Jump to ContentJump to Main Navigation
Constitutional Courts and Democratic ValuesA European Perspective$

Victor Ferreres Comella

Print publication date: 2009

Print ISBN-13: 9780300148671

Published to Yale Scholarship Online: October 2013

DOI: 10.12987/yale/9780300148671.001.0001

Show Summary Details
Page of

PRINTED FROM YALE SCHOLARSHIP ONLINE (www.yale.universitypressscholarship.com). (c) Copyright Yale University Press, 2021. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in YSO for personal use. Subscriber: null; date: 16 April 2021

The Impact of the European Court of Justice

The Impact of the European Court of Justice

(p.122) Chapter Eleven The Impact of the European Court of Justice
Constitutional Courts and Democratic Values

Víctor Ferreres Comella

Yale University Press

Abstract and Keywords

This chapter describes the two powerful courts that constitutional tribunals now encounter: the European Court of Justice (ECJ) in Luxembourg and the European Court of Human Rights (ECHR) in Strasbourg. The jurisprudence of these two courts empowers ordinary judges in ways that undermine, or could potentially undermine, the centrality of constitutional tribunals when it comes to reviewing national legislation. This chapter focuses on E.U. law and the ECJ, and the next chapter is devoted to the ECHR. The chapter cannot, and does not, address all the constitutional issues that the existence of these two European courts raise. It simply aims to offer some thoughts on how well or how poorly the Kelsenian model fits with the judicial arrangements that have been established in the supranational sphere and what should be done to guarantee a better fit.

Keywords:   powerful courts, constitutional tribunals, European Court of Justice, ECJ, European Court of Human Rights, ECHR

IF CONSTITUTIONAL COURTS worked in a domestic legal vacuum, the only source of potential erosion to worry about would be the interpretive powers that ordinary judges can exercise if they are not contained in the right way, as we explored in the previous chapter. But the international forces that have given birth to the European Union and to the Council of Europe have complicated the picture. These organizations may affect in serious ways the centralized Kelsenian system of judicial review that exists at the domestic level. (In this discussion, I often refer to the centralized model as “Kelsenian” rather than “European” to avoid confusion with the judicial system that has been articulated at the European supranational level.)

Constitutional tribunals now encounter two powerful courts, the European Court of Justice (ECJ) in Luxembourg (in the context of the European Union) and the European Court of Human Rights (ECHR) in Strasbourg (in the context of the Council of Europe). The jurisprudence of these two courts empowers ordinary judges in ways that undermine, or could potentially undermine, the centrality of constitutional tribunals when it comes to reviewing national legislation. In this chapter I focus on E.U. law and the ECJ, and the next chapter is devoted to the ECHR. The aim of these two chapters is rather modest. I cannot address all the constitutional issues that the existence of these two European courts (p.123) has raised. My aim is simply to offer some thoughts on how well or how poorly the Kelsenian model fits with the judicial arrangements that have been established in the supranational sphere and what should be done to guarantee a better fit.

The Legal Architecture of the European Union

To frame our discussion, I need to say a few words about the legal architecture of the European Union. This organization, which currently includes twenty-seven members, is based on three so-called pillars, the most important of which is the first, which includes the European Community and the European Atomic Energy Community (Euratom). The second pillar, which deals with common foreign and security policy, and the third, which is devoted to police and judicial cooperation in criminal matters, are of a more intergovernmental character. For purposes of our discussion here, we can focus on the problems that arise in the context of the European Community.1

The Treaty of Rome, the foundational treaty of 1957 that created the European Community (which was originally called the European Economic Community), established a complex distribution of legislative powers between the Community institutions and the national authorities. The European Community can regulate only certain matters that are enumerated in the treaty; the rest remain in the hands of the member states. Several institutions were created to exercise the competences assigned to the Community. When it comes to lawmaking, the European Commission, the Council of Ministers, and the European Parliament are the most relevant entities. Depending on the subject matter, the Community can issue regulations, which are equivalent to ordinary legislation, or directives, which are framework statutes that bind member states as to certain results but leave to states the choice of the methods of achieving those results. The law of the European Community (E.C. law) thus comprises both the foundational treaties (primary law) and the legal provisions enacted by the Community institutions (secondary law).

The interpretation and application of E.C. law in ordinary legal disputes is basically left to the courts of the different member states. These courts work under the guidance of a central court that sits in Luxembourg: the ECJ.2 This court is composed of twenty-seven judges (one per member state) and eight advocates general. The latter are responsible for (p.124) presenting, with complete impartiality and independence, an “opinion” in the cases assigned to them. Depending on the type of case to be decided, the court may sit as a full court (though it has long ceased to do so), in a grand chamber of thirteen judges, or in chambers of three or five judges.

The ECJ's jurisdiction covers different types of cases. The court, for example, can hear complaints that a particular state has failed to fulfill its obligations. It can also review the validity of decisions and laws enacted by the Community institutions or the legality of an institution's failure to act. Actions for damages against the Community for noncontractual liability are also part of its jurisdiction. Its most important function, however, is to guarantee that E.C. law is interpreted and applied in a uniform manner in all member states. For these purposes, the court has been conferred the authority to answer questions sent by national judges concerning E.C. law. This procedure is called “preliminary reference.”3 Whenever a national court deciding a case has doubts about the proper interpretation of E.C. acts, it can raise a preliminary reference to the ECJ for clarification. Similarly, if the national court has doubts about the validity of any of those acts, it can certify a question to the ECJ.4

This mechanism is widely regarded to be of key importance to the proper operation of the system—it is “the jewel in the Crown.”5 It is so central, actually, that the E.C. treaty provides that when a national court is deciding a case as the court of last resort, it is no longer free to ask the ECJ but is actually required to do so whenever a question concerning the interpretation or validity of E.C. law arises.

