrichment generally divide into four types:
2 Cases where a national intervention agency (NIA) levies money on behalf of the EU, but this turns out to be contrary to EU law;
3 Cases where a Member State pays out money to private parties in breach of EU law (a) as an invalid state aid or (b) via an NIA; and
4 Cases where the EU pays out money to an individual in breach of EU law.
Though there are also cases which do not fall perfectly into this pattern, for example because EU money is granted on the basis of certain conditions and the grant is rendered invalid when those conditions are not fulfilled.85 In practical terms (perhaps unsurprisingly)86 as will be seen in what follows, the ECJ has demonstrated itself to be strictest in Cases 1 and 3(a). Indeed, cases in category 1 are the most frequent,87 and have arisen since the days of the European Coal and Steel Community Treaty.88 However, in conceptual terms there is a potential difference between Cases 1 and 2 on the one hand, and Cases 3 and 4 on the other, as a result of the way in which such cases tend to be reasoned.
First, however, it is important to note that there are two key factors which operate against the introduction of Model 2 reasoning in this area. The first of these factors is that it is often the case that the right to recover, or the right even to an ‘actionable story’ at Member State level, comes not from jurisprudence at all, but from a directly effective piece of European legislation. Thus in the case of state aids, when these are found by the Commission to be invalid under the terms of Article 107 TFEU (ex Art 87 EC), Regulation 659/99 EC89 lays down the conditions for their recovery. If the Member State fails to terminate or recover the aid, the Commission can bring Article 108(2) TFEU (ex Art 88(2) EC) infraction proceedings against it, and the Member States’ only defence in such circumstances is that the Commission’s decision was unclear or unambiguous,90 or that recovery is absolutely impossible.91 It is therefore clear that at the European level, the recovery of the aid is separate from the decision finding the aid to be unlawful: it does not arise immediately upon such a finding. For example, in Belgium v Commission (Tubemeuse), AG Tesauro noted that:
It is true that the Commission normally notifies recovery of aid unlawfully granted in the same Decision which rules on compatibility. However, those two aspects of the question are independent of each other, to such an extent that, as the Court has observed, recovery ab initio of aid may be requested even if the aid is, at the same time, declared compatible, but only with effect from the future . . . The order to recover the aid is not an automatic consequence of unlawful implementation. It is the Commission which decides whether it should or should not adopt such an order on the basis of assessments in regard to which it has a wide discretion.92
The national court’s obligation to recover the aid thus comes from the separate exercise by the Commission of its discretion under Article 108(2), and there is no unjust enrichment event at the European level, only a legislative requirement that the Member States provide their own ‘actionable story’ to recover the money. Similarly, in the case of levies by National Intervention Agencies, some levies on behalf of the EU but in contravention of EU law are covered by secondary legislation granting a right to reimbursement. Thus, for example, in the case of FMC v Intervention Board for Agricultural Produce,93 Regulation No 1922/92 (amending Article 4 of Regulation No 1633/84) set out the applicable rules for calculating the amounts which operators who could rely on the invalidity could claim by way of reimbursement. Another example is Regulation No 450/2008 of the European Parliament and of the Council of 23 April 2008 which laid down the Community Customs Code (Modernised Customs Code).94 Here again, while the Member States may translate the relevant requirements into unjust enrichment actions, the European rules are legislative and require no more than that some kind of ‘actionable story’ is provided at national level.
