László Blutman[1]:

Preclusive Effect of Directives: An Explanatory Framework for the

Post-Lisbon Era

The preclusive effect of directives is of special significance when they are applied in legal disputes between private litigants. The relevant case law starting with Pafitis seems to run against the principles laid down previously in the Marshall/Foster/Dori doctrine. This article is seeking to establish a conceptual framework for a possibly more consistent explanation of the Pafitis case law than those having been offered so far. The author takes stock of those general legal effects that a directive may produce in internal litigation, explores the features and functioning of preclusive effect, reaches back to and elucidates the distinction between direct effect and direct applicability and identifies four levels as to how directives operate in domestic litigations. The conclusions drawn from these considerations make way for fitting the Pafitis case law into the Court’s general jurisprudence relating to directives. 

I.  The Pafitis case law 

Since the beginning of the 1970s a question has been looming permanently on the horizon, namely, under what conditions directives can have direct effect in the domestic law of Member States. After van Duyn remaining isolated from the subsequent case law, the relating and well-known principles have been developed and reinforced by the European Court of Justice in the Marshall(I), Foster and Faccini Dori cases (and later repeated and refined many times in other cases).[2] The Marshall/Foster/Dori doctrine based on the distinction between horizontal and vertical legal effects has been facing a serious challenge since the 1990s as legal practice has gradually brought to the surface some shortcomings of this construction in legal disputes between private parties (horizontal disputes).[3] Commentators usually attribute the outset of the erosion of these principles to CIA Security International, though this decision (dated 30 April 1996) was preceded by two other decisions in Pafitis (dated 12 March 1996) and Bernáldez (dated 28 March 1996) where the directives in question had also been assigned specific legal effects in horizontal litigation.[4] I therefore will call the cases raising the problem of the horizontal, preclusive effect of directives Pafitis case law.[5] 

Main features of these cases lie in that the Court considers directives applicable in disputes between private parties and they preclude the application of national provisions contrary to them. This preclusion, however, usually prevents a private party to litigation from asserting its rights rooted in the national legal rule which the application of the directive has precluded from applying in the case. Thus, this preclusive effect of directives comes suspiciously close to direct effect which is not allowed in disputes between private parties under Marshall/Foster/Dori principles and has brought tension in the case law. By referring to “sufficiently precise and unconditional” directive provisions (e.g. in CIA Security International) the Court highlighted the usual criteria of the direct effect, and thus gave the impression that the preclusive effect of directives in such circumstances coincides with direct effect. So, with the Pafitis case law various speculations have come to light as to how these decisions can be reconciled with previous practice of the Court. 

This article is not about arguing for or against the horizontal direct effect of directives but seeking a relatively consistent explanation for the Court's relevant case law. The basic problem is how cases belonging to the Pafitis case law can be made consistent with the Marshall/Foster/Dori doctrine, and whether the distinction between vertical and horizontal effect can be maintained.[6] Various attempts have been made to give comprehensive answer to the problems these cases raised, but each of them has more or less failed.[7] Even the most elegant and viable theory based on a distinction between exclusionary and substitution effects of directives has suffered serious and sound attacks.[8] Hartley took the view that “no theory provides a full explanation of the twists and turns of the Court’s case law.”[9]  

In what follows I will investigate whether there may exist other explanation, or at least a conceptual framework for other explanations, more comprehensive and consistent than the previous ones. For this purpose I shall reinterpret the concept of direct applicability against direct effect in light of legal effects of directives. As the central concept of the Pafitis case law is the preclusive effect of directives an evident starting point is to locate its place among other general legal effects of directives in national legal proceedings. 

II.  Six general effects of directives in national legal proceedings 

In explaining the Pafitis case law at least four essential factors should be reconciled:[10] (1) some directives can be applicable in legal disputes between individuals; (2) the directive applied this way can have a preclusive effect on the applicability of a national provision; (3) nevertheless, the directive has no direct effect, i. e. obligations of private parties cannot be based on a directive; (4) and distinction between vertical and horizontal disputes and/or legal relations must be maintained. It seems to me that two problems cannot be evaded: that of direct applicability and the question of when obligations are imposed on a private person in a legal dispute, which is a standard point recurring in the Court's case law. 

