Is non-Community Union law, i.e.second and third pillar law, supreme over national law in the same sense as mainstream Community law? This highly important question has not been analysed by the Court of Justice as yet, and it may take a while until it is referred, given the restrictions on preliminary references under those pillars. Some remarks can be made.
First, from an international perspective, there is no doubt that non- Community Union law is as supreme over national law as mainstream Community law, and before the Court of Justice, an international court, a Member State would not be allowed to invoke national law to escape its obligations under the second and third pillar. But more important is the question whether second and third pillar law should also be supreme over national law before the national courts.
Now, does non-Community Union law deriving from the second and third pillar take precedence over conflicting national law? In the context of the third pillar, the Treaty itself excludes direct effect of framework deci- sions and decisions. This issue has been discussed before. However, the Treaty is silent on the supremacy or primacy of these same acts, and of any of the other acts adopted under Title VI, or indeed on the supremacy of the relevant Treaty provisions themselves. The exclusion of direct effect does not of itself entail the absence of supremacy. As discussed below, also in mainstream Community law, non-directly effective provisions as such are supreme over conflicting measures of national law. The difference is that the courts cannot draw the same consequences from this primacy, as they cannot ‘apply’ the non-directly effective provisions. They can, on the other hand, and are under an obligation to, interpret conflicting measures
of national law in conformity with Community law, including non- directly effective provisions; and they have jurisdiction to hold the State or other governmental bodies liable in damages for harm caused by its infringements of Community law, including (some) directly effec- tive provisions. Direct effect and supremacy are accordingly separate and independent issues.
One way to answer the question is to go back to Costa v ENELand to check whether the reasons adduced by the Court of Justice to proclaim the precedence of Community law as a general principle apply with the same force to second and third pillar law. In my opinion, they do not, at least not as forcefully. In Costa v ENEL, the Court of Justice derived the princi- ple of precedence from ‘the special and original nature of the law stemming from the Treaty, an independent source of law’. ‘By creating a Community of unlimited duration, having its own institutions, its own personality, its own legal capacity and capacity of representation on the international plane, and more importantly, real powers stemming from a limitation of sovereignty or a transfer of powers from the States to the Community, the Member States have limited their sovereign rights, albeit within limited fields, and have thus created a body of law which binds both their nationals and themselves’. Furthermore, the terms and general spirit of the Treaty made it impossible as a corollary to accord precedence to unilateral and subsequent measures. And the Court completed its argumentation with references to provisions of the Treaty:
Article 5(2) of the EEC Treaty (now Article 10 EC), Article 7 of the EEC Treaty (prohibition of discrimination); the fact that several provisions pro- vided for specific procedures if Member States wanted to derogate from the Treaty; and the fact that regulations are ‘directly applicable’ under Article 189 of the EC Treaty (now Article 249 EC).
In the context of non-Community Union law, it seems that these crite- ria apply to a much more limited extent.130The terms and spirit of the sec- ond and third pillar would rather argue against the acceptance of the principle of supremacy. It was precisely to escape the intervention by the Court of Justice, and in order not to open up co-operation in the areas of common foreign and security policy and justice and home affairs to the same characteristics of mainstream Community law, that they were put in separate ‘pillars’. Likewise, some of the other criteria cannot support a claim of primacy with the same force, at least not formally speaking:
the ‘Union’ does not have its own institutions,131it does not have legal
130See also Chr Timmermans, ‘The Constitutionalisation of the European Union’, (2002) YEL, 1, at 9.
131But see the unitary view defended for instance by B De Witte, ‘The Pillar Structure and the Nature of the European Union: Greek Temple or French Gothis Cathedral?’, in T Heukels et al (eds), The European Union After Amsterdam (London, Kluwer, 1998) 51;
D Curtin and I Dekker, ‘The EU as a ‘Layered’ International Organization: Institutional Unity in Disguise’, in P Craig and G de Búrca (eds), The Evolution of EU Law (Oxford, OUP, 1999) 83. This is developed further below, in Part 2.
personality,132it does not have capacity of representation; more impor- tantly, there is no sense of limitation of sovereignty or transfer of powers:
in common parlance, the second and third pillars are intended not to be supra-national, but instead were kept separate because the Member States preferred to confine these areas to intergovernmental co-operation. In addition, Article 5 of the EC Treaty (now Article 10 EC), which is nowa- days considered the main Treaty basis for the principle of precedence, has no equivalent under the Union Treaty.133And finally, the text argument of Article 189 of the EEC Treaty (now Article 249 EC), equally leads to the opposite result, given the express exclusion of direct effect of framework decisions and decisions. The argumentation of the Court of Justice in Costa v ENELaccordingly does not offer the same support in favour of a principle of supremacy of non-Community Union law.