The second important reason why it is difficult to find examples of Model 2 reasoning in relation to unjust enrichment is that when given the direct choice, the ECJ has tended to opt specifically for Model 1. Metallgesellschaft and Hoechst v IRC95 is of course well known to English lawyers as the starting point for the Deutsche Morgan Grenfell case discussed above in part one.96 To reiterate the facts,97 Metallgesellschaft and Hoechst concerned two separate cases involving companies resident in England whose parent companies had their seats in Germany. Companies whose parent is resident in England as well as companies whose parent is not, had to pay Advance Corporation Tax (ACT)98 and Mainstream Corporation Tax (MCT).99 The ACT was payable first, but could then be set off against the MCT paid later. Under the relevant legislation, two companies resident in the UK, one of which held at least 51% of the other, could make a group income election (GIE). The result of such an election was that the subsidiary could choose not to pay ACT on dividends which it paid to its parent company. Obviously, since the applicant companies in this case did not have parent companies resident in the UK, it was not possible for them to make a GIE, and thus ACT was payable on all the distributions made by them to their parent companies. This meant that, although this ACT was later offset against MCT, in the meantime they did not have use of the funds used to pay the ACT. Subsidiaries whose parent companies were also resident in the UK would have been able to retain the money in comparable circumstances by making a GIE. Metallgesellschaft and Hoechst (M&H) instituted proceedings before the High Court against the IRC, in which they sought a ruling that they had suffered loss by virtue of the fact that the distribution of dividends by the subsidiaries to their parent companies had been subject to ACT,100 maintaining that they had suffered a ‘cashflow disadvantage which subsidiaries of parent companies resident in the UK did not incur’,101 a disadvantage which they argued was indirect discrimination on the grounds of nationality contrary to the EC Treaty. The companies were not therefore claiming return of the principal sum: the ACT had already been offset against MCT by the time the proceedings were brought. The case instead concerned a claim for interest on the sum for the period between payment of the ACT and payment of the MCT. The High Court stayed the proceedings and sent various questions to the ECJ for a preliminary ruling under Article 267 TFEU (ex 234 EC), including the question whether the facts gave rise to a restitutionary right on the part of the claimant, or whether the interest was available, if at all, only through the Court’s case law on damages for breach of EC law.102 The High Court also asked whether the failure of the claimants to rely on EC law to challenge the unavailability of the GIE provided an answer to the claim.103
The ECJ ruled that it was indeed contrary to the right of establishment enshrined in Article 49 TFEU (ex Art 43 EC) of the Treaty for the tax legislation of a Member State to afford companies resident in that Member State favourable tax treatment when their parent company is also resident in that Member State, but not when the parent company has its seat in another Member State.104 This set the scene for the answer to the second question, whether the companies could claim the interest, either as a claim for restitution, or as a claim for compensation for damage resulting from breach of EC law. The UK Government argued that ACT should not be regarded as a tax levied contrary to EC law, since the subsidiaries were in any event bound to pay the tax. The companies’ complaint, they argued, was the lack of availability of the GIE option, and this should be considered as giving rise to non-contractual liability.105 This seemed to be motivated by the fact that the rules on interest and perhaps on mitigation would have been more favourable to the UK had the claim been classified as a wrong.106 Second, the UK Government argued that if they were correct on the first point, and the claims were to be treated as claims for recovery of sums paid in breach of EC law, it was settled law that it is for the national system to determine whether interest is payable.107
In answer to the first aspect of the reasoning, on ‘categorisation’, the Court answered that:
It must be stressed that it is not for the Court to assign a legal classification to the actions brought by the plaintiffs before the national court. In the circumstances, it is for Metallgesellschaft and Others and Hoechst to specify the nature and basis of their actions (whether they are actions for restitution or actions for compensation for damage), subject to the supervision of the national court.108
However, the ECJ then reiterated the well-known case law109 on the right to refunds of charges ‘levied in a Member State in breach of rules of Community law’, indicating that an unjust enrichment analysis may be available. Turning to the question of national procedural autonomy on the question of interest, the Court continued by distinguishing this case from Roquette Frères110 and Express Dairy Foods,111 in which questions relating to interest were held to be for the national court. It pointed out that:
In the main proceedings . . . the claim for payment of interest covering the cost of loss of use of the sums paid by way of ACT is not ancillary, but is the very objective sought by the plaintiff’s actions in the main proceedings. In such circumstances, where the breach of Community law arises, not from the payment of the tax itself but from its being levied prematurely, the award of interest represents the reimbursement of that which was improperly paid and would appear to be essential in restoring the equal treatment guaranteed by . . . the Treaty.112
As a result, the case was not one of a claim for interest on a principal sum in the way such a claim is usually understood, but one in which the interest was the principal sum.113 The Court concluded that Article 49 TFEU (ex Art 43 EC) entitled the companies to obtain interest accrued on the ACT paid before the MCT was payable ‘and that sum may be claimed by way of restitution’.114 On the other hand, it might not have to be so claimed, since if the case were to be considered as one of compensation for loss under the principle of state liability in damages, in order fully to repair the damage caused by the breach of Article 49 a sum representing the amount of interest would also be available. The UK Government’s argument that the plaintiffs could not be awarded interest if they sought compensation in a claim for damages could not be accepted since it would render exercise of the right (practically) impossible. Thus while the ECJ refused to classify the relevant event in this case, it did make it clear that from its point of view both of the categories were open, and that the UK Government could not succeed in either case.