The Court of Justice established in the Pafitis-type cases a legal effect (preclusive effect) that is less than direct effect (i.e. full effect), but is more than the effect traditionally termed as indirect effect (duty of consistent interpretation). Since one of the legal effects of directives prevailing in horizontal relations have to be explained and put in context it must, in some way, be outlined what general legal effects a directive may have in domestic legal disputes. On the other hand, as direct effect is the full effect (l'invocabilité maximale) of an EU measure (directive), and in horizontal relations the preclusive effect of directives is something less (l'invocabilité partielle), direct effect will be the evident reference point in evaluating these effects.[11]  

The various general legal effects of directives can overlap, but they can be adequately specified in six forms according to the case law of the Court:[12] (1) a party to the dispute can rely on the directive;[13] (2) the directive can confer rights on a party;[14] (3) the directive can impose obligation on the party;[15] (4) the national court is required to set aside a national provision which is contrary to the directive;[16] (5) the national court is bound, so far as it is possible, to use the relevant directive in interpreting and construing national legal rules (conforming interpretation);[17] (6) the directive precludes the application of a national provision contrary to it (preclusive effect). 

In examining these legal effects in terms of direct effect, it becomes clear that case (5) does not presume direct effect as, in case of conforming interpretation, the EU norm can only assert its legal effect through and by force of the interpreted national provision and in this sense the legal effect is “indirect” as opposed to direct one.[18] 

The Court of Justice frequently examines whether a private person can invoke a directive in internal litigation, see case (1). There were times when this possibility of invoking an EU legal provision in an internal dispute in itself indicated direct effect in the Court's case law.[19] Nevertheless, the von Colson, Grimaldi, Marleasing decisions laying down the principle of conforming interpretation washed away the importance of invocation of  EU law in abstracto and made it obvious that case (1) cannot simply be the manifestation of just the direct effect, since a private party can rely on an EU norm with a view to fostering the adequate interpretation of a domestic legislation, i.e. to producing indirect effect. For this reason, the Court tends to refer to the possibility of relying on an EU norm in less and less general terms. Rather, it makes use thereof in connection with whether someone can invoke the norm in order to enforce a right or obligation, thus the Court connects case (1) with (2) and (3). The possibility to rely on an EU legal norm (directive) has become a kind of ancillary formation which does not exclude per se its independent appearance. But Blumann and Dubouis point out correctly that talking about the invocability of a directive in the abstract just conceals the nature of the legal effect for which the provision is invoked.[20] 

Case (4) is considered to be the classic formula given in Simmenthal (II), and it essentially looks identical with the preclusion in (6). For instance, in Bussone also decided in 1978, the Court of Justice identified the legal effect of a directly effective regulation as “precluding the application of any legislative measure”.[21] This formally means as if a court would set aside the national provision. Thus, the legal effect in (4) and (6) can be identified with each other. 

Furthermore, case (2) appears not to be a necessary criterion of the direct effect, although the Court of Justice traditionally, from van Gend & Loos on, connects direct effect of a provision with its creating rights for the parties to the dispute. When it held in Danske Slagterier that the treaty provision at issue “has direct effect in the sense that it confers on individuals rights...” it just repeated a regular statement of direct effect doctrine.[22] This, however, is not always a condition for a directly effective directive. There were a number of environmental protection cases where owing to the absence of environmental impact assessment private persons tried to enforce obligation based on directive against State entities in vertical disputes without having any substantive, determinable, correlative right. Despite this, the Court found the directive emerged in such cases applicable: lack of a right identifiable on the part of the private person invoking it did not make an obstacle to the direct effect of the directive.[23] So conferring rights by a directive provision indicates its direct effect, but this is not a necessary condition thereof. 

As a result, only case (3) can be the one that is the indispensable concomitant of the direct effect: inasmuch as an obligation based on a directive falls on one party in a national legal dispute, it is without doubt the essential mark of direct effect. This is the most restricted, but indispensable legal effect of a norm with direct effect. If the Court of Justice wishes to uphold the Marshall/Foster/Dori doctrine, it cannot permit that in a horizontal dispute a directive be applied so that obligations will be imposed on a private party originating in the directive provisions. 

The central question is the relation of case (6) to case (3): if, according to the Court, a directive “precludes” a national provision, what effect may be attributed to it then? This is a problem of key importance because in the Pafitis case law the particular legal effect of directives can be identified as preclusive effect.[24] 

III.  Preclusive effect of directives – in context 

Establishing the preclusive effect of directives in horizontal disputes is not without any antecedent.[25] On the contrary, given the Court’s previous case law, the 1996 breakthrough cannot be regarded as unexpected (at least in hindsight). Before the Pafitis case, there had been at least two decades of the Court's jurisprudence where it had attributed preclusive effect to (directly effective) EU measures in internal legal disputes. 

But before the Pafitis case the Court had referred even in various horizontal disputes to the fact that a directive provision “precluded” a national practice or national provision. The Dekker/Habermann-Beltermann case law can be regarded as such, in relation to Directive 76/207/EEC.[26] Nevertheless, these cases did not bring about any breakthrough because the Court used the expression “preclusion” for contrasting the content of the directive with that of the national legislation, and not in a sense of legal effect for narrowing the scope of application of an internal legal rule.[27] But later on, the Court of Justice started referring eventually or in a downright manner to the preclusive legal effect of a directive provision using the expression “preclusion”. This “slipover” marks the essential and conceptual change. 

The rise of the Pafitis case law happened among particular circumstances as far as the first case (the Pafitis case) is concerned. The Court had applied directly the company law Directive 77/91/EEC in question in two earlier (vertical) cases and a provision of the directive had precluded certain national legal rules.[28] In Pafitis the same provision of the same directive (Art. 25) had been brought up under very similar circumstances, in this case now involving a horizontal dispute. The Court of Justice also established the preclusive effect of the said directive provision as it had done in the previous cases, although this case was one between private litigants.[29] Transferring previous conclusions drawn in vertical disputes to a horizontal dispute indicates the “slipover” I have referred to which shows essential change on the merits.

Preclusion of a national provision is not denoted only by the term “preclude”, but also by the formula “setting aside” used by the Court in paragraph 22 of the Simmenthal (II) judgment or in other instances the expression of “non-application” or “disapplication”, too. All of them refers to the ad hoc restriction of the application of a national rule and the effect of the applied EU measure, but moreover the term “preclude”, by its double meaning, can also refer to the material relationship of a directive and a national provision. Due to this double meaning of “preclusion” it is uncertain in many cases in what sense the Court uses the term (as opposed to “setting aside” or “disapply” which clearly refer to the effect of the EU measure).  

As the preclusive effect of directives (and thus the preclusion of any contrary national provision) is based upon Article 4(3) EU and Article 288 FEU (ex Articles 10 and 249 of the EC Treaty), and all things considered, upon primacy of EU law,[30] the preclusion of a contrary national provision appears this way as such special “measure” to which now Article 4(3) EU (ex Article 10 EC) is referring. Nevertheless, the preclusive effect of directives is restricted: special character of the directive as a legal source (framework law directly addressed to the Member States themselves), the principle of legal certainty and limits of the EU competence must be taken into consideration as restrictive factors, which prevent directives from imposing obligations on a private party, i.e. which prevent preclusive effect of directives from turning into direct effect in horizontal disputes.  

All this leads to the issue of the relation of preclusive effect to the direct effect itself. The Court of Justice did expressly refer to this in the relatively early Marimex case: a regulation having direct effect has a legal effect to preclude the application of a national provision contrary to it.[31] This, however, does not exclude the possibility that norms disposing of legal effect other than direct one can have such preclusive effect. This position is shared by Advocates General Léger and Saggio, as well as certain commentators in their wake, who think that the preclusive effect is not just the feature of direct effect, but that of the primacy of EU law. [32] 

I am inclined to agree with this approach, but in different conceptual framework. It is clear if we wish to maintain the Marshall/Foster/Dori doctrine, then preclusive effect cannot be attributed exclusively to directly effective measures. In Pafitis case law the Court speaks of preclusive effect in a sense that it nevertheless does not acknowledge direct effect of directives applied this way. The preclusive effect cannot therefore be an effect of solely those EU legal norms which give rise to direct effect.[33] Though I admit that preclusive effect is an evident consequence of the primacy of EU law, but in horizontal litigation directives without being directly effective must be directly applicable to be able to produce preclusive effect. So I regard preclusive effect as a feature of directly applicable EU measures (including directives). That is I will be arguing for in the following. 