On the other hand, it can be argued that the quality and characteristics of the Community legal order have some radiation effect (Reflexwirkung) on the Union’s legal system.134The case law of the Court of Justice gives an example of such radiation effect in the context of the Brussels I Convention on the recognition and enforcement of judgments in civil and commercial matters.135It is argued that the characteristics of the Community legal order spread out and affect the second and third pillar law. The unity thesis gives additional force to the argument: in fact, it is argued, the Union is not sep- arate from the Communities: both organisations use the same institutional structure, they are based on common principles and aim to achieve com- mon objectives; they are in fact the same actors, acting in different capaci- ties and with varying competences and under varying procedures.
However, these elements cannot do away with the fact that the second and third pillars are just that: separate pillars,136which have not been brought under the Community system, precisely because the High Contracting Parties did not want the law deriving from these pillars to have the same characteristics, and to be governed by the same principles.
132That is to say, it has not expressly been awarded such personality in the corpusof the Treaty.
One can argue, however, that the Union does have de factolegal personality, on the basis of the principles as laid down by the International Court of Justice in the Reparation for Injuries Case, seeInternational Court of Justice, Advisory Opinion of 11 April 1949, Reparation for Injuries Suffered in the Service of the United Nations, available on www.icj-cij.org.
133But see for an argument in favour of the development of a similar principle of loyalty in Union law, D Curtin and I Dekker, ‘The Constitutional Structure of the European Union:
Some Reflections on Vertical Unity-in-Diversity’, in P Beaumont, C Lyons and N Walker (eds), Convergence and Divergence in European Public Law (Oxford, Hart Publishing, 2002) 59.
134See Chr Timmermans, ‘The Constitutionalisation of the European Union’, (2002) YEL, 1, at 10.
135See Case 288/82Duijnstee [1983] ECR 3663. The Convention has now been transformed into a regulation.
136Despite the fact that, I agree, the image of the Greek temple with three pillars overstates the differences rather than the commonality between the various forms of cooperation.
In this context, however, these differences outweigh the commonality.
Let us now change the perspective, and try and argue the case starting not from the Community orthodoxy, but from a wider angle. What was unique and novel in Costa v ENELwas not the fact that the Court of Justice, an international court, accorded priority to a Treaty over conflicting national law: Pacta sunt servanda, and it is only natural for an international court to emphasize that. It may not even have been so spe- cial that the Court of Justice stated that Community law was also to have precedence before a national court: if asked, any international court would come to the same decision, because the State would (probably) infringe international obligations if the courts did otherwise. But what made the difference for Community law was that there was a court which could hold, for all the Member State courts alike, and in the course of a procedure before a national court, that Community law takes precedence, in the sense that Member States cannot deviate unilaterally from what they have agreed in common. Under classic international law, an international court will only have to decide whether the State as such (and including all its organs) has violated an obligation under international law, ex post facto. So while it has been maintained for a long time that international law does not oblige national courts to apply international law and award prece- dence to international obligations,137it is also clear that if a court does indeed deny precedence to these obligations, it most likely contributes to the State’s violation of the Treaty and thus causes the international liabil- ity of the State to arise. The preliminary rulings procedure, however, makes it possible for ‘the clock to be stopped’:138 in the context of Community law, it is not necessary to wait until the end and ask the ques- tion whether indeed the national court has contributed to causing the State’s international liability to arise: the Court of Justice can interfere at an earlier stage, and prevent the national courts from contributing to the violation of the Treaty.
The same may happen in the context of Title VI, where the Court has limited jurisdiction to give preliminary rulings. However, there are important differences with the situation in mainstream Community law:
the Court only has jurisdiction to answer preliminary rulings in so far as a Member State has accepted this jurisdiction. While most Member States have done so, it is by no means obligatory, and in addition, the Member States could chose between various options, as to whether lower and/or highest courts could or must make references. What would be the effect
137But see arguments to the contrary, for instance the Danzigcase referred to above, and recently the La Grand case, discussed below. See for a discussion of more modern approaches in international law concerning the principle of direct effect, A Nollkaemper,
‘The Direct Effect of Public International Law’, in JM Prinssen and A Schrauwen (eds), Direct Effect. Rethinking a Classic of EC Legal Doctrine (Groningen, Europa Law Publishing, 2002) 157.
138See D Wyatt, ‘New Legal Order, or Old?’, (1982) ELR147.
of a decision of the Court of Justice, awarding precedence to a particular provision of an act adopted under Title VI, if some national courts can- not make references on the issue? Are they to the same effect bound by that decision? Underlying the principle of supremacy of Community law is pacta sunt servanda, and the notion of uniformity of Community law:The executive force of Community law cannot vary from one State or another in deference to subsequent domestic laws, without jeopardising the attainment of the objectives set out in Article 5(2) [now Article 10 EC] and giv- ing rise to the discrimination prohibited by Article 7’.139Hence, it should be binding on all courts.