AG Fennelly, on the other hand, was not as reticent as the ECJ on the matter of the categorisation of the claim. He similarly concluded that the restriction on the availability of the tax advantage was incompatible with Article 49 of the Treaty,115 and held that in terms of effectiveness, the mere fact that the loss was the result of only a temporary violation of EC law could not justify the refusal of a claim for around £800,000. He too, therefore, rejected the submission of the UK that the claim could not be regarded as restitutionary in nature.116 So far AG Fennelly had gone no further than the ECJ, but he continued:
I believe that it is more correct and more logical to treat the plaintiff’s claim as restitutionary rather than as a compensatory claim for damages. ACT was . . . extracted from them in contravention of Community law and, therefore, unlawfully. In the period between payment of ACT and its being taken into account in respect of the [M]CT liability of the subsidiaries, it should have been repaid to the plaintiffs by the United Kingdom. If it had been possible to bring legal proceedings during this period, the plaintiffs would, in my view, have been entitled to interest. It is neither logical nor just to deprive them of this entitlement merely because, in the meantime, the liability of the United Kingdom to repay the principal sum has been discharged. In a practical sense, also, the claim for interest is closer to a restitutionary rather than a compensatory claim. The underlying sums are known and indisputable. All that is necessary is for the national court to establish an appropriate interest rate for the relevant period.117
In the alternative, AG Fennelly concluded that, should the Court disagree that a ‘restitutionary claim’ should be available, Article 49 did create rights for individuals;118 that there was a causal link between the statutory exclusion of the companies and the loss suffered by the claimants; but that there was some question whether the breach was ‘sufficiently serious’ as required by Brasserie/Factortame III, although this was a matter for the national court to decide.
In answer to the fifth question on mitigation, the Court held that, had the companies applied for a GIE, they would in any case have had their request refused, and had they appealed successfully after payment of the ACT they would still not, under English law, have had any right to reimbursement. Thus any reduction of the amount awarded on this basis would be contrary to EC law.119 Similarly, AG Fennelly, who referred to this question as laches, concluded that the Member State may not plead that the taxpayer should have made use of statutory remedies and/or the primacy of EC law in order to reduce or exclude a claim, since ‘a Member State must not be allowed to profit from its own wrong’.120
In terms of the analysis proposed here, AG Fennelly’s opinion is relatively straightforward, being a clear example of Model 2 reasoning in which he clearly considers the two categories of causes of action and chooses the event of unjust enrichment (although in terminology he refers instead to a ‘restitutionary claim’)121. The reasoning of the ECJ is more difficult, and in fact appears to fall somewhere between the two models. Since the ECJ clearly distinguishes between claims for state liability in damages under Francovich122 and claims for restitution, to this extent its reasoning falls into Model 2, but significantly, while acknowledging the difference between the types of claim, the ECJ refuses to classify the actions itself, stressing that this is for the national court. In terms of the specific rules relating to recovery in the case, the Court merely returns to its previous case law on equivalence, effectiveness and practical possibility. Furthermore it does not, even through these tools, define the event of unjust enrichment.123 It is therefore clear that the ECJ’s preference was for Model 1 reasoning, in which the whole ‘actionable story’ is to be provided at national level.