IV.  Directives: how to relate preclusive effect to direct applicability? 

1.  Reassessing direct applicability against direct effect 

Under the Pafitis case law it seems there is an increased role of the distinction as to what is meant by directly applicable norms as opposed to norms producing direct effect.[34] However, the clear distinction between direct applicability and direct effect is not a primary question: neither in the case law of the Court of Justice nor in the works of commentators.[35] The Court frequently uses direct effect and direct applicability interchangeably.[36] Some authors consider these two phenomena - if not necessarily in theory, but along a more practical approach - to be identical.[37]  

In my view, identifying direct applicability with direct effect, even from a practical aspect, cannot be maintained when dealing with horizontal effects of directives. Therefore I assume, as for example Steiner and Woods do, that direct applicability is a necessary (but not sufficient) precondition of direct effect,[38] as it would be hard to conceive that a legal provision produces legal effects in a case but at the same time it does not apply to this particular case.[39] This claim finds support in the Court's case law: e.g. in Grad the Court has taken the view that “... [regulations] are directly applicable and therefore by virtue of their nature capable of producing direct effects...” regarding direct applicability as a precondition of the direct effect.[40] If we claim that a directly non-applicable rule cannot have direct effect, the reverse question seems also important, too: is there, and if so, under what conditions, a directly applicable EU measure which on the other hand does not have a direct effect? If yes, what legal effect can it produce?[41] This distinction is essential to be able to explain the Court’s case law in connection with the horizontal effects of directives. 

By direct applicability I mean that a legal provision being legally sufficiently complete[42] is able to enter national legal (judicial and administrative) practice and apply by itself in particular legal proceedings without other (e.g. transforming, implementing or supplementary) legal rules. This description does not rule out that a measure might require further measures for its proper or full implementation, but not for its entry and independent (albeit sometimes limited) application in national legal disputes. This is a more restrictive description than that we usually come across and which roughly follows the Court's statements made in Variola and Zerbone emphasizing the capacity of directly applicable measures of entering into national law without any further measure. [43] This approach fails in case of directives at least in two respects. Under such definition one does not find answer to the question of whether or not non-transposed or mis-transposed directives are part of a Member State's national legal system? Shaw's answer is in the positive,[44] but if so, such directives or some of their provisions are directly applicable measures like regulations? They are clearly not. Furthermore, it does not explain “indirect applicability”, which should exist if direct applicability exists and should also be accounted for. Thus, I rather take direct applicability as a capacity of a legal provision of entering national legal practice and applying independently of other legal rules in particular legal disputes.[45] 

On the other hand, by direct effect I mean in most traditional sense that a legal provision is capable of giving rise to rights and obligations enforceable by a court.[46] But, in what sense can a legal effect be direct or indirect? In my view a provision is directly effective if it can be used as an independent legal basis for the legal assessment of the particular factual situation in terms of rights and obligations. A provision is only indirectly effective, if its legal effect is aimed at and influences only other legal rules applicable in the legal dispute: an indirectly effective provision may constitute a basis for only assessing another legal rule, therefore it only "indirectly" reaches the facts of the case through influencing the meaning or application of other rules.[47] In this sense both preclusive effect and the duty of conforming interpretation is an indirect legal effect of directives. 

In what sense can the application of a legal provision be direct or indirect? According to the description I have given previously a provision is directly applicable if in a legal dispute it has the capacity of standing and applying all by itself without the help or medium of other legal rules. It follows that EU directives used by national courts to perform their obligation of conforming interpretation cannot be treated as directly applicable ones because in such disputes they serve as means of determining the meaning of another legal provision which the national court construes or interprets. In such situations the directive can only apply in the case indirectly i.e. by virtue of another (national) provision under interpretation. On the other hand, a directive producing preclusive (indirect) effect does not apply by force of a national provision, but it applies independently of and against another national provision to be precluded. In this sense a directive having preclusive effect is directly applicable because it stands alone against the contrary national legal provision to be precluded and neutralized. 

In sum, the directly applicable legal rule without direct effect makes its way into the field of judiciary, thus enters into the courtrooms and reaches the judge’s pulpit: the judge has an obligation to take it into consideration and, if necessary, apply it for specific purposes (e.g. using it for interpretation of another rule). This rule being without direct effect, however, does not reach the legal tilt-yard and the parties in the sense that the rule cannot provide rights to or impose obligations on the litigants, i.e. it may produce legal effects only indirectly, through other rules. On the other hand, a directly effective rule does not only reach the pulpit, but also the parties, and from then on, rights and obligations can be derived directly therefrom.[48] 

From all this it follows that one can distinguish three levels on which a directive (until its proper transposition) may produce general legal effects towards Member States: 

Level(I) - It obliges the State only as a legislator if the directive is applicable neither directly nor indirectly;

Level(II) - It obliges the Member State (its legislative, executive and judicial branches) if the directive is indirectly or directly applicable, but does not have direct effect;

Level(III) - It obliges the Member State as legislative, executive and judicial organization and also the State as a litigant if the directive is directly effective. 

In horizontal disputes the preclusive effects produced by directives can be located on Level (II), as its essence consists in that a directive prevents the application of a contrary domestic rule without creating rights for or imposing obligations on the parties. In this case the directive has no direct, only indirect effect, but is directly applicable. (From another angle, since a directive may not have direct effect in horizontal legal disputes according to the Marshall/Foster/Dori doctrine, the preclusive effect of a directive may only prevail in these cases if by virtue of its provisions no right or obligation is created by precluding a contrary domestic rule.) Of course, preclusive effect appears both on Level (II) as well as Level (III) because it goes not only with directly applicable but as well as directly effective rules; in fact preclusive effect being a standard form of direct effect is a natural concomitant of directly effective EU legal rules.[49] On Level (II) may another widely-known form of indirect effect appear when, according to the von Colson/Grimaldi/Marleasing case law, an obligation is imposed on the national court to interpret national law consistently with EU measures applying indirectly and producing only indirect effects.

2.  Applicable measures (directives) without direct effect in the Court's case law 

In spite of conceptual uncertainties, there is certain tradition of making distinction between direct applicability and direct effect in the Court's case law (although a great number of examples contrary to this could be mentioned). In some earlier decisions the Court of Justice distinguished the two phenomena in the case of regulations,[50] or even has been forced to do so confronting the difficulty that not all of the regulations can give rise to direct effect because lack of precision or the regulation itself may empower Member States to adopt further measures for its implementation.[51] That was the case in Monte Arcosu where a provision of the regulation at issue was directly applicable without, however, producing direct effect in the case.[52] 

A new aspect of this distinction definitely started to shape in the Grosskrotzenburg case a couple of months before the Pafitis, Bernáldez and CIA Security International decisions.[53] Here the Court in an infringement action under Article 226 EC (now Article 258 EU) held that the State obligation stemming from a directive is objective in nature and can be applied and enforced by State authorities even if no private persons rely on its provisions in defense of their rights. The Court separating in procedural sense the obligation falling upon the State from any correlative rights to be invoked by an individual opened up the way for the well-known Kraaijeveld/Linster/Wells case law.[54] The importance of these Level (III) cases lies in the Court's making it clear that private litigants may rely on untransposed or mistransposed directives and obligations based on them against the State or its entities without being able to assert any correlative rights. So on Level (III), directives may produce direct effect against a State in an internal legal dispute without conferring rights on other litigants,[55] or put another way, conferring rights on litigants is a sufficient but not necessary precondition or characteristic of a directly effective legal provision.[56] 

Separating a general and objective State obligation to follow and apply a directive from the actual legal position of the private parties in a legal dispute (in terms of conferring rights or imposing obligations),[57] the Court reinforced the applicability of directives also in horizontal disputes on Level (II) even if these directly applicable directives produce legal effects stopping short of direct effect. This will take us to the situation where the preclusive effect of a directive may, in theory, prevail in horizontal disputes. Let us suppose the directive at issue does not confer rights and impose duties on the private litigants, but there is a contrary national provision the application of which can be conclusive to the outcome of the case. One party invokes the directive so as to preclude the application of the national provision detrimental to the invoking party's legal interests without thereby imposing any obligations based on the directive on the other private party. What can the national court do? The Member State and its court has a general and objective obligation to insure the prevalence of the directive under Article 4(3) EU (ex Article10 EC) irrespective of whether or not the application of the directive serves as a tool for enforcing private litigants' rights, that is irrespective of the private litigants actual legal position in the dispute, except in one respect, that application of directive may not give rise to obligations against the other party. This general and objective obligation can be discharged in an ad hoc manner if the court excludes the application of the national provision in question from the case. In this instance, the directive becomes applicable, but it will have no direct effect whatsoever. This is the situation which emerged in a range of cases such as in Candolin, Pafitis and Bernáldez.[58] 

