judicial Politics
93
the legal realm, where the strength of a government's legal argument will at least in part be determined by the legal discourse defined by the court. To the extent the court can select amongst multiple justifiable legal reasonings, it can compel the governments to behave differently than they would if left to their own devices.
Second, international courts can rely on sub-state actors such as lower courts and interest groups to circumvent national governments. For example, Burley and Mattli (1993) argue that the CJEU was able to persuade national courts to directly enforce EU law and treat it as superior to national law. The preliminary ruling procedure, where a national court asks the CJEU if the interpretation or validity of an EU law is in question, on issues when a decision is necessary for the court to be able to give a ruling, or where there is no judicial remedy under national law, was intended for private actors to challenge the validity of EU law in national courts (Alter, 2001), but the legal doctrine the CJEU developed transformed into a vehicle for the enforcement of EU law.
Third, international courts may solve collective action problems. The states that are parties to the international treaty may see the court as a facilitator of co-operation. Assuming that states want co-operation in cases when it is benefit-able, but not if it is too costly, Carrubba and Gabel (2015) propose that the court is an information clearing house that allows states to reveal the cost of compliance. In infringement cases, other states can field briefs in support of either the plaintiff or the defendant. If compliance is seen as costly by other states, they will field briefs in support of the defendant. In contrast, if the cost of compliance is seen as acceptable, they will field in support of the plaintiff. The role of the court is, then, to weigh these against each other. Thus, the role of the court, in this sense, is to facilitate non-compliance in cases where the cost of compliance is too high.
Fourth, international courts may enable states to make credible commitments through mechanisms that raise the ex post cost of violations, thereby mitigating the incentives to renege. The argument, here, is that the cost of reneging on an international agreement is higher if an international court has ruled that an action is invalid than if there is no such court to make such a ruling (Pollack, 2003).
The EU Legal System and the Court of Justice of the European Union
'EU law' (which here will be used as shorthand for the legal acts of the European Communities and the European Union) constitutes a separate legal system that is distinct from but closely integrated with international law and the legal systems of the EU member states. This law derives from three main sources.
The Political System of the European Union
Judicial Politics
First, there are the 'primary' acts between the governments of the EU member states, as set out in the various treaties and the other conventions reforming the basic institutional structure of the EU.
Second, there are the 'secondary' legislative and executive acts of the Council, the European Parliament and the Commission, which derive from the articles in the treaties. The EU treaty sets out five kinds of secondary act as follows:
• Regulations, which have general application and are binding on both the EU and the member states;
• Directives, which are addressed to any number of member states, are binding in terms of the result to be achieved and must be transposed into law by the national authorities;
• Decisions, which are addressed to member states or private citizens (or legal entities such as firms) and are binding in their entirety;
• Recommendations, which can be addressed to any member state or citizen but which are not binding; and
• Opinions, which have the same force as recommendations.
However, these descriptions are somewhat misleading, particularly the distinction between regulations and directives. Directives are often so detailed that they leave little room for discretion in the transposition of the legislation by the member states. Also, through a series ot judgements, the CJEU has made directives more akin to regulations in terms of their ability to confer rights directly on private citizens.
Third, added to these two formal sources of EU law are the 'general principles of law'. As in all legal systems, primary and secondary sources of law are unable to resolve all legal issues. The EU treaty instructs the CJEU to ensure that 'the law is observed', which the CJEU has interpreted to mean that when applying the primary and secondary acts it can apply general legal principles derived from the EU's basic principles (as expressed in other articles in the treaty, such as the preamble) and from the constitutions of the member states. There are four main types of general principle:
• Principles of administrative and legislative legality, which are drawn from various member states' legal traditions, such as 'legal certainty' (laws cannot be applied retroactively, and litigants should have legitimate expectations about EU actions), 'proportionality' (the means to achieve an end should be appropriate) and 'procedural fairness' (such as the right to a hearing and the right of legal professional privilege);
. Economic freedoms, which are drawn from the EU treaty and include the 'four freedoms' (freedom of movement of goods, services, capital and persons), the freedom to trade, and the freedom of competition;
, Fundamental human rights, which are set out in the Charter of Fundamental Rights of the European Union, which is attached to the treaty; and
• Political rights, which have been introduced in declarations by the member states and are referred to in the EU treaty, such as 'transparency' (access to information) and subsidiarity (the EU can only act in policy areas not included in the Treaties if the policy aims cannot be achieved adequately at the national level).
Composition and Operation of the Court of Justice of the European Union
To apply these sources of law, the member states established the CJEU in Luxembourg (not to be confused with the European Court of Human Rights in Strasbourg, which is the court of the Council of Europe). The CJEU has one judge per member state and eleven advocates-general. The treaty lays down how the judges are appointed, as follows:
The Judges and Advocates-General shall be chosen from persons whose independence is beyond doubt and who possess the qualifications required for appointment to the highest judicial offices in their respective countries ... they shall be appointed by common accord of the governments of the Member States for a term of six years ... Every three years there shall be a partial replacement of the Judges and Advocates-General ... The Judges shall elect the President of the Court of Justice from among their number for a term of three years. He [sic] may be re-elected.