Through this procedure, the ECJ has had the opportunity to announce basic principles. It took an important first step in 1964, for example, when it proclaimed that E.C. law (both primary and secondary law) must prevail over national legislation in case of conflict. This is the so-called principle of primacy. Even a statute passed by the democratic national parliament, therefore, has to yield to the higher authority of E.C. law.6

From a political point of view, this step was relatively easy for the ECJ to take. As Joseph Weiler has explained, most E.C. laws at that time were enacted through a procedure that required the unanimous consent of the national governments sitting on the council. The primacy of E.C. law was not hard for the governments of the member states to swallow if they retained a veto over the content of that law. When unanimity was (p.125) later abandoned in favor of supermajority rules in an increasing number of areas, however, the principle of primacy started to show its bite: a member state that voted against a piece of E.C. law nevertheless had to accept its higher status over national legislation.7

For a long while, however, the ECJ did not say how the primacy of E.C. law had to be enforced at the domestic level. But everything changed in 1978, when it handed down its historic judgment in the Simmenthal case.

The Revolution that Simmenthal Brought About

Before 1978, different solutions seemed to be available to the member states for reacting against national laws that allegedly offended E.C. law. The Italian Constitutional Court crafted a doctrine that was very faithful to the Kelsenian spirit animating the centralized model. It held that a national statute that contradicts E.C. law is not only directly contrary to E.C. law but also indirectly contrary to article 11 of the Italian Constitution, which links the Italian republic to international organizations. This holding reinforced the primacy of E.C. law in Italy. The most important consequence, however, was that it preserved the monopoly of the Italian Constitutional Court within the Italian system. If an ordinary judge concluded that a national statute violated E.C. law, the judge had to raise the question to the Italian Constitutional Court. The judge could not set the statute aside on his or her own authority.

This was a very coherent move given the assumptions upon which the centralized model rests. If, as we have seen in earlier chapters, there are good reasons for Italy and many other European countries to have created a centralized system of judicial review of legislation when the higher norm to be safeguarded is the national constitution, those reasons would also seem to be sufficiently powerful to justify a similar centralization when the set of higher norms to be guaranteed is E.C. law.

The ECJ, however, rejected this Kelsenian solution developed by the Italians. In the Simmenthal case,8 the ECJ laid the foundations for a decentralized system of judicial review. It asserted that a national court that is called upon to apply provisions of Community law “is under a duty to give full effect to those provisions, if necessary refusing of its own motion to apply any conflicting provision of national legislation.” Therefore, “it is not necessary for the court to request or await the prior setting aside of such (p.126) provision by legislative or other constitutional means.”9 That is, all national judges in charge of specific disputes are empowered, on their own authority, to disregard any domestic statute that contradicts E.C. law. They need not stay the proceedings and wait for the formal invalidation of that statute by the national constitutional court.

Obviously, this deeply transforms the role of ordinary judges in Europe: they are now authorized to review, by themselves, the validity of parliamentary enactments under higher norms. Constitutional courts have lost the monopoly they used to have. They still retain their monopoly over the determination of the validity of statutes under the national constitution, but they have lost their general monopoly over statutes. Ordinary courts too can now check them under E.C. law. The holding in Simmenthal seems a crime of high treason against the Kelsenian order of things! Ironically, here we have a European supranational court that has “Americanized” the practice of judicial review of legislation in the diverse member states of the European Union.

This revolution was not quickly accepted in all jurisdictions. It has taken some time for the doctrine to take domestic root.10 The different countries in Europe, moreover, have translated the Simmenthal doctrine into national constitutional terms in different ways.11 The doctrine is now quite settled, however, and we should inquire into its justification and explore the extent to which it fits with, or departs from, the rationales that support the Kelsenian system.

Why Simmenthal?

There are basically two substantive arguments that the ECJ relied on to justify its holding in Simmenthal.12 The first appeals to efficiency. The ECJ reasoned that the effective enforcement of E.C. law would be impaired if the ordinary courts in charge of handling specific disputes were not authorized to immediately set aside national statutes that contravene E.C. law. If courts had to stay the proceedings and ask the national constitutional court to intervene, there would be a delay in the resolution of cases. This delay would amount to an impediment to the full effectiveness of E.C. law, even if “only temporary” in nature. The application of E.C. law, the court held, should be “direct and immediate.”13

This may sound like a convincing argument, but the countries that have established constitutional courts should not be impressed by it. The (p.127) centralized model of judicial review is based on the assumption that this delay is a price worth paying—provided, of course, that the delay is not unreasonably long. As was argued in the previous chapters, the issue raised by the statute can be definitively settled by the constitutional court from the very beginning; there is no need to wait for the case to get decided by the highest courts after various appeals. A single decision by the constitutional court striking down the statute, moreover, has the power to bind all judges immediately. These advantages of centralization seem perfectly applicable when domestic legislation is to be reviewed for its conformity with E.C. law.14

There is a second and more powerful argument to justify the holding in Simmenthal, however. The major task of the ECJ, as we know, is to answer national judges' questions regarding the proper interpretation of E.C. law. What would be the point of requiring those judges to raise a question to their national constitutional court whenever they deemed a particular statute to be inconsistent with E.C. law? If the case law of the ECJ is not sufficiently clear for judges to ascertain whether the applicable statute is valid, it is the ECJ that must clarify things—not the national constitutional court.15 If the latter were entitled to intervene, moreover, it would probably decide the issues concerning E.C. law without first consulting the ECJ. (As a matter of fact, constitutional courts have been very reluctant to send preliminary references to Luxembourg. Only recently have some of them done so, in a few cases.)16

So the Simmenthal decision was strategically wise. If national constitutional courts had been permitted to centralize (for domestic purposes) the function of reviewing national legislation under E.C. law, they would have gone their own ways. They would have done the checking themselves, without first asking the ECJ. This would have been disruptive of the E.C. legal order. The ECJ is the supreme interpreter, but it needs to be petitioned by national judges in order to speak its supreme judgment. (Folklore has it that the judges in Luxembourg opened several bottles of champagne when they received their first case in 1954.) Since the ECJ had reason to believe that ordinary judges would be more eager than constitutional courts to use the preliminary-reference mechanism, a centralized system, like the one that Italy had set up before the Simmenthal decision, had to be rejected.