3.  What renders a directive without direct effect directly applicable? 

Not every directive can prevail in the above mentioned way, since direct applicability - and not only direct effect - comes with preconditions. As we take direct applicability as being in itself a prerequisite of direct effect, it is not so obvious for the first glance whether the often invoked “clear, precise and unconditional” requirements are conditions of the direct effect or those of the direct applicability, which become the conditions of the direct effect because the direct applicability itself is a characteristic of directly effective rules.[59] I assume that the condition of direct applicability of not directly effective directives may also lie in that the legal provision at issue be “unconditional and sufficiently precise”.[60] 

If we take as a starting point that the specificity of a directly applicable directive is that it enters the field of judiciary and the courtrooms on Level (II), but not creating directly rights in favour of or obligations against the litigants, an evident question may come in: Are the requirements of clarity, precision and unconditionality needed for a directive (or generally an EU legal norm) to be able to enter the field of the judiciary, or just to create rights and obligations? When the question is put this way, we can reach some particular conclusions. 

A norm must be sufficiently clear and precise in order for its content to be identifiable and applicable directly on Level (II), whereas in case of direct effect in order for the rights and obligations included in it to be identifiable and judicially enforceable on Level (III). The “sufficiently precise and clear” condition can thus be generally a condition with both directly effective and directly applicable provisions - only the degree and form of precision and clarity can be different at most.[61] Unequivocality as the Court used the term in paragraph 52 of the judgment in Marshall (I) means a degree of precision which indicates a precondition for direct effect. But a directly applicable provision not producing direct effect is also conditional upon certain clarity and precision. Dutheil de la Rochère pointed out that a directive should be sufficiently precise to be capable of precluding a national provision, i.e. of having preclusive effect.[62] In von Colson, which was a Level (II) litigation, the directive at issue did not contain precise and unconditional obligation to produce direct effect, but it was precise enough to be able to provide basis for the national court's interpretation, i.e. to have indirect effect.[63] Naturally, to identify an obligation in a directive provision requires much higher degree of precision of the rule than to discern the purpose of the whole measure in order to use it for interpreting a national legal rule or for unveiling its inconsistency with a national provision. 

As far as unconditionality is concerned, the conclusion is of a similar nature. In Klattner the Court of Justice concisely set out the essentials of this condition: “A legal provision is unconditional where it lays down an obligation which is not qualified by any condition and is not made subject, in its implementation or effects, to the adoption of any measure either by the EU institutions or by the Member States...”[64] This “measure” can either be an adoption of a further legal provision or the exercising of some discretionary power left open by the legal norm. When direct applicability is at stake the question is whether the legislator has maintained the possibility of adopting a legislative act enabling the full application of the EU measure.  

The “unconditionality” is a criterion of both direct applicability and direct effect, having partial disparities in its content, though. Eridania and Monte Arcosu suggest that the Member State’s power ensured by a regulation to pass implementing measures may prevent the regulation having direct effect but is not bound in all cases to be prejudicial at the same time to its direct applicability.[65] For example, exercising a rather significant power of discretion with regard to the application of an EU measure can deprive the norm of its direct effect as exercising discretion may subject the existence of particular obligations to be based on this measure to various conditions, which cannot be directly derived from its provisions.[66] It is not, however, so much the case with direct applicability because the potential discretion does not necessarily deprive the norm of the possibility of its application in a case for the purposes of precluding a contrary national legal rule on Level (II).[67] 

As a conclusion we can claim that the sufficient precision, clarity and unconditionality are just as much of general conditions of direct applicability as that of the direct effect, but the specific content of these conditions is not the same in the two instances. Naturally, this conclusion is quite general and its practical application is not without problems.[68] Inferring from the Court’s case law, it is the in-depth analysis of norms on a case-by-case basis that decides these questions. Based on this conclusion, however, it becomes understandable why we need not think of the Court’s alluding to direct effect when it examined the conditions of ‘sufficient precision’ and ‘unconditionality’ in the CIA Security International case. Another point that follows from this is that the only essential criterion of the difference between a provision with direct effect and a directly applicable rule without direct effect is whether it can impose duties on private litigants. This is also deemed to be the limit of direct applicability in accordance with the Court's case law.