(Article 253)
The staggered terms of office of the judges ensure continuity. The other elements of the article are somewhat misleading. In practice, 'by common accord of the member states' means that each member state proposes a judge, whose nomination is then ratified by the other member states. Also, by convention the large member states each appoint one advocate-general, with the remaining places rotating between the smaller member states. In addition, the independence and qualifications of the judges are sometimes compromised. There is little evidence of explicitly political appointments to the CJEU. But, several member states have tended to appoint 'academic lawyers' and civil servants instead
96 The Political System of the European Union
of recruiting judges from the senior ranks of the judiciary (Chalmers, 2015). To ensure the suitability of Judges, the Lisbon Treaty prescribes that candidates have to appear before a panel of former CJEU judges and national courts whose role is to issue an opinion of the suitability of candidates. The key element of the panel's consideration is a one-hour closed hearing, where the focus is on the candidate's legal and professional experience. In the first term in office (2010-14), the panel issued negative opinions in more than 10 per cent of all hearings (Sauve, 2015).
Unlike in the US Supreme Court, there are no provisions for dissenting opinions of judges in the CJEU to be recorded. In fact, the CJEU judges swear an oath to preserve the secrecy of the vote.
The workload of the CJEU has increased dramatically. The number of cases brought before it rose from 79 in 1970 to 279 in 1980, 384 in 1990, 526 in 2000, 531 in 2010 and 735 in 2020. To cope with this increase, the Court of First Instance (CFI) was created in 1989. But it soon became as backlogged as CJEU. Then, with the Lisbon Treaty, the Court of First Instance changed its name to the General Court. Its rules of procedures lays out that:
"the purpose of the General Court is to ensure a uniform interpretation and application of EU law. It decisions can be appealed to the CJEU, but only on the matter of interpretation of the law. The General Court is responsible for dealing with:
• cases lodged by individuals or companies against acts by EU institutions that are addressed to them or that affect them directly and individually, including cases on employment relations between the EU institutions and their staff and actions related to intellectual property rights;
• actions brought by EU Member States against the European Commission;
• actions brought by EU Member States against certain acts of the Council;
• actions seeking compensation for damages caused by EU institutions or bodies.
The number of judges in the General Court is two per EU Member State. The judges are appointed for a renewable term of 6 years and elect amongst themselves a president and a vice-presidentwho serves for three years. Cases brought before the General Court are heard by Chambers sitting with 3 or 5 judges, or, in some cases, a single judge. Judges elect Presidents of Chambers. For each case, one judge is appointed Judge-Rapporteur (who prepares the initial draft of the judgment). The General Court may also sit as a Grand Chamber (15 judges), when required due to the legal complexity or importance of the case."
judicial Politics 97
To deal with the caseload, the CJEU has established procedures to allow cases to be handled in a chamber of three or five judges instead of the full plenary. The Treaty of Nice extended this practice by formally reversing the precedence between the chamber system and the full court, whereby the CJEU now sits in chamber as the general rule, and the 'Grand Chamber' of fifteen judges or the fully plenary of the court only meet on special occasions (Johnston, 2001: 511-12). The Nice Treaty also introduced provisions for the establishment of specialized 'judicial panels' (by unanimity in the Council, following a proposal by the Commission of the CJEU, and after consultation with the European Parliament). The reason for this new practice was the need for a new procedure to deal with EU staff-related cases.
Figure 4.2 shows the number of cases heard by the different benches of the court between 1997 and 2017. During this period, the court headed almost 10,000 cases. The most noticeable trend is that the number of cases almost doubled in the period. More than 50 per cent of all cases were heard by the Chamber of five judges. Until 2005, the full court heard more than 10 per cent of all cases annually. However, since 2005, it has been rare for the full court to hear cases. Moreover, the number of cases heard by the small plenary, the grand chamber after 2003, has been stable in absolute numbers, and approximately halved as a share of the total number of cases. Instead, there has been a steady
hull coiifi
i Small plenary,' Grand chamber - Chamber (5 judges) ; Chamber (3 judges) . (Vice } President
Figure 4.2 Cases heard by bench type, 1997-2010 Source: CJEU annual reports.
98
The Political System of the European Union
increase in the both the absolute number and the proportion of cases heard by the Chamber of three judges. Hence, the main effect of the introduction of the chamber system is that a larger proportion of cases is heard by only three judges rather than the full court. While this change has increased the ability of the Court to deal with the increase in the number ot cases, it may, to the extent that there is variation in the opinion of the judges, decrease the predictability of the rulings from the court. In fact, Malecki (2012) uses the chamber system to locate the preferences of judges on a Europhilia-Euroscepticism scale, finding evidence of heterogeneity in the preferences of judges, and at the same time demonstrating that Chamber composition may matter for how the court rules.
Another innovation by the CJEU was the introduction of the US practice of 'docket control', whereby the court can refuse to hear a case that it thinks should be resolved by a national court. This was initially used on an informal basis. The Treaty of Nice then introduced a new procedure, whereby the CFI had the right to reject referrals from national courts if the referal did not fall under the jurisdiction of the treaty article covering preliminary references from national courts. However, the treaty reform left the wording of the treaty article untouched, thereby rejecting calls by several member states and some members of the CJEU to allow only the domestic courts of last instance to refer cases to the CJEU (see Court of Justice, 1999; Tumor and Mufioz, 2000).
Justice via the CJEU is a long and drawn-out process. In 2013. the average length of proceedings was twenty-seven months, with state aid and competition cases taking on average forty-seven months to complete while cases related to intellectual properly and appeals were substantively faster, about sixteen months. In the period from 2013 to 2017, the duration fell to about sixteen months. The reduction in the duration of state aid and competition cases, down to about twenty-three months, explains most of this reduction in duration length. Various suggestions have been made for speeding up the judicial process, such as creating 'circuit courts' modelled on the US federal legal system (cf. Weiler, 1993). However, further reform would require a substantial overhaul of the EU court system and the national court referrals procedure, which to date the member state governments have refused to contemplate (cf. Craig, 2001).
Jurisdiction of the Court of Justice of the European Union
As defined in the EU treaty, the CJEU has jurisdiction in three main areas. First, the court hears actions brought against member states for failure to comply with their obligations under the EU treaties and EU legislation. These actions, known as 'infringement proceedings', can be brought either by the Commission
Judicial Politics yy
under Article 258, by another member state under Article 259, or in the area of state aid by either the Commission or a member state under Article 108. Article 260 also asserts that the member state concerned 'shall be required to take the necessary measures to comply with the judgement of the CJEU'. The ability of the CJEU to enforce rulings against the member states is limited. Until the Maastricht Treaty the Commission was only able to introduce new infringement proceedings against a state in an effort to embarrass it into submission. However, the Maastricht Treaty enabled the CJEU to impose financial sanctions on a member state if the Commission brought an additional action for failing to comply with the CJEU's original infringement judgement.
Second, like many national constitutional courts, the CJEU has the power of judicial review of EU legislative and executive acts. Under Article 263 the CJFU can review the legality of legislative acts (of the Council and European Parliament) as well as other acts of the Council, the Commission and the European Central Bank. Under this article any member state, the Council, the Commission and the European Parliament can bring an action to the CJEU either on the ground of lack of competence or because of an infringement of the treaty or a procedural requirement. In contrast, the Court of Auditors, the European Central Bank and the Committee of the Regions can only bring actions to protect their own prerogatives. Finally, private citizens can bring actions against a decision by an EU institution that is of direct concern to them. A further aspect of the CJEU's power of judicial review is hearing actions against EU institutions for failing to act when they have been called upon to do so by the EU treaty or a piece of secondary legislation (such as the delegation of powers to the Commission) under Article 265. These actions can be brought by any natural or legal person.
Third, under Article 267 the CJEU has jurisdiction to give preliminary rulings on references by national courts. Under this procedure, any national court can ask the CJEU to issue a ruling on a case that relates to any aspect of EU law. The national courts then have some discretion in determining how they should use the CJEU ruling when making their judgement on the case in question. At face value this suggests that it is the national courts that give the final ruling on many cases of EU law, which was probably the intention of the drafters of the Treaty of Rome. In practice, however, the jurisdiction of the CJEU under this article has been far more significant for the development of EU law and the constitutionalization of the EU than the CJEU's jurisdiction in any other area. The CJEU often interprets EU law in a manner that allows little discretion to be exercised by national courts when applying the CJEU's interpretations. Also, Article 267 rulings constitute the majority of CJEU judgements. On the one hand, this reveals a high penetration of EU law into the national legal systems. On the other
100 The Political System of the European Union
hand, by enabling national courts to enforce CJEU judgements, the preliminary references procedure has the effect of making national courts the lower tier of an integrated EU court system, and the CJEU the quasi-supreme court at its pinnacle.
The CJEU also has jurisdiction over a number of other miscellaneous areas for which a small number of cases are heard each year. These include actions for damages against the EU institutions by a member state or a private individual (under Article 268), acts of the European Council relating to a breach of fundamental rights by a member state (under Article 269), and employment disputes between the EU and the staff of the various EU institutions (under Article 270).
In sum, the Treaty of Rome created a new legal system and a powerful supranational court to police the system. Nevertheless, when signing the treaty the founding fathers probably did not realize the potential long-term implication of their action: the gradual constitutionalization of the EU through the operation of the legal system and the judgements of the CJEU.
Constitutionalization of the European Union
In a judgement in 1986 the CJEU described the founding treaties as a 'constitutional charter' (case 294/83, Parti Ecologiste 'Les Veríš' v. European Parliament [1986], EC R 1339). This was the first time the court had used the term 'constitution' to describe the treaties, although academic lawyers had been pointing to the constitutional status of the treaties for some time (Green, 1969). Nevertheless, the EU constitution lies less in the founding treaties than in the gradual constitutionalization of the EU legal system (Stein, 1981; Hartley, 1986; Mancini, 1989; Weiler, 1991, 1997; Shapiro, 1992). The two central principles of this constitution are the direct effect and the supremacy of EU law, which are classic doctrines in federal legal systems.
Direct Effect: EU Law as the Law of the Land for National Citizens
The direct effect of EU law means that individual citizens have rights under EU law that must be upheld by national courts. This makes EU law 'the law of the land' in the member states (Weiler, 1991: 2413). The CJEU first asserted the direct effect of EU law in a landmark judgement in 1963 (case 26/62, Van Genu en Loos v. Nederlandse Administrate der Belastingen [1963], ECR 1). In this case a private firm sought to invoke EC law against the Dutch customs authority in a Dutch court, and the Dutch court consulted the CJEU for a preliminary ruling on
judicial Politics iui
whether EC law applied. Four of the then six member states argued to the court that the specific article in the EC treaty to which the case referred did not have direct effect. Despite the opposition of the majority of the signatories of the treaty, the CJEU ruled that individuals did have the right to invoke EC law because 'the Community constitutes a new legal order... the subjects of which comprise not only member states but also their nationals'. This was accepted by the Dutch court. This ruling meant that direct effect applied to primary treaty articles, and in subsequent judgements the CJEU expanded the doctrine to all categories of legal acts of the EU
However, direct effect works differently for regulations and directives. Regulations have a vertical and a horizontal direct effect, meaning that citizens can defend their rights against both the state (vertical) and other individuals or legal entities (horizontal) But, in the case of directives the CJEU has taken the view that these only have a vertical direct effect because they must be transposed into national law by the member states (case 152/84. Marshall I [1986], ECR 723; case C-91/92, Faccini Dori [1994] ECR 1-3325).
To compensate for the lack of a horizontal direct effect of directives the CJEU has developed the doctrine of 'states' liability'. This implies that the state is liable for all infringements of EU directives. For example, when an Italian firm became insolvent and failed to make redundancy payments to its employees, the CJEU ruled that the Italian state should foot the bill because it had not properly transposed Directive 80/987, which required the establishment of guarantee funds for redundancy compensation (cases C-6,9/90, Francovich I [1991], ECR 1-5357)
The central implication of direct effect is that EU law is more like domestic law than international law. The subjects of international law are states: if a state fails to abide by its obligations under an international convention, individuals cannot invoke the convention in their national courts unless the convention has been incorporated into domestic law. In contrast to international law, the subjects of domestic law and EU law are private citizens who can invoke their rights in domestic courts.
The establishment of the doctrine of direct effect led to a dramatic increase in the number of cases brought by individuals to national courts to defend their rights under EU law. The effect, as Weiler (1991: 2414) argues, was that 'individuals ... became the "guardians" of the legal integrity of Community law within Europe similar to the way that individuals in the United States have been the principal actors in ensuring the vindication of the Bill of Rights and other federal law'.
102
The Political System of the European Union
judicial Politics
103
Supremacy: EU Law as the Higher Law of the Land
Unlike the US constitution, the Treaty of Rome did not contain a 'supremacy clause' stating that in the event of a conflict between national and EC law, EC law would be supreme. However, shortly after establishing the direct effect of EC law the CJEU asserted the supremacy of EC law, and like direct effect this doctrine was confirmed and reinforced in subsequent rulings.
The landmark judgement on this doctrine was in the case of Coste v. ENEL in 1964 (case 6/64 [1964], ECR 585). An Italian court asked the CJEU to give a preliminary ruling on a case in which there was a clear contradiction between Italian and EC law. The CJEU duly argued:
By creating a Community of unlimited duration, having its own institutions, its own personality, [and] its own legal capacity . . 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. The integration into the laws of each member slate ot provisions which derive from the Community ... make it impossible for the states, as a corollary, to accord precedence to a unilateral and subsequent measure over a legal system accepted by them on a basis of reciprocity.
In other words, the CJEU held that the doctrine of supremacy was implicit in the transfer of competences to the EC level and the direct effect of EC law.
Formally speaking, EU law takes superiority over national law only in those areas in which EU law applies. But, as the competences of the EU have expanded into almost all areas of public policy, the application of supremacy no longer applies to the 'limited fields' to which the CJEU referred in 1964. Also through successive judgements the CJEU has established that supremacy applies to all EU norms, be it an article in the treaties, a secondary act by the EU institutions (no matter how minor, such as administrative regulations of the Commission) or even a 'general principle of EU law', as defined by the CJEU
As a result, the supremacy doctrine has further distanced the EU legal system from international law. Direct effect is insufficient by itself to establish the EU logal system as a system of domestic law. When international conventions are incorporated into domestic law, individuals can invoke them in domestic courts. But, if a domestic legislature subsequently adopts a national law that contravenes the international convention, the provisions of the international law no longer apply. With the supremacy of EU law, in contrast, national legislative majorities are permanently bound by the provisions of EU law. Weiler (1991: 2415) therefore concludes that parallels of this kind of constitutional order ... may be found only in the internal constitutional order of federal states'. By establishing the dual
I
doctrines of direct effect and supremacy of EU law, the CJEU has transformed the EU from an international organization to a quasi-federal polity.
Integration through Law and Economic Constitutionalism
The application of these basic doctrines has enabled the CJEU to play a central role in the economic and political integration of the EU. For example, in the area of economic freedoms Article 34 states simply that 'quantitative restrictions on imports and all measures having equivalent effect shall be prohibited between the member states'. While this article seems pretty innocuous, through a series of judgements the CJEU has transformed the EU's economic system on the basis of the article (Alter and Meunier-Aitsahalia, 1994).
In 1974, in the Dassonville decision (case 8/74 [1974], ECR 837), the CJEU declared illegal any national rule that was 'capable of hindering, actually or potentially, directly or indirectly intra-Community trade'. Such hindrances included not only quotas and other restrictions on imports, but also internal rules that affected the competitive position of imported goods. The implication of this interpretation became clear with the Cassis de Dijon judgement in 1979 (case 120/78 [1979], ECR 837). In this decision the CJEU ruled that a German law specifying (hat a 'liquor' must have an alcohol content of at least 25 per cent could not prevent the marketing of the French drink Cassis de Dijon in Germany as a liquor, despite it having an alcohol content of less than 20 per cent. This is known as the principle of 'mutual recognition': that is, any product that can be legally sold in one member state can be legally sold anywhere in the EU. Mutual recognition subsequently became one of the basic principles in the establishment of the single market (see Chapter 8).
This interpretation of the treaty obligations is inherently deregulatory. It has forced member states to delete numerous social and economic rules that in many cases were established as expressions of particular social, cultural and ideological preferences. The effect is a specific type of 'economic constitution', whereby competition between different national regulatory regimes has the potential of facilitating a 'race to the bottom' (cf. Joerges, 1994; Chalmers, 1995; Streit and Mussler, 1995; Maduro, 1997) (see Chapter 8).
State-Like Properties: External Sovereignty and Internal Coercion
As we discussed in Chapter 1, the EU is not a state. In particular, until the Lisbon Treaty, the EU did not have the right to sign international treaties. And, the EU still does not have a legitimate internal monopoly on the use of coercion to enforce
104
The Political System of the European Union
1
its decisions. Nevertheless, the CJEU has been instrumental in developing statelike properties for the EU in both these areas.
First, on the external side the EU has always had the power to make treaties with third parties under Articles 207 (common commercial policy) and 217 (association agreements). Even in these limited fields, though, most member states originally considered that the articles merely allowed the Commission to negotiate agreements on behalf of the member states, and that sovereignty remained with the member states. However, in 1971 the CJEU established the principle that when making agreements with third countries the EU would be sovereign over any existing or future acts between the individual member states and the third countries in question (case 22/70, ER TA [1971], ECR 263). In the same judgement the CJEU argued that the jurisdiction of the EU in the international sphere covered all areas of EU competence, not just those included in these two treaty articles. In other words, in one stroke the CJEU conferred new treaty-making powers to the EU and deprived the member states of their own independent powers relating to EU competences. In a sense this CJEU interpretation was a precursor of Article 216 in the Lisbon Treaty, which enables the EU to conclude agreements with any third countries or international organizations which are legally binding for both the EU and its member states.
Second, on the internal side, Article 4 of the treaty instructs the member states to 'take any appropriate measure ... to ensure the fulfilment of the obligations arising out of the Treaties or resulting from the acts of the institutions'. Most member states originally assumed that this article took effect only in relation to the other treaty articles and EU law. However, the CJEU has used it as a substitute for the lack of direct enforcement powers in the EU system (cf. Shaw, 1996: 208-13; Weatherill and Beaumont, 2004), For example, it ruled that member states must adapt all relevant national rules to the requirements of EU law (cases 205-215/82, Deutsche Milchkontor GmbH v. Germany [1983], ECR 2633), and that Article 10 should be applied to all state organs at all levels of government (Case C-8/88, Germany v. Commission [1990], ECR 1-2321).
Furthermore, the CJEU has broadened the definition of the types of action a member state musl use to enforce EU law. For instance, in 1997 it ruled that the French government should have used the state security forces more effectively to ensure the free movement of goods in the internal market (case C-265/95, Commission v. France [1997]). The court acknowledged that member states should retain exclusive competence as regards the maintenance of public order and the safeguarding of internal security', but it went on to argue that:
it falls to the Court ... to verify ... whether the member state concerned has adopted appropriate measures for ensuring the free movement of goods. ...
Judicial Politics 105
[In the present case] the French police were either not present or did not intervene ... the actions in question were not always rapid ... [and] only a very small number of persons has been identified and prosecuted.
In other words, the EU did not need a police force of its own in order to exercise coercive power. According to the CJEU, the member states were obliged to take all reasonable measures to enforce EU law, including the use of security forces
Kompetenz-Kompetenz; Judicial Review of Competence Conflicts A key weapon in the arsenal of supreme courts in any multilevel political system is the ability to police the boundary of competences between the states and central government: what German constitutional laws call Kompetenz-Kompetenz. Prior to the Lisbon Treaty, which sets out a catalogue of competences, the EU treaty gave no formal powers to the CJEU to undertake this task. The treaty refered to the principle of subsidiarity, meaning that the EU can only act in areas that are not better tackled at the national level. The European Council then agreed to a set of rules on how this principle should apply; for example, the Commission must prove in the draft of any legislation that the legislation does not breach the principle of subsidiarity. However, it was open to question whether the subsidiarity principle was justiciable before a national court or the CJEU.
In the absence of a catalogue of competences, the CJEU gradually developed a power to police the vertical allocation of competences. Most significant in this respect was the CJEU's decision in 2000 to annul a directive on tobacco advertising and sponsorship (case 376/98, Germany v. European Parliament and Council [2000]). In 1998, the Council and European Parliament had adopted this directive under Article 114 of the EC treaty, covering the harmonization of laws for the completion of the single market. However, the CJEU ruled that 'Article [114] should be available as a legal basis only in cases where obstacles to the exercise of fundamental freedoms and distortion of competition are considerable'. Thus, a ban on tobacco advertising could only be adopted under Article 114 if it allowed products that circulated in the internal market (such as newspapers or magazines) to move more freely than if there were different national tobacco advertising rules. Since the proposed ban was more widespread than simply covering these goods, the CJEU pointed out that 'the national measures affected are to a large extent inspired by public health policy objectives'. However, the public health competences in the treaty (Article 168) only allowed for the adoption of EU legislation on common safety standards in organizations, and hence did not extend to the harmonization of national public health standards more generally.
The Political System of the European Union
Some observers were surprised by the judgement to annul the directive as the CJEU had applied Article 114 quite broadly in the past (Hervey, 2001). However, the ruling can be interpreted as a strategic signal by the CJEU to the governments that it could be trusted in competence-conflict decisions: in this case between the harmonization of rules in the single market (an exclusive EU competence) and public health standards (an exclusive competence of the member states). By ruling that the EU could only harmonize rules in the single market if there was a clear case of market distortion, the CJEU effectively defined a boundary between the federal powers of the EU and the rights of the member states.
This was particularly significant because the CJEU judges were aware that the Convention on the Future of Europe was about to begin and that one of the key issues in the design of an EU constitution would be a catalogue of competences in the EU and who would be responsible for policing such a catalogue. Several member states had already proposed a new quasi-judicial body for this task: a special EU constitutional court composed of either national parliamentarians or judges from the highest courts in the member states. By ruling against the legislative majorities in the Council and European Parliament, the CJEU demonstrated that it couid be trusted to protect the rights of states that were on the losing side in the EU's legislative process.
The Lisbon Treaty subsequently established a catalogue of competences, with areas defined as either exclusive competences of the EU (Article 3), shared competences between the EU and the member states (Article 4), areas for co-ordination of economic and social policies (Article 5) and areas where EU action can supplement the actions of the member states (Article 6). The Lisbon Treaty also established that the CJEU has the sole right to police the boundaries between these competences. In other words, following the tobacco advertising ruling the member states decided to grant exclusive Kompetenz-Kompetenz to the CJEU rather than to a new body. Without the tobacco advertising judgement, it is unlikely that the governments would have been able to formally establish this important new power.
Penetration of EU Law into National Legal Systems
The penetration of EU law into national legal systems has developed both quantitatively and qualitatively. On the quantitative side there has been a substantial increase in the use of the Article 267 procedure for requesting
judicial Politics
preliminary rulings from the CJEU by national courts, and on the qualitative side national courts have gradually accepted the existence and supremacy of the EU legal system over national law and constitutions.
Quantitative: National Courts' use of CJEU Preliminary Rulings
Figure 4.3 shows the number of Article 267 references by all member state courts to the CJEU in 1961-2017. During this period, while the EU grew from six member states to twenty-eight, the number of references to the CJEU rose from one or two a year in the early 1960s to over 250 a year by the late 1990s. Then, the number of referrals was stable until the 2010s, when it again doubled to above 500 by 2017. The rapid rise in references in the 1970s followed the establishment of the doctrines of direct effect and supremacy, which encouraged national courts to use the references procedure to strengthen their positions in their domestic political systems, and encouraged private litigants to use the procedure to invoke their rights in their domestic courts. Although the number of annual referrals fell in the years prior to the Central, Eastern and Southern enlargement in 2004, the volume of referrals has since increased dramatically. Moreover, the substantive importance of the topics that the court currently addresses also increased and evolved. As Chalmers (2015: 72) puts it, these now range from:
/ - '.\/v
..-v\/
I960 1970 1980 1
Figure 4.3 Growth of CJEU referrals, 1961-2019 Source: CJEU annual reports.
108 The Political System of the European Union
[the legality of] the European Stability Mechanism (C-370/12 Pringle), start of life (C-34/I0 Brüstte), the characteristics of body shape (C-354/13 Kaltoft), parental access rights to children (C-400/10 PPU McB), the parameters of national security (C-300/11 ZZ v Secretary of State for the Home Department), and collective provision for the old (C-399/09). These issues have an icono-graphic pre-eminence that, no matter how they are dressed up, cases of customs duties on ureaformaldehyde or unpaid utility bills simply do not have.
However, not all member states' courts have used the references system to the same extent. As Figure 4.4 shows, the number of referrals from most member states has risen over time The figures suggest a 'learning curve', however, with the original member states making more references in each period than the member states that joined later. Nevertheless, several other factors cross-cut this trend. First, within each wave of EU members, the larger states have made more references than the smaller states: Germany, France and Italy have made more references than the Netherlands, Belgium and Luxembourg; the UK have made more than Denmark and Ireland; Spain has made more than Portugal; Austria and Sweden made more than Finland; and Poland has made more than the other member states who joined in 2004 and 2007, Second, despite the learning curve, British courts made fewer references in the early 1990s than did Dutch and Belgian courts, which perhaps reflects the sceptical attitude towards the EU among the public and the elite in the UK (Golub, 1996). However, the courts
7
Figure 4.4 Smoothed annual number of CJEU referrals, by selected member states Source: CJEU annual reports.
Judicial Politics 109
in Ireland, Portugal and Luxembourg, where the public and elites are strongly pro-European, also made few references to the CJEU.
Stone Sweet and Brunell (1998a, 1998b, 2000) consequently argue that other factors might explain the variations in member state usage of the referral mechanism. In particular, a significant proportion of the variation is explained by the size and openness of a member states' economy. In other words, the larger the market and the larger the volume of imports, the greater the incentive for importing firms to take cases to the CJEU to guarantee market access for their goods and services. In line with this, the increase in the number of referrals since 2004 cannot be attributed to the arrival of twelve new member states. Instead, it might be due to the fact that Germany, the Netherlands and Belgium, who are the major net exporters in the EU single market, have substantively increased their number of referrals. Together with Italy and France, these five countries have been responsible for more than 50 per cent of all referrals since 2010.
The subject matter of references to the CJEU by national courts has also changed significantly. In the early period most references related to the Common Agricultural Policy. However, by the early 1990s, issues relating to the operation of the internal market - such as the free movement of goods, the tree movement of workers, taxes, freedom of establishment and the approximation of national laws - comprised over half of all references. This reflects the fact that the majority of laws governing the regulation of the market were set at the European rather than the national level (see Chapter 8). Then, by the end of the 2010s, cases on freedom, security and justice, intellectual and industrial property, taxation and transport came to be the dominant subject matters of new cases, reflecting the fact that EU law by now goes to the heart of what was used to be considered the core responsibilities of the state (see Chapters 10 and 11).
Finally, there has been considerable national variation in the extent of compliance with EU law Table 4.1 shows the total number of infringement cases brought before the CJEU in 2015-19. Although it may have been the case in earlier times, these figures no longer support the conventional wisdom that the southern EU states are generally less likely to enforce EU law as effectively as the northern states. To better understand these national differences in compliance rates, Falkner et at (2005) conduct a large qualitative study, resulting in a typology of 'worlds of compliance'. They argue that the are differences in the cultures of compliance in member states' political and administrative systems. The 'world of law obedience', represented by the Nordics, there is a culture of respect for the rule of law which usually ensures fast and correct transposition. In contrast, in the 'world of neglect', represented by countries like Greece, France and Portugal, the absence of such a culture typically leads to long periods of bureaucratic
110 The Political System of the European Union
Table 4.1 Infringement cases declared and dismissed, 2015-19
	Declared	Dismis
Greece	18	0
Germany	13	1
Italy	10	1
Spain	9	1
Poland	9	0
Portugal	9	0
France	7	0
Belgium	7	0
UK	6	2
Bulgaria	w	1
Luxembourg	4	0
Hungary	4	0
Ireland	4	1
Czech Republic	3	1
Slovenia	3	0
Netherlands	2	1
Latvia	2	0
Romania	2	0
Austria	V	2
Sweden	I	0
Denmark	1	1
Slovakia	1	2
Cyprus	1	0
Lithuania	1	0
Malta	1	1
Croatia	1	0
Finland	0	0
Estonia	0	0
Total	127	15
Note: These figures reflect the number of cases in which the CJEU declared that a member state failed to fulfil its treaty obligations, in 2015-19, Source: CJEU annual reports.
Judicial Politics 111
Inertia and apolitical transposition processes. In the 'world of domestic politics' where administrations usually work dutifully, but where a culture of compliance is absent, transposition of EU law depends on the fit with the preferences of the governing parties and other key actors. In this cluster, Falkner et al. (2005) grouped countries like Belgium, Germany, the Netherlands, Spain and the UK. Extending the typology to include Central end Eastern Europe, Falkner and Treib (2008) identified the 'world of dead letters', where transposition is politicized, conducted swiftly with severe shortcomings in actual policy delivery. This cluster consists of the Czech Republic, Slovenia, Hungary and Slovakia.
This classification inspired several authors to look beyond factors related to the administrative and political system for explaining pattern of non-compliance (for an extensive overview of the compliance literature, see Treib 2014). For example, Thomson (2010) highlighting the role of the Commission and used disaggregated measures. He found that protracted non-compliance was rare and that member states policy preferences, rather than culture, affect the likelihood of transposition problems, but only conditional on the behaviour of the Commission. More recently, Thomson et al. (2020) further highlighted the role of incentives in the transposition process, distinguishing between national governments and domestic stakeholders' incentives to deviate, conform and exceed beyond the minimum standards set out in the legislation, finding that incentives to comply or exceed are more common than incentives to deviate. Finally, Pircher and Loxbo (2020), investigated transposition over a twenty-year period from 1997 to 2016, demonstrated that the pattern of transposition has evolved over time. Although their data support the notion of different 'worlds of compliance' in the start of the period, such country-clusters have not been visible in the more recent time period. Differences in incentives rather than cultures are more likely to explain variations in transposition patterns. More recently, though, Fjelstul and Carrubba (2018) demonstrate that it is not clear how much we can conclude from variations in such counts due to strategic behaviour of the member states, the Commission and the court.
Qualitative: National Courts' Acceptance of the EU Legal System
Not all national courts have capitulated to the emerging constitutionalization of the EU (Mattli and Slaughter, 1998a, 1998b). In particular, the German Constitutional Court (Bundesverlassungsgericht) at first accepted direct effect and supremacy, but more recently has challenged the legitimacy of the EU framework (cf. Kokott, 1998; Alter, 2001). In the Brunner judgement, the German court ruled that the German Basic Law limited the transfer of powers to the EU, and argued that the
112 The Political System of the European Union
EU was a sui generis organization and not a state based on democratic norms. The court claimed that because it was commanded by the German constitution to defend the basic rights and principles of democracy set out in the German Basic Law, it had the jurisdiction to declare acts of the EU ultra vires (beyond the legal authority of the EU) if they breached the Basic Law (but it would seek to co-operate with the CJEU if faced with such a prospect). Having said this, the court declared that the Maastricht Treaty could be ratified by Germany because the German parliament maintained the right to transfer (or withdraw) German government competences to the EU. The court warned, however, that the EU could only legitimately become a state if it were fully democratic, with the necessary institutions of parliamentary democracy, a clearly defined hierarchy of rights and a single demos (Weiler, 1995).
In the UK, in contrast, the courts accepted direct effect immediately upon the country's accession to the EU in 1973. However, it was difficult for the UK to accept the supremacy of EU law as this conflicted with the central constitutional concept of parliamentary sovereignty - that is, acts of parliament immediately overrode ail existing law or legislation (Craig, 1998). Nevertheless, in 1990 the House of Lords found a way to reconcile parliamentary sovereignty and EU supremacy. On a reference from the House of Lords, the CJEU ruled that a 1988 parliamentary act was in breach of EU law, and the House of Lords accepted this judgement on the ground that in passing the 1972 act of accession to the EU the British parliament had voluntarily accepted the EU legal system, of which the supremacy of EU law was a central part. The House of Lords also argued that this did not compromise parliamentary sovereignty, as a future British parliament could repeal the act of accession, and thus withdraw the UK from the EU.
Another interesting case is Sweden, where there was dispute over whether the domestic constitution would have to be changed if Sweden became a member of the EU (Bernitz, 2001). The constitution had been amended in 1965 to allow for the conclusion of treaties with the then European Community. But most experts in Sweden took the view that because of the development of the EU and its competences in the 1980s and 1990s, this provision was not enough to allow for the substantial transfer of power that would result from accession. The new constitutional provision that was finally agreed upon by the Riksdag (the Swedish parliament) was significantly less extensive than many legal experts had proposed. The constitutional amendment imposed constraints on EU law that flowed directly from the German Brunner judgement. If an EU law conflicted with a fundamental right that was protected by the national constitution and backed by a national democratic majority, a Swedish court would be forced to reject the EU law unless it was clear that the relevant right was sufficiently protected at the EU level by a European charter of fundamental rights.
1
Judicial Politics 113
In sum, EU law has been accepted as an integral part of national legal systems and as sovereign over national law. However, in several member states the highest national courts maintain that this is conditional on national constitutional norms: for example, that parliaments retain the right to revoke the supremacy of EU law by withdrawing the transfer of sovereignty to the EU (as in Germany, the UK and France). One could argue that this solution has been driven primarily by the desire of national courts not to renounce their previous positions on the EU, or to declare basic constitutional principles null and void (such as parliamentary sovereignty in the British case). Only in Germany did the national constitutional court withdraw from its previously unconditional acceptance of supremacy, but this had profound effects on the other member states, as in the Swedish case, and forced the EU to address the protection of fundamental rights and the democratic accountability of the EU institution, and ultimately to the incorporation of the Charter of Human Rights into EU law in the Lisbon Treaty.
Explanations of EU Judicial Politics: Is the CJEU a Runaway Agent?
Early work from a legal-formalist perspective regarded the CJEU as a heroic promoter of European integration against the wishes of the member states (see Shapiro, 1981; Weiler, 1994). More recently, two competing views of the CJEU have emerged (Alter, 2008). Drawing in the intergovernmental approach, one school of thought understands the CJEU as an agent of the member states; empowered to enforce integration on behalf of the governments. In this view, member states are powerful, the judges are national delegates and the court is sensitive to the interests of the governments, fearing treaty-reforms and non-compliance if seen as acting beyond its mandate. The alternative view, derived from the supranational politics approach, argues that the CJEU has been able to work with self-interested lower national courts, integrationist judges and private interests to move EU integration beyond the intentions of the governments.
Legal-Formalist View: The Hero of European Integration
Legal scholars of the EU have traditionally emphasized the internal logic of law and the legal process. As Weiler (1994: 525) explains: 'The formalistic claim is that judicial process rests above or outside politics, a neutral arena in which courts scientifically interpret the meaning of policy decided by others.' In other words, the CJEU simply applies EU law as set out in the EU treaties and in secondary legislation, without any conscious desire to promote