Now, what are the costs of the decentralized arrangement that the (p.128) ECJ imposed? How are the values that inform the Kelsenian model affected?

Legal Certainty Under Simmenthal

Let us start with legal certainty. The first thing to note is that, in general, E.C. law is quite different from national constitutions. A constitution is a very special text that often expresses broad and morally charged principles, the interpretation of which is deeply controversial. The conflict between a statute and the constitution is often the conflict between a detailed legal provision and a rather abstract and fundamental principle. In contrast, E.C. law is basically made up of “ordinary legislation” that is expressed in rather specific terms. The conflict between a national statute and E.C. law is often a clash between two pieces of ordinary legislation. To the extent that this is so, controversies over whether national statutes comply with E.C. law will not be as deep and extended as constitutional controversies. A decentralized system in the field of E.C. law is thus more tolerable.

This contrast cannot be taken too far, however. Some parts of E.C. law do resemble constitutional norms. The market freedoms that are protected by E.C. law, for example, can be restricted by member states in the name of a compelling public interest. Judges must apply the principle of proportionality and decide whether the restriction is ultimately justified.17 Fundamental rights are also protected under E.C. law—both as foundational “general principles” and as part of secondary legislation. Member states are bound by such rights when they act in an area governed by E.C. law.18 Domestic judges, of course, are likely to disagree among themselves when they have to check national legislation for consistency with E.C. law in such cases.

Fortunately, however, there is an important mechanism that ensures uniformity: the preliminary-reference procedure. As has already been noted, national judges can directly consult the ECJ whenever they need guidance. What's more, the national courts whose decisions are not subject to appeal are required, as we know, to send a preliminary question to the ECJ if an interpretive problem arises under E.C. law.19 The ECJ's intervention is of great import, since its rulings are widely believed to be legally binding. The answers that it provides in the form of preliminary rulings are to be followed not only by the judges who raised the pertinent questions (p.129) but also by all other judges in all member states. The ECJ has clearly claimed this authority, and the majority of scholars have accepted it.20

Parenthetically, it is interesting to observe that although there is reluctance in civil-law countries to officially accept the binding character of the doctrines established by the respective domestic supreme courts, there is no such reservation when it comes to the ECJ. Why this difference? In part, the reason is—ironically enough—that the ECJ has no jurisdiction to quash the decisions of national courts that depart from its rulings. Precisely because of this feature of the current judicial architecture, it has become imperative for the legal community to internalize the idea that the ECJ's precedents must be followed and to strongly criticize those judges who fail to do so. As was explained in chapter 3, countries within the civil-law tradition have not felt a strong need to emphasize the role of precedents set by the highest courts, in so far as the latter can impose their views by regularly overturning the decisions of lower courts. This is in contrast to the situation in the common law, we noted, where the doctrine of precedent is a cultural stabilizer that compensates for the fact that it is hard to bring cases to the supreme court. In the context of the European Community, the situation is really extreme: the supreme interpreter of E.C. law—the ECJ—has no power whatsoever to overrule the decisions rendered by national courts. The need to inculcate the culture of precedent in lawyers and judges is therefore absolutely vital.

Actually, the role of precedents in making the system operational is of increasing importance. The ECJ has introduced an exception to the duty of national courts of last resort to raise a preliminary question: they need not ask for a preliminary ruling when there is no reasonable doubt about the meaning of the relevant E.C. legal provision. This is the so-called acte clair doctrine (clear-act doctrine). The absence of doubt is usually linked to the existence of precedents. If the question that is being examined by the national court is identical to a question already decided by the ECJ, there is no need to consult the ECJ. Even when the question is different, there is no need to send a reference if the answer can be derived from ECJ precedents.21

Because the ECJ is more and more overloaded with preliminary questions, it is necessary to insist on the responsibility of national courts to apply the existing case law themselves. The ECJ should focus on those preliminary references that pose new problems and rely on the capacity (p.130) of national courts to read the existing precedents on their own and apply the rules that have been clearly established.22

Notice that this way of understanding the preliminary-reference procedure is similar to the system that I proposed in chapter 10 as an internal correction to the Kelsenian model of constitutional review of legislation at the domestic level. I suggested that ordinary judges should be allowed to set aside legislation on their own authority provided that they interpret the national constitution in light of the precedents furnished by the constitutional court that have sufficiently clarified the underlying issue. A similar arrangement is in place in the European Community under the clear-act doctrine.

In general, the ECJ is quite careful when it makes its rulings. True, its reasoning is too “cryptic” and “Cartesian,” as Joseph Weiler has observed critically.23 The more detailed opinions of the advocates general may compensate for this argumentative deficit to a certain extent, but it would be better if the ECJ were more explicit about the underlying reasons that support its holdings.24 For purposes of securing legal certainty, however, what matters is the extent to which the court generates a consistent body of rules. In this regard, its overall performance seems to be acceptable.25

For all these reasons, legal certainty is not at grave risk, even though all national judges are empowered under Simmenthal to check the conformity of domestic legislation with E.C. law.

An Internal Correction to the Simmenthal Doctrine

There is room for improvement, however. First, constitutional courts should be allowed to intervene to review national legislation when the ECJ delegates the resolution of important legal issues to the national judiciary. Let me explain.

In some areas of E.C. law, the ECJ thinks it advisable not to frame very detailed answers to the preliminary questions that domestic courts raise. This is especially true when fundamental rights are at stake. As Daniel Sarmiento has explained, the ECJ seems to move to a different beat when it handles legal questions of a constitutional nature, such as fundamental rights. It abandons its general position as a sort of supreme federal court and opens itself to interpretive pluralism.26 This is understandable: it makes sense for the ECJ to delegate certain matters to the (p.131) interpretive and fact-finding powers of the national judiciary. In the Bauer Verlag case,27 for example, an Austrian court had to determine whether a national statute prohibiting businesses from offering consumers free gifts linked to the sale of goods or the supply of services was consistent with E.C. law. In that particular lawsuit, a German publisher would have been prohibited, under that statute, from selling a weekly magazine in Austria that offered readers the chance to take part in games for prizes. The ECJ was consulted by the Austrian court on several points. The ECJ reasoned that the national statute restricted both freedom of movement of goods and the fundamental right to freedom of expression. It also held that to justify this restriction, the state could legitimately invoke its interest in maintaining press diversity. Diversity could indeed be in danger if small publishers were unable to offer prizes like those offered by the largest companies. However, it was for the Austrian court, the ECJ held, to decide whether the restriction was necessary in order to achieve this goal. “It is for the national court to determine whether these conditions are satisfied on the basis of the study of the Austrian press market,” the ECJ said.28

It is perfectly legitimate, in cases of this sort, for the ECJ to prefer not to give a detailed answer. The national courts may indeed be in a better position to decide certain issues. But if this is the case, I submit, constitutional courts should be allowed to intervene. Recall that the Simmenthal doctrine is based on the assumption that ordinary courts at the national level will get enough guidance from the ECJ. But if the ECJ declines to give enough guidance because it feels that the national judiciary is in a better position to determine certain points of law, then the constitutional court should be allowed to step in and unify the domestic legal answer. In the Bauer Verlag case, for instance, the Austrian Constitutional Court should have been authorized to review the statute under E.C. law. It would thus have given a unified answer to the several important questions that the ECJ had transferred to the national judiciary. To the extent that the Simmenthal doctrine seems to exclude this type of intervention by the constitutional court in a case like this, the doctrine should be reconsidered.

Second, the Simmenthal power of ordinary judges to disregard national legislation that contravenes E.C. law should not be taken to exclude the possibility of establishing additional systems of legislative review. The (p.132) advocate general, Gerhard Reischl, was explicit about this in his opinion in Simmenthal. He argued that even if ordinary courts were given the power to immediately disregard legislation on their own authority, there would be no need to eliminate the procedures of constitutional review that the Italian legal system had devised to formally cancel legislation that breaches E.C. law. Proposals along these lines have been made by various scholars, and some constitutional courts have been sensitive to them.29

Simmenthal and Democratic Values

What about the other values that the Kelsenian model seeks to serve? In previous chapters I argued that the model allows the governmental majority to offer its justification of the legislative measures it has enacted. I claimed, in addition, that the model better satisfies the need for courts in charge of review to be sufficiently linked to the democratic processes. How well are these democratic values preserved under the Simmenthal arrangement?

The Right of the Governmental Majority to Be Heard

Under Simmenthal, the ordinary court handling a particular lawsuit is empowered to set aside the statute even if, in most cases, the government is not a party to the case and will not be notified of the proceedings. Is this fair?

As a general rule, the answer is yes. The government has the opportunity to defend its point of view—not before the ordinary court, but before the ECJ. Suppose that a statute raises an issue of E.C. law and that there is no ECJ case law that sufficiently resolves the matter. The national judge then asks the ECJ for a preliminary ruling. (Recall that when the case gets to the highest court, this petition is compulsory.) The ECJ then notifies the E.C. institutions and the governments of all the member states, who can submit their views.30 Since the decision to be rendered by the ECJ will give concrete meaning to the relevant provisions of E.C. law, all the member states are interested in influencing the outcome and are consequently allowed to express their points of view.31 In this way, the governmental majority whose statute is being examined by the ECJ has the chance to argue in its defense: it can give the reasons it thinks that E.C. law is not being breached. The other states can support the government's (p.133) argument or object to it. If they have enacted similar legislation, they have an incentive to participate and insist that the measure under examination is not at odds with E.C. law.

Sometimes, however, the ECJ is not addressed by the national judge. If, for example, the ECJ has laid down precedents that sufficiently clarify the underlying issue raised by the national statute, the domestic judge will normally refuse to certify a preliminary question. In this situation, the governmental majority that supports the statute will not be heard by the ECJ. Is this unfair? Not really. The government had an opportunity to express its point of view in the past, when the ECJ established the relevant precedent (or set of precedents). At that time, the governments of all the member states were asked to express and justify their interpretive positions regarding E.C. law. That was the key moment when those governments had the chance to contribute to the process through which E.C. law is construed.

Imagine now that the following situation arises. The governmental majority is perfectly aware that a new statute it has enacted is inconsistent with E.C. law as interpreted by the precedents established by the ECJ. The majority insists, however, that the ECJ's interpretation is wrong. How could the system allow the government to make its case? Maybe a jurisdictional clause similar to the one that I suggested in chapter 10 could be included in the controversial statute. The clause would explicitly require the national court deciding a case to send a preliminary reference to the ECJ if the court concluded, given the ECJ's precedents, that the statute violated E.C. law. Of course, the ECJ could still give a quick answer to the national court through a reasoned order that makes reference to the past precedents.32 The ECJ, however, might give some weight to the disagreement that the national parliament was expressing through that clause and might be more willing to hold a full hearing in which the E.C. institutions and the governments of all member states would be allowed to express their positions. This institutional device would seem particularly acceptable in cases where the applicable jurisprudence of the ECJ is deeply controversial, as may be the case when certain fundamental rights are involved, for example.

Finally, what happens when the existing case law is insufficient to decide an issue, the national court certifies a question to the ECJ, and the ECJ's answer is very open-ended because the ECJ believes that the (p.134) national judiciary is better equipped to deal with particular aspects of the underlying legal problem? I have earlier argued that in this type of case, the constitutional court should be authorized to step in on grounds of legal certainty. The right of the governmental majority to be heard presses in the same direction: if the ECJ has given discretion to the national judiciary to decide whether the statute violates E.C. law, the constitutional court should intervene so that the government can express its arguments in support of the statute.

The Links between Courts and the Democratic Branches

Let us now turn to the necessary links between courts in charge of legislative review and the democratic branches. How does the Simmenthal arrangement affect these links? To answer this question, we need to distinguish between two types of situations, as we did before.

Suppose, first, that the ECJ has offered a rather detailed elaboration of the meaning of E.C. law. In such cases, we must focus on the ECJ and consider the appointment of ECJ judges and their terms of office, as well as the ways in which the political branches can respond to its jurisprudence.

As to the appointment and tenure of ECJ judges, there is not a systematic deficit on this score, in comparison with constitutional courts. The political branches participate in the judicial selection process: ECJ judges are appointed by common accord of the governments of the member states. And the judges do not sit for life: they hold their offices for a limited period of six years (though the term can be renewed).33 If there is a loss from a democratic point of view, it is because there is not a European public opinion comparable to the national public opinion that operates at the domestic level. The decisions to appoint the judges of the ECJ may be less visible than the decisions to appoint the members of a national constitutional court.

On the other hand, as I argued in chapter 9, democratic links should not be designed in ways that can harm judicial independence. The process of interpreting E.C. law must be sensitive to the points of view of the democratic branches, but it must also include a supranational dimension. The ECJ must aim for an interpretation of E.C. law that protects the general interests of the Community, and this is possible only if the ECJ enjoys a sufficient degree of independence from the national democratic (p.135) branches. In this regard, the fact that judges are appointed by common agreement, and not simply by each national government, is an important feature. In contrast, the fact that they can be reappointed after the six-year term has expired is detrimental to their independence. (The prohibition against filing dissenting opinions diminishes the risks, though, since the court speaks with one voice and the national governments cannot easily monitor the individual performance of the judges.)

With respect to the capacity of the political branches to respond to the ECJ, things are not too bad either. I argued in previous chapters that it is desirable for judicial decisions regarding the constitution to attract debate and be subjected to public scrutiny and political checks. The intuition behind the Kelsenian model is that, given the legal and political conditions in many European countries, this goal is facilitated by centralization, as well as by the fact that the decisions rendered by constitutional courts have general effects (when the courts check legislation). In the context of the European Community, this goal is sufficiently satisfied. When the ECJ answers a preliminary reference, it does not decide a particular case. It answers a general question, and its answer is binding on all. The question and the answer may be more or less specific, but the ECJ makes a ruling that has erga omnes effects similar to those that result from a decision by a constitutional court striking down a statute. Actually, the impact of the ruling by the ECJ is even wider: the ECJ does not hold that a particular statute passed by a particular parliament violates E.C. law; it instead holds that a particular normative content violates E.C. law, and any statute that has that content, whichever country enacts it, is automatically judged to breach E.C. law. Even statutes that have a different content but are similar for purposes of the ECJ's holding are also to be regarded as contravening E.C. law. As a result, it is quite clear that the public and the political branches must focus their attention on the ECJ.

Of course, the extent to which the ECJ is actually checked by the political branches and by public opinion does not depend on the architecture of courts exclusively. It also depends on the features of the political system. There is an open debate about the seriousness of the democratic deficit of the E.C. political process. It is not easy to build a supranational organization that is as democratic as member states are. There is probably ample room for improvements on this level. But at least the judicial architecture does not make it difficult for member states to activate the (p.136) relevant supranational democratic checks—to the extent that they exist. A well-known example of a political response to the ECJ is the positive-action provision that was inserted by the 1997 Treaty of Amsterdam into article 141 of the E.C. treaty. The political branches reacted in this way against the ECJ's controversial decision in the Kalanke case, which established a very restrictive jurisprudence on the legality of affirmative action programs for women.34

So when the ECJ establishes in a rather specific manner the interpretation to be followed by national courts, there is no strong reason to criticize the current system on democratic grounds.

In contrast, when the ECJ does not establish specific rulings but grants discretion to the national judiciary, we have good reason to criticize the current arrangement under Simmenthal. If, given Simmenthal, the constitutional court is not authorized to speak as to the validity of the national statute and ordinary courts instead must address the issue on their own authority, some of the democratic principles that support the centralized model are negatively affected. As we know, in many European countries, the selection process for ordinary judges is not as democratic as the selection process for members of the constitutional court. Moreover, the capacity of the democratic branches to respond to the judicial interpretation of E.C. law is made more difficult in a decentralized system, for different courts may disagree and as a result no clear message will be sent to the democratic branches until the supreme courts gradually establish the relevant case law. If, in contrast, the constitutional court were the entity responsible for reviewing the statute, the conversation with the political institutions would be easier.

National Constitutional Checks on the European Court of Justice

A final point on the advantages of constitutional courts in the context of the European Union is in order. A recurring theme in the debates about the construction of Europe is the issue of primacy: Is E.C. law, or are national constitutions, the supreme law of the land? May national judges disregard E.C. law if it violates a basic principle or rule embodied in the national constitution? In particular, may national judges refuse to apply a piece of E.C. legislation on the ground that it conflicts with fundamental rights as protected by the national constitution—even if (p.137) the ECJ upholds that legislation as consistent with fundamental rights at the European level?

This is an interesting and complex issue that has attracted the attention of many scholars. It is beyond my present purposes to examine this question.35 But there is a marginal point worth making here: when such extraordinary cases arise, it seems reasonable to resort to the constitutional court to channel the dialogue (and the tension) between the national judiciary and the ECJ.36 The constitutional court can concentrate all the potential “disobedience” to the latter. This system has several advantages. First, it provides clarity. It would be more chaotic if all the different ordinary courts could decide themselves whether to accept the enforceability of E.C. law depending on whether they thought that the ECJ human-rights case law was good enough. It is better for the constitutional court to be in charge of this delicate matter. The position of the national judiciary should be expressed in a clear way. The constitutional court should be the entity that talks to Luxembourg in the name of ultimate principles rooted in the national constitution. Second, if a constitutional court made public its own reservations about the acceptability of the ECJ's jurisprudence in the area of rights, it would send a strong warning to Luxembourg. As Dieter Grimm explains in the context of the German Constitutional Court, a preliminary reference raised by the German court to the ECJ could be used to signal the existence of a potential conflict that needs to be worked out. The warning would be stronger if it came from the constitutional court than if it emerged from lower ordinary courts.37 Finally, the constitutional court is more likely to be prudent in these potential fights. Precisely because its decisions have a greater impact than those of ordinary judges deciding concrete cases, the constitutional court is encouraged to think twice before it holds that E.C. law conflicts with fundamental rights protected by the national constitution.