However, while this may be the case whenever the ECJ is able to exercise a conscious choice, there are circumstances in which it may not wholly be able to do so. It will be recalled from the discussion above that the ECJ’s current approach to the imperatives of ‘equivalence’ and ‘effectiveness’ is to balance the reasons in favour of the relevant national rules against any detrimental impact they have on the application of EU law. This means that in some cases the ECJ will inevitably have to become involved in the reasoning behind the relevant national rules in order to determine whether that reasoning outweighs the EU concerns. In this context, the extent to which the ECJ must therefore move inevitably towards something more like Model 2 reasoning depends in part on the nature of the national rule in question. The result is that in some instances the nature of the question is such that while the Court still reiterates explicitly the principle established in Metallgesellschaft and Hoechst, implicitly, as a matter of substance, it has had to move away from that declaration. For this reason, there has been a slight difference between situations in which money has been levied in contravention of EC law by a Member State or NIA and those in which it has been wrongly paid out by those entities.
Cases in which a Member State or a National Intervention Agency
Levies Money in Contravention of EU Law
In the very early cases in this context the reticence of the EC was such that in Rewe-Zentralfinanz124 and Comet,125 (where, as noted above, the principles of equivalence and effectiveness were first established) AG Warner held that he did
not think that any such independent right of action was conferred on the plaintiffs by Community law: I think that it was for their national laws to lay down the remedies to which they were entitled in consequence of the invalidity of the fiscal legislation in question.126
This is a view which he reiterated two years later in Pigs and Bacon Commission v McCarren and Co Ltd.127 Similarly, in Roquette Frères,128 where the money had been levied by an NIA rather than the Member State, the Court held that:
Disputes in connexion with the reimbursement of amounts collected for the Community are thus a matter for the national courts and must be settled by them under national law in so far as no provisions of Community law are relevant.129
However, in Pigs and Bacon, the ECJ held that since the pigmeat levy at issue was ‘not due’, ‘any trader who is required to pay the levy has therefore the right to claim the reimbursement of that part of the levy which is thus devoted to purposes incompatible with Community law’.130 This then paved the way for the establishment of the famous San Giorgio ruling that
entitlement to the repayment of charges levied by a member state contrary to the rules of community law is a consequence of, and an adjunct to, the rights conferred on individuals by the community provisions prohibiting charges having an effect equivalent to customs duties or, as the case may be, the discriminatory application of internal taxes.131
Nevertheless, this still only meant that the Court was following Model 1 in requiring the national system to provide an actionable story for recovery of the money. The slightly more significant change came when the Court began to examine the impact of various national defences on the effectiveness of the remedies that those Member States were providing in this regard. Where the situation has involved a private party claiming from a public defendant, whether that public defendant was a Member State or an NIA, the defence often pleaded132 is that of passing on, as the public entity seeks to argue that the claimant did not bear the charge personally, but passed it onto its clients or customers in its prices. In other words, as argued in chapter five, the essence of the defence is to argue that even if the defendant was enriched, this was not at the expense of this particular claimant as would be required for a successful unjust enrichment action. Originally this defence, confusingly known in the EU context as ‘the defence of unjust enrichment’ (on the basis that recovery would lead to the unjust enrichment of the claimant) was only recognised as a possibility by the Court in cases such as Hans Just,133 Amministrazione delle Finanze dello Stato v Denkavit Italiana134 and Amministrazione delle Finanze dello Stato v Ariete,135 the ECJ holding that ‘Community law does not prevent the fact that the burden of the charges which have been unlawfully levied may have been passed on to other traders or to consumers from being taken into consideration’.136 However, in San Giorgio,137 the Court held that the passing on defence was contrary to EC law in as much as it was ‘subject to rules of evidence which render the exercise of that right virtually impossible, even where the repayment of other taxes, charges or duties levied in breach of national law is subject to the same restrictive condition.138 For this finding of incompatibility, as compared to the cases in which the Court had simply let the defence stand, AG Mancini had to consider the defence of passing on in more detail than, for example, AG Reischl in Hans Just. Thus AG Mancini held that in principle: