EU legislation is said to have `direct effect' when its provisions can be relied on in national courts. That is, directly-effective provisions create rights and duties between individuals, which can be enforced in the national courts. There is a related, but not identical, concept -- DirectApplicability -- which states that certain provisions of EU law become national law without further enactment. `Direct applicability' is an institutional concept -- it concerns how law is incorporated. `Direct effect' is a remedial concept, concerning whether the law can be relied upon by individuals. The distinctions between direct effect and direct applicability have come about in an evolutionary way; most likely they were originally intended to have the same meaning (if, indeed, much thought was given to how direct applicability would be implemented in practice). In any event, it is now established that some provisions are directly effective but not directly applicable, some are directly applicable but not directly effective, and some are both directly effective and directly applicable.
`Direct effect' is not a concept that comes from the Treaties, or even from legislation. The Treaties are largely silent on how member states are to give effect to rights and duties under EC law. Although some member states have the administrative machinery in place to allow international treaties to become part of domestic law, most (including the UK) do not. In the UK it is normally necessary for Parliament to specify by legislation how a treaty is to be interpreted by the courts. For example, the UK ratified the ECHR more than fifty years ago, but it only became a source of law on the enactment of the HumanRightsAct1998. It is very likely that the original signatories to the ECTreaty believed they were entering into the same kind of agreement -- an international treaty requiring specific implementation in domestic law. But this is not the way the institutions of the EC themselves viewed the matter.
The doctrine of direct effect can be traced back to the decision of the ECJ in VanGendEnLoos1963, a case concerning whether what is now Art. 25 (abolition of customs duties) could be used as a defence against a prosecution for non-payment of duties. Although the Treaty did not state that the Art. 25 was directly effective (or even, for that matter, directly applicable), the ECJ held that it was in the spirit of the Treaty that this provision should be directly effective. Van Gend was a landmark decision of the ECJ, and it is clear from the submissions of the member states at the time that such a decision -- implying that a treaty should have immediate legal consequence in the courts of the member state -- would not be welcome. The Van Gend decision is an early example of the ECJ's `teleological' interpretative proess, in which it seeks to give effect to the purpose of the Treaty rather than the specific words of the Treaty. It is a very different style of law-making than that practiced by UK courts, at least openly.
In order to reduce the burden on national courts, and perhaps to mollify the reluctant member states, the ECJ imposed certain restrictions on the Treaty articles that would be able to have direct effect. In particular, a provision would only be directly effective if:
In practice, these restrictions have not always been as limiting as they might initially appear. For example, although a measure that imposes a positive obligation on a member state might not be justiciable, the ECJ has been willing to separate Treaty provisions into separate `negative' and `positive' elements, and allow member state courts to enforce the negative elements. For example, in CostaVENEL1964, the ECJ had to consider the direct effect of (what is now) Art. 31 of the Treaty. Para (1) imposes a positive obligation, to ``adjust state monopolies'', which was held, in itself, not to be directly effective. However, para (2) says: ``Member states shall refrain from introducing any new measure which is contrary to the principles laid down in paragraph 1''. This meant that the applicant could seek a declaration that a monopoly had been created unlawfully.
Moreover, even though the Van Gend decision stipulated that a provision had to be `clear' and `negative', in DefrenneII the ECJ gave direct effect to Art. 141, a provision which neither the member states nor even the Commission believed to be sufficiently precise to be directly effective. In that case, the ECJ held that the provision was not clear enough to allow complex questions of fact to be determined but, given that the factual situation was clear, the principle -- of equal pay for men and women -- was clear enough to be justiciable.
In short, it seems that Treaty provisions may have direct effect if they are unconditional, intended to create rights and duties, leave no futher implementation to the member states, and direct effect is necessary to give effect to important principles underlying the Community. There is, however, little specific support for this stance in the Treaties -- it is a reflection of the ECJ's purposive approach to interpretation.
The early cases on direct effect concerned Treaty articles. However, once it was established that Treaty articles could have direct effect, there seemed little doubt that the ECJ would apply direct effect to Regulations, where it made sense to do so. After all, Art. 249 states that a Regulation shall be ``binding in its entirety and directly applicable''. And this proved, indeed, to be the case.
Two problems arise in extending the principle of direct effect to directives: first, whether a directive is sufficiently precise that they can be directly effective at all; second, whether -- assuming that directives can, in general, have direct effect, they can have horizontal direct effect. To have horizontal direct effect is to be capable of being invoked in an action in the national courts between two private individuals. Most of the landmark cases on direct effect until Van Duyn (see below) in 1974 had concerned actions against the state. The question whether the same rules would apply to actions between private individuals does not seem to have been considered in detail. It now seems settled law that Treaty articles and Regulations can have horizontal direct effect, and the issue remains controversial only for Directives.
By definition, a Directive is something that must be implemented by the member states using their own legislative machinery. If that is the case, can a Directive -- which of itself is not a piece of legislation -- have direct effect? The Van Gend formula can be applied, in at least some case, to Directives, as can the purpose expressed in that case for giving direct effect to provisions of EC law. On the other hand, the very nature of Directives often means that the obligations they create will be less precise than those in Treaty articles and Regulations.
In another controversial case -- VanDuyn1974 -- the ECJ suggested that a Directive -- in this case prohibiting restrictions on free movement -- was directly effective, against the state at least, provided its obligations were clear enough for it to be justiciable. This decision, like Van Gend in its day, was not welcomed by the member states. They reasoned that if the Treaties drew a distinction between non-discretionary Regulations, and discretionary Directives, there must be some true discretion. Making Directives justiciable at the suit of the individual would override any discretion the state might possess. However, a few years later the ECJ, in deciding the case of Ratti1978, was unequivocal about the direct effect of directives. It stated that it would be contrary to EC law for a member state, when asked by a litigant who had complied with EC law to rule in his favour, to prefer its own rules to those of the Community.
However, in the days of Van Duyn the ECJ could not impose financial penalties on states for failing to implement Directives. Consequently, the only way for the ECJ to compel recalcitrant states to meet their obligations was to allow actions brought against them by private individuals to succeced where otherwise they would have failed (As an aside, it should be noted that the ECJ has extended direct effect to provide for an individual remedy in cases where the Community has failed in its obligations, as well as the member states: e.g., ReynersVBelgium1974).
This punitive nature of direct effect was emphasised by the ECJ in Ratti_, which also imposed a limitation on the scope of direct effect of directives. The reasoning the Court applied in Ratti used was derived from the `estoppel' principle -- if member states were free to resile from their obligations to implement Directives, many important Directives would remain unimplemented. Member states would therefore be estopped from using the strict discretionary nature of Directives as a defence against an allegation of non-implementation.
From this reasoning it followed that a Directive would only have direct effect, if it had any at all, once the deadline for its implementation had passed.
If the purpose of granting direct effect to all the various kinds of legislation is to ensure that the Community's project is not hindered by non-compliance of the member states (and the institutions of the EU), and if Regulations and Treaty articles can have horizontal direct effect, Directives should also have horizontal direct effect. There seems little basis in the Treaties for distinguishing Directives in this regard. Nevertheless, the ECJ has consistently refused to allow individuals to rely on Directives in actions against other individuals. The paradigm case is MarshallI1986, in which the Court refused to extend the Van Duyn principle to provide for horizontal direct effect, although it did accept that vertical direct effect could exist against state-run bodies, as well as the central organs of state. There is no logical reason why a Directive should be effective against the state, but not against an individual -- the various objectives to horizontal direct effect put forward by the ECJ apply also to many Regulations. For example, it was suggested in Marshall I that Directives are not made publically available, and it would be unjust to expect private individuals to organize their affairs according to laws of which they are unaware. However, Art. 254 now imposes an obligation to publish Directives. Moreover, arguments that Directives leave too much implementation to the discretion of member states to be effective against private individuals also mitigate against direct effect against the state, and it now seems settled that Directives can have this vertical direct effect.
Most likely the ECJ's willingness to allow for vertical, but not horizontal, direct effect of Directives is a compromise. While logic would suggest that Directives ought to be horizontally directly effective if Regulations are, member states have -- reluctantly -- conceded direct vertical effect, but would be very unwilling to accept direct horizontal effect. All the same, there are indications that the ECJ would wish to extend horizontal direct effect to directives, as it has achieved similar effects by other methods.
The first of these methods has been to take a broad view of what constitutes the state. In Foster v British Gas (C-188/89), the ECJ held that a gas supplier was an organization against which a Directive could have direct effect -- at that time British Gas held a monopoly on gas supply and was under effective control of the state. Nevertheless, it was a private company, and the `estoppel' argument used in Ratti would not apply.
The second method of achieving horizontal direct effect has been to impose an obligation to comply with EC Directives on the national courts, to which vertical direct effect would apply with less controversy. Important cases in this area are Marleasing1989 and Von Colson. In these cases, the ECJ held that, as the courts of member states are under an obligation to decide cases in accordance with EC law, and a Directive is a source of EC law, then a decision in a case between two private individuals which was in contravention of a Directive could itself be challenged by the individual against the state.
If it is the case that a national court must take into account provisions of Directives when deciding cases between individuals (and this was suggested in Ratti, ten years before Marleasing_), then Directives essentially have horizontal direct effect, in all but name, and there is less uncertainty on this point than many commentators suggest. However, if the Marleasing principle is taken to its logical conclusion, it ought to allow for Directives to be cited by the state, against_ the individual. This is a path that even the ECJ seems reluctant to tread, and a number of more recent cases have sought to justify the exclusion of the Marleasing principle when its use would create impose additional criminal liability on an individual.
The third situation in which the effects of horizontal direct effect have been observed without its express application is where the horizontal direct effect is incidental. The most striking case is CIASecurity1996 in which, by holding that a provision of national law that conflicted with an unimplemented directive could not be applied, the ECJ incidentally changed the balance of authority in a case between to private individuals. This case, and others like it, demonstrate the incoherence of the ECJ's expressed position on horizontal direct effect.
In the Faccini Dori case, there were strong suggestions from the Advocates General that the ECJ should abandon its artifical distinction between horizontal direct effect and vertical direct effect. However, the Court steadfastly refused to do this. Nevertheless, cases like CIA Securities show how close the ECJ is to recognizing horizontal direct effect, at least covertly, and it has to be wondered whether the distinction serves any logical purpose. To some people, the view of the ECJ on direct effect was perfectly clear in the Ratti decision twenty years ago.
Finally, it should be noted that since the Francovitch1991 decision, it has been possible for an individual to obtain a remedy in damages against the state for non-implementation of directives. If, as is generally assumed, the purpose of allowing direct effect to Directives is a punitive one, serving to encourage compliance of the member states with their obligations under EC law, then Francovitch renders the doctrine obsolete. However, it's unlikely that we'll see the Van Duyn principle overturned any time soon.
Law glossary index
©1994-2006 Kevin Boone, all rights reserved