In this chapter I evaluated the extent to which the Kelsenian model of constitutional review at the domestic level is undermined by, or is in tension with, the decentralized system that the ECJ established in Simmenthal. As we have seen, certain features of the E.C. legal order make the potential tension between the two systems a relatively tolerable one. However, to the extent that the ECJ is sometimes very open-ended in its (p.138) rulings because it prefers to grant discretion to the national judiciary, the Simmenthal regime may end up being too disruptive of the foundational values of the Kelsenian model. In those cases, a stronger presence of constitutional courts is to be welcomed.

In general, we should find ways to ensure that the logic of the system of judicial review of legislation under E.C. law is not too different from the logic of the domestic system of constitutional review patterned after the Kelsenian model. If the difference is too large, the two systems may not be able to coexist smoothly. Their combination may become explosive at some point. It is necessary to introduce some adjustments in order to make them less distant from each other. In chapter 10 I suggested a marginal correction to the Kelsenian model, in a decentralizing direction, that would enlarge the role of ordinary judges in the area of constitutional review. This correction puts the domestic system of constitutional review closer to the institutional arrangement that stems from Simmenthal. But, as I have suggested in this chapter, some adjustments should also be introduced in the other direction: the Simmenthal system needs to be refined so that it does not depart too radically from the Kelsenian intuition in favor of centralization.


(1.) For an overview of the evolution that led to the establishment of these three pillars, and to several rearrangements within each of them, see Paul Craig and Gráinne de Búrca, E.U. Law: Texts, Cases, and Materials (Oxford: Oxford University Press, 2008), 1–37. It should be noted that, as of this writing, the Treaty of Lisbon, which was signed on December 13, 2007, has not yet been ratified by all member states. Although this treaty will change E.C. law in various ways, the legal doctrines and the judicial architecture that I examine in this chapter will not be affected at the foundational level.

(p.212) (2.) There are actually three E.C. judicial bodies in Luxembourg: the European Court of Justice (ECJ) (which was originally the only court), the Court of First Instance (which was established in 1988), and the Civil Service Tribunal (which was established more recently, in 2004). They are all part of the so-called Court of Justice of the European Communities. In the text, I focus on the ECJ, since this is the only court that currently has the authority to render preliminary rulings, which play a key role, as we will see.

(3.) The “general version” of the preliminary reference is regulated in article 234 of the E.C. treaty (and article 150 of the Euratom treaty). Apart from this general version, there are two other types. One is the preliminary reference that operates in the context of articles 61–69 of the E.C. treaty, which deals with visas, asylum, immigration, and other policies concerning the free movement of persons. This type of preliminary ruling can only be sought by a national court against whose decisions there is no judicial remedy in national law. The other type was introduced by article 35 of the Treaty on European Union and covers police and judicial cooperation in criminal matters under the third pillar. The ECJ can give preliminary rulings on the interpretation and validity of certain measures adopted under this pillar, but only if the member state accepts the ECJ's jurisdiction by making a declaration to this effect. The member state has the further choice whether a preliminary ruling can be sought by any court or only by a court against whose decisions there is no judicial remedy in national law. In the text, I discuss only the “general” procedure regulated by article 234 of the E.C. treaty. For a detailed description of all these mechanisms, see Craig and de Búrca, E.U. Law, 460–501.

(4.) Actually, the ECJ has held that national courts cannot set aside E.C. acts on their own authority, even if they conclude that the acts are invalid. If a national court concludes that an E.C. act is invalid, the court must petition the ECJ to pass judgment on that act. See Case 314/85, Foto-Frost v. Hauptzollamt Lübeck Ost, 1987 E.C.R. 4199.

(5.) Craig and de Búrca, E.U. Law, 460.

(6.) Case 6/64, Costa v. ENL, 1964 E.C.R. 585.

(7.) Joseph Weiler, “The Transformation of Europe,” in The Constitution of Europe (Cambridge, U.K.: Cambridge University Press, 1999), 10–101.

(8.) Case 106/77, Amministrazione delle Finanze dello Stato v. Simmenthal S.p.A, 1978 E.C.R. 629.

(9.) Ibid., par. 24 (emphasis added).

(10.) The Italian Constitutional Court, for example, finally endorsed the Simmenthal doctrine in the Granital case, decision 170/1984, June 5, 1984.

(11.) On the different and complex ways in which domestic legal systems have incorporated this basic principle of E.C. law, see Anne-Marie Slaughter, Alec Stone Sweet, and J. H. H. Weiler, The European Court and National Courts: Doctrine and Jurisprudence (Oxford: Hart Publishing, 1998).

(12.) Although these arguments are rather underdeveloped in the court's decision, (p.213) they are more explicit in the opinion by Advocate General Gerhard Reischl.

(13.) Case 106/77, Simmenthal, pars. 23, 26.

(14.) There is an additional way to criticize the ECJ's efficiency argument. Since it requires some digression, I present it here. The ECJ saw no efficiency problem when it later held, in the Foto-Frost case, that when a national court believes that a legal provision enacted by the European Community (or even an administrative or executive decision of the European Community) is invalid under E.C. law, the court cannot simply set the provision aside, but must instead send a preliminary question to the ECJ. This important doctrine gives the ECJ a monopoly when it comes to determining that an E.C. act is invalid. “National courts have no jurisdiction themselves to declare that acts of Community institutions are invalid.” Case 314/85, Foto-Frost, par. 20. One of the rationales for this centralized arrangement is uniformity and legal certainty, exactly the same values that figure so prominently in the classical justification of the Kelsenian model of constitutional review at the domestic level. As the ECJ said in its Foto-Frost decision, “Divergences between courts in the Member States as to the validity of Community acts would be liable to place in jeopardy the very unity of the Community legal order and detract from the fundamental requirement of legal certainty.” Ibid., par. 15. This centralization is very strict, moreover, as strict as the Kelsenian system typically is. In a later case, Gaston Schul Douane-expediteur B.V., the ECJ made clear that the national court is not exempted from the duty to raise a question to the ECJ, even in the extreme case in which the E.C. act to be scrutinized for its validity is very similar to another act that has already been held invalid by the ECJ. The national court is still required to seek a ruling from Luxembourg on a question relating to the validity of the provisions of a regulation “even where the Court has already declared invalid analogous provisions of another comparable regulation.” Case C-461/03, Gaston Schul Douane-expediteur B.V. v. Minister van Landbouw, 2005 E.C.R. I-10513, par. 25. This means, of course, that the national court will have to stay the proceedings (although it can decree interim measures, see Cases C-143/88 and C-92/89, Zuckerfabrik Süderdithmarschen AG v Hauptzollamt Itzehoe and Zuckerfabrik Soest GmbH v Hauptzollamt Paderborn, 1991 E.C.R. I-415) and therefore wait until the ECJ rules on the validity of the relevant E.C. act. The ECJ, however, asserted: “Reducing the length of the proceedings cannot serve as justification for undermining the sole jurisdiction of Community Courts to rule on the validity of Community law.” Case C-461/03, Gaston Schul, par. 23. (It should be noted that the average time that the ECJ needs to issue its ruling is currently 20.4 months.) The ECJ was probably right in Foto-Frost: it is more efficient for the ECJ to be the only court that can decide that an E.C. act is invalid, and to do so through a single decision that has general effects, than to allow all the different national courts to decide for themselves (p.214) whether to set aside the E.C. act in question. (See, however, the conclusions by Advocate General Dámaso Ruiz-Jarabo in the Gaston Schul case. He claimed that it is too rigid for the ECJ to insist on its monopoly when the act to be reviewed is very similar to one that has already been declared invalid.)

(15.) In its decision, the ECJ noted that the treaty “provides that any court or tribunal of a Member State is entitled to make a reference to the Court whenever it considers that a preliminary ruling on a question of interpretation or validity relating to Community law is necessary to enable it to give judgment,” Case 106/77, Simmenthal, par. 19, and it concluded: “The effectiveness of that provision would be impaired if the national court were prevented from forthwith applying Community law in accordance with the decision or the case-law of the Court.” Ibid., par. 20. Advocate General Gerhard Reischl was more explicit in his opinion: if the national court needs clarification, it must refer the case to the ECJ, not the national constitutional court.

(16.) See, for example, the references raised some years ago by the constitutional courts of Belgium and Austria in the cases C-93/97, Federation Belge des Chambres Syndicales de Médecins ASBL v. Flemish Government, 1998 E.C.R. I-4837, and C-143/99 Adria-Wien Pipeline GmbH and Wietersdorfer & Peggauer Zementwerke GmbH v. Finanzlandesdirektion für Kärnten, 2001 E.C.R. I-8365. For a more recent preliminary reference by the Belgian Constitutional Court, see C-303/05, Advocaten voor de Wereld VZW v. Leden van de Ministerraad, 2007 E.C.R. I-3633. See also Corte costituzionale decision no. 102/2008, February 13, 2008 from Italy.

(17.) Balancing is quite prominent in the jurisprudence of the ECJ in this area. On the reasoning of the court, see Joxerramon Bengoetxea, Neil MacCormick, and Leonor Soriano, “Integration and Integrity in the Legal Reasoning of the European Court of Justice,” in The European Court of Justice, ed. Gráinne de Búrca and Joseph Weiler (Oxford: Oxford University Press, 2001), 43–85.

(18.) Because the member states have been transferring more and more powers to the European Community, the actions of the European Community need to be constrained by fundamental rights. In the 1970s, the ECJ started to recognize fundamental rights as unwritten “general principles” of E.C. law and drew inspiration from the common constitutional traditions of the member states and from various international instruments, the most relevant of which is the European Convention on Human Rights. For a general view of this evolution, see Bruno de Witte, “The Past and Future Role of the European Court of Justice in the Protection of Human Rights,” in The E.U. and Human Rights, ed. Philip Alston, Mara Bustelo, and James Heenan (Oxford: Oxford University Press, 1999), 859–897. States are also bound by the fundamental rights that flow from E.C. law, but only when they are acting within the scope of application of E.C. law. (p.215) See Craig and de Búrca, E.U. Law, 395–402. It must be noted, incidentally, that the Charter of Fundamental Rights of the European Union, which was proclaimed in the Nice summit of December 2000 and was included in the failed treaty to establish a “Constitution for Europe,” has increased the visibility of E.C. fundamental rights among national judges. (The new Treaty of Lisbon, signed on December 13, 2007, formally incorporates the charter into E.U. law.)

(19.) The ECJ's decision in Case C-224/01, Köbler v. Republik Österreich, 2003 E.C.R. I-10239, is a good step to reinforce this duty. Under the Köbler doctrine, a member state can be liable for damages caused by grave infringements of E.C. law stemming from a court's decision. One of the relevant factors in determining whether there has been a serious breach of E.C. law is “non-compliance by the court in question with its obligation to make a reference for a preliminary ruling under the third paragraph of Article 234 EC.” Ibid., par. 55. And the court added: “In any event, an infringement of Community law will be sufficiently serious where the decision concerned was made in manifest breach of the case-law of the Court in the matter.” Ibid., par. 56.