Nevertheless, and irrespective of this procedural problem, it would seem anomalous for the Court of Justice not to accept the supremacy of non-Community Union law – to the extent that it has jurisdiction. With the same force as for Community law, it must be accepted that the Member States cannot unilaterally detract from legal rules accepted on the basis of reciprocity.140This is a simple application of the principle of pacta sunt servanda. Once the clock is stopped, an international court will naturally state that treaty obligations take precedence, otherwise the international liability of the State will arise. The difficulty is that the clock is not stopped in the same way in the various Member States, and that the decision of the Court may not have the same (uniform) effect for each and every national court.
More problematic is the case for an absolute and unconditional version of supremacy. Consider the objections raised by the national courts against the principle of supremacy of mainstream Community law. Some of these objections concerned the place of the courts in the constitutional structure, and these have been overcome in one way or another. But oth- ers were more principled, and were most powerful in the context of con- flicting provisions of national constitutional law and Community law. In the case of Internationale Handelsgesellschaft, the Bundesverfassungsgericht argued against the primacy of Community law, that the Community legal order lacked a sufficient protection of fundamental rights; accordingly, this protection had to be offered at the national level. In the context of Community law, the Court of Justice has been able to counter this argu- ment by the development of the theory of general principles of Community law which include fundamental rights: the protection offered at the national level was replaced by protection at the Community level and accordingly there was no need for the national courts to retain juris- diction to review Community law. It is well-known that the Court of Justice has been able to convince the Bundesverfassungsgerichtto a large extent. Now, the argument was powerful in the context of Community
139Case 6/64Costa v ENEL[1964] ECR 585, at 594.
140See, once again, Case 6/64Costa v ENEL[1964] ECR 585, at 594.
law, where the Court of Justice can indeed state that it is able to replace the national courts in the protection of fundamental rights. However, that is not the case in the context of Title VI where the Court of Justice has only very limited jurisdiction, and cases may not reach the Court, because the more limited version of the preliminary rulings procedure, and of actions for annulment, which are in any case precluded for private applicants. In addition, there are other deficiencies in the system of Title VI, which would seem to add force to objections of national courts against the supremacy of acts adopted under this Title. The very limited democratic legitimation of acts adopted under Title VI is probably one of the most important.141
Where does all this leave us? The question of supremacy can arise before the Court of Justice, and the natural tendency of the Court of Justice will go in favour of the acceptance of supremacy also in the area of Title VI. However, the context is so different from that of the first pillar, that there are good reasons to argue against applying the same absolute and unconditional version of supremacy.
4.3. DIRECT EFFECT AND SUPREMACY: THE‘SIMMENTHAL MANDATE’ Direct effect and supremacy constitute the groundwork of the Community mandate of the national courts. The essence of the Community mandate is contained in those two doctrines, the culmination of which for the mandate of the national courts is Simmenthal, where the Court held that every national court must in a case within its jurisdiction apply Community law in its entirety and protect rights which individuals derive from it, and must set aside any provision which may conflict with it, whether prior or subsequent to the Community rule.
4.3.1. ‘Setting Aside or Disapply’
The first questions relating to the exact duties and obligations of the national courts acting as Community courts were put before the Court soon after Van Gend en Loos and Costa-ENEL. In Lück, the Finanzgericht Düsseldorf sought a clarification of the consequences of the principle of primacy with regard to provisions of national law incompatible with
141It may be be objected that when Costa v ENEL and Simmenthal were decided, the Community was not more democratic than is the case now in the second and third pillar.
However, one may and must accept a higher level of democratization now, with pro- gressing integration and maturing of the system. In addition, the decisions adopted in the third pillar probably touch upon individuals’ lives more directly and more intrusively than was the case with economic decisions adopted in the early days.
Community law: what is their fate? Would they be ‘void’, ‘non-existent’,
‘to be annulled’? The Court held that Article 95 of the then EEC Treaty merely had the effect of ‘excluding the application of any national measure incompatible with it’. Disapplying the conflicting measure is the most general remedy that individuals may claim from national courts.142The supremacy of Community law requires, in the case at hand, that the con- flicting rule find no application,143is set aside, ‘disapplied’, or declared
‘unenforceable’.144 Community law does not automatically render the conflicting rule null and void. The national norm remains in existence and can be applied to cases in which they do not lead to an infringement of Community law. Community law only dictates the non-application of the conflicting measure, whether prior or subsequent, in cases where Community law would otherwise be infringed.145 The precedence of Community law imposes an obligation de résultatrather than an obligation de moyenson the national courts.146