V.  What cannot be side-stepped: obligations of private party based on directive 

Serious disputes have arisen in various legal studies as to whether the Court attributed direct effect to the applied directives in the Pafitis case law.[69] Prima facie, some doubt may arise in this regard.[70] For instance, in the Unilever Italia case the application of the directive fundamentally changed the parties' legal position, and as a result of the application of the directive, the defendant (Central Food) had to pay the purchase price. In Candolin, the insurance company had to compensate for the damage incurred due to the preclusive effect of the directive. Were not obligations imposed on the private parties by reason of the application of the directive at issue? In this respect such cases rise two basic, but abstract questions as follows. How to assess the situation when the preclusive effect of a directive prevents a private party from asserting its rights rooting in the excluded national rule: Is it an obligation imposed on this litigant? On the other hand, do the directive applied in this way impose obligation on a private litigant if it precludes a national provision having hindered the other litigant to assert some substantive right in the proceedings? These questions concern the very nature of legal obligation at an almost philosophical level.[71] 

Though the Court seems to be very determined in that a directive must not create obligation in horizontal relations, as this is the last bastion yet standing between regulation and directive, the Court has in fact taken a very restrictive stance in what “imposing obligation” means in such legal settings. This extremely restrictive approach is grounded in and expressly or implicitly makes use of three distinctions. (1) Following the Court’s case law, a distinction should be made whether a legal obligation can be enforced against a private party resulting from the application of a directly applicable directive or this obligation is directly based on the applied directive. The limitation established in the Marshall/Foster/Dori case law does only apply for the second instance: an obligation entering in on account of the preclusive effect of a directive should not be directly based on this directive. (2) It also seems important that the directive precludes a general rule or just an exception, restriction, or a rule specific to a general or background national rule. (3) A third distinction which usually plays role is between the imposition of the legal obligation on a party and the legal effect of a directive merely aggravating the legal position or affecting adversely the legal interests of a private party.[72] 

If we share the assumption of Judge Lenaerts and Corthaut that there is a dividing line, albeit a very thin one, between precluding national legislation and imposing obligations on private parties,[73] that must run along these three distinctions that I have set out.[74]

The question is whether these distinctions or their combinations are capable of giving a general and meaningful account for the Court's decisions made up till now in relation to horizontal effects of directives. 

No obligations based on directives seem to be imposed on the litigants where the directive and the precluded national provision are of different subject matter and their interactions are either procedural in nature (CIA Security International, Unilever Italia) or just contingent even if they are substantive (Océano Groupo, Ingmar). Here, a private litigant will become subjected to an obligation or an additional legal burden by virtue of the preclusive effect of the directive, nevertheless, this obligation is not flowing directly from the directive but from other existing national legal rules (or contracts).[75] In these cases the exclusion/substitution distinction may work well having a strong explanatory power, and some elements of the Court's reasoning falls close to this theory. 

Another pattern appears in cases, where the domestic law contains an exception or restriction to a general legal rule, which the directive at issue does not acknowledge and does not allow either. The exception or restriction by virtue of which a private party would be released from an obligation is lifted by the preclusive effect of the directive, and the party will remain bound to perform the obligation. In these circumstances the directive may preclude an exceptional national rule passed for extraordinary occasions in contravention of the directive (Pafitis), or a conflicting national marketing restriction which the directive does not allow (Unilever Austria), or an additional condition set forth by a national provision unlawfully limiting the exercise of a right protected by the directive (Bellone, Centrosteel, Axa Royale) or an exemption absolving an insurer company contrary to the directive from general liability based on national legal provisions (Bernáldez). However, if an exception to the general national provision is granted and contained by an unimplemented directive itself and the exception would exempt one of the parties from performing obligations based on national legal provisions, this exception shall not be relied on in the dispute (Faccini Dori).[76] 

A directive may also have preclusive effect against such national provisions in a horizontal case which set limitations contrary to the directive on the degree of liability for compensation or the period of time for enforcing a claim in relation to economic activities regulated by EU law. It may preclude national legislature from applying which, in contradiction to the directive, lays down maximum limits of compensation or allows the compensation to be limited in a disproportionate manner in insurance relations (Mendes Ferreira, Candolin) or prescribes a limit for the amount of compensation claimed on account of infringement of the right protected by the directive (Draehmpaehl). Regarding other types of conflicting limitations the doorstep-selling directive (Council Directive 85/577/EEC) also precludes a national provision from imposing a time limit within which the right of cancellation may be enforced, if the consumer has not received information on this right (Heininger).[77] However, apparently running against Mendes Ferreira and Draehmpaehl, the Riunione Adriatica judgment cannot be explained within this construction, where Directives 73/239/EEC and 92/49/EEC could not remove limitation on the amount of insurance premium provided by domestic law contrary to them, because in the Court's view it would have imposed obligation on defendant.[78] 

In most of these cases the obligation imposed on a private party could be enforceable by reason of application of a directive, this obligation, however, was not derived from the directive itself. This is a rather narrow construction of when a legal duty is imposed on a private party, and it can well be criticised on account of the uncertain dividing lines. Nevertheless, the Court of Justice has deemed applicable the Marshall/Foster/Dori doctrine under these conditions in the past ten or so years. It is not clear how susceptible national courts will be to this niceties that can be sifted out from the Court’s case law. (To wit, it is the Court of Justice that regularly draws the national courts' attention to the requirement that no obligation derived from directives should be imposed on private parties.)[79] 

VI.  Connecting the dots: preclusive effect as an indirect effect of directly applicable directives 

A conclusion can be drawn from the foregoing, namely, that in the cases of the Pafitis line the application of directives in horizontal disputes can only be explained if, contrarily to what has been done so far, a distinction is made between the direct effect of a directive and its direct applicability. The directly applicable norm (directive) without direct effect has a feature of precluding the legal effects of conflicting domestic provisions, which, however, does not have the result of imposing obligations on private parties based on the preclusive directive provisions. The preclusive effect may mean an interim step in upholding EU law in national horizontal disputes. The various legal effects produced by directives in legal disputes are conditional upon the conditions set out in the Court’s case law. The degrees to which the prevalence in internal litigation of untransposed or mistransposed directives is measured can be summed up in the following way:[80] 

(1) Indirect effect of indirectly applicable directives: duty of consistent interpretation (provided that the directive and the national law can be reconciled by interpretation). A directive differing in content from a national provision can produce indirect effect in horizontal legal disputes insofar as the domestic court shall interpret and apply the national provision in conformity with the directive (von Colson/Grimaldi/Marleasing case law). The indirect character comes from that a directive provision is capable of influencing the establishment of the content of domestic law, but it can only prevail by virtue and by the mediation thereof.[81] 

(2) Indirect effect of directly applicable directives: preclusive effect (if the directive and the domestic law cannot be reconciled even by means of interpretation). In such instances the directive becoming directly applicable precludes the application of contrary national provision if the preclusive provision of the directive is sufficiently clear, precise and unconditional and does not impose obligation on a private party (the Pafitis/Bernáldez/CIA Security International case law). Naturally, it depends on the particular content of the provisions of a directive whether they are able to produce preclusive effect so their application cannot impose legal obligations on private litigants. 

(3) Direct effect of (directly applicable) directives (if the directive and the national law cannot be reconciled even by interpretation). Directives do not have direct effect in horizontal disputes. Direct effect can only emerge if the directive imposes an obligation on the State (or the emanations of State) that did not, or at least not satisfactorily, implement the directive (a Marshall/Foster/Dori case law). 

(4) Absence of Level (II) or (III) internal effects of ( inapplicable) directives giving rise to State liability (if the applicable domestic law is in irreconcilable conflict with the directive). If the above effects of directives are not produced because of the lack of a condition, the directive cannot exert Level (II) or (III) effects in national litigation and the national court has to decide according to the contrary national provision, which gives rise to the Member State’s liability under EU law (the Francovich case law).