(20.) The ECJ announces on its Web site that its decisions are binding on all the national courts of the member states. The Web site includes a page with answers to frequently asked questions. Question 8 reads: “Are national courts obliged to follow the interpretation of the Court of Justice?” And the answer is: “Yes. Whenever the Court decides that a Community act is incompatible with the treaties, or whenever it gives an interpretation of Community law, that decision is legally binding and the national court which made the reference and all the other courts in the Member States are obliged to follow that ruling. The national courts are therefore bound by the interpretation of the Court. The same is true for all other public authorities.” Curia, “FAQ: Your Questions about the Court of Justice of the European Communities,” http://curia.europa.eu/en/instit/services/dpi/faq1.htm (question 8). The textbooks that European students read in law school, moreover, insist on the binding character of the ECJ's doctrines. As Karen Alter has explained, the role of E.C.-law experts in strengthening the authority of the ECJ has historically been of great significance. These experts started to create associations to promote knowledge of E.C. law and to spread the belief in the authority of the ECJ. One of their strategies was to publicize and criticize national judicial decisions that departed from the ECJ's case law. See Karen Alter, Establishing the Supremacy of European Law: The Making of an International Rule of Law in Europe (Oxford: Oxford University Press, 2001), 58.

(21.) These exceptions were defined by the ECJ in Case 283/81, CILFIT v. Ministry of Health, 1982 E.C.R. 3415.

(22.) Various proposals have been made to deal with the ECJ's excessive workload, which is a consequence of the increasing number of preliminary (p.216) references brought to it. See Alan Dashwood and Angus Johnston, eds., The Future of the Judicial System of the European Union (Oxford: Hart Publishing, 2001), and Paul Craig, “The Jurisdiction of the Community Courts Reconsidered,” in The European Court of Justice (see note 17), 177–214. There seems to be a general consensus that in the near future, national judges will have to share more responsibility in the interpretation of E.C. law so that the ECJ can focus on the most important and controversial questions. See Thomas de la Mare, “Article 177 in Social and Political Context,” in The Evolution of E.U. Law, ed. Paul Craig and Gráinne de Búrca (Oxford: Oxford University Press, 1999), 228–233.

(23.) Joseph Weiler, “Epilogue: The Judicial Après Nice,” in The European Court of Justice (see note 17), 225.

(24.) On this bifurcation, see Mitchel Lasser, Judicial Deliberations: A Comparative Analysis of Judicial Transparency and Legitimacy (Oxford: Oxford University Press, 2004), 103–141, 203–238.

(25.) Sometimes, however, the ECJ does not offer national judges sufficient guidance. For a criticism of the ECJ on this ground, see Julio Baquero, “De la cuestión prejudicial a la casación europea: Reflexiones sobre la eficacia y la uniformidad del Derecho de la Unión,” Revista española de Derecho Europeo 13 (2005): 35–58.

(26.) Daniel Sarmiento, Poder judicial e integración europea: La construcción de un modelo jurisdiccional para la Unión (Madrid: Civitas, 2004), 294–338.

(27.) Case C-368/95, Vereinigte Familiapress Zeitungsverlags und vertriebs GmbH v. Heinrich Bauer Verlag, 1997 E.C.R. I-3689.

(28.) Ibid., par. 29.

(29.) In France, for example, Dominique Rousseau has suggested that the French Constitutional Council should include E.C. law as part of the bloc de la constitutionnalité that it must guarantee against ordinary legislation. Actually, in some decisions, the court has already taken E.C. law into account in determining the validity of French statutes. See Dominique Rousseau, Droit du contentieux constitutionnel (Paris: Montchrestien, 2006), 115–122, 204–207. The Italian Constitutional Court can invalidate statutes under E.C. law in the context of abstract challenges (in connection with legislative conflicts between the state and the regions or between regions). For an updated description, see Federico Sorrentino, “Il diritto europeo nella giurisprudenza della Corte costituzionale: Problemi e prospettive,” available at http://www.cortecostituzionale.it/informazione/eventi/eventi.asp?anno=2006. Ricardo Alonso García has suggested some types of cases in which the intervention of the Spanish Constitutional Court to strike down a statute in violation of E.C. law would be appropriate. See Ricardo Alonso García, El juez español y el Derecho comunitario (Madrid: Consejo General del Poder Judicial, 2003), 53–60. In an important report released in February 2006, the Spanish Consejo de Estado, in its capacity as an advisory body of the government, has advocated reforms that would allow the (p.217) Constitutional Court to determine the validity of Spanish statutes under E.C. law, as a complementary mechanism to the Simmenthal arrangement. The report, together with doctrinal commentary on it, can be found in Francisco Rubio Llorente and José Álvarez Junco, eds., El informe del Consejo de Estado sobre la reforma constitucional: Texto del informe y debates académicos (Madrid: Consejo de Estado y Centro de Estudios Políticos y Constitucionales, 2006), 116–117.

(30.) See article 23 of the Statute of the Court of Justice.

(31.) Thomas de la Mare has rightly emphasized the relevance of this participation, on the basis of discourse theory. If the court is to issue a ruling that will be binding on all member states, the latter should be allowed to take part in the procedure that will yield the ruling. See Thomas de la Mare, “Article 177 in Social and Political Context,” 240–249.

(32.) See article 104.3 of the Rules of Procedure of the Court of Justice.

(33.) See article 223 of the E.C. treaty.

(34.) Case C-450/93, Kalanke v. Freie Hansestadt Bremen, 1995 E.C.R. I-3051.

(35.) For some ideas and proposals, see Mattias Kumm and Victor Ferreres Comella, “The Primacy Clause of the Constitutional Treaty and the Future of Constitutional Conflict in the European Union,” International Journal of Constitutional Law 3 (2005): 473–492, and Aida Torres Pérez, Conflicts of Rights in the European Union: A Theory of Supranational Adjudication (Oxford: Oxford University Press, 2009).

(36.) In this regard, the French version of the centralized model of judicial review, until its recent reform in 2008, made it more difficult for the Constitutional Council to intervene in a conversation with the ECJ. The court could review statutes only in abstract review proceedings, and only before they were promulgated. In contrast, the German Constitutional Court, for example, could speak to E.C. legal issues in the context of concrete cases. On this difference, see Karen Alter, Establishing the Supremacy of European Law.

(37.) Dieter Grimm, “The European Court of Justice and National Courts: The German Constitutional Perspective after the Maastricht Decision,” Columbia Journal of European Law 3 (1997): 238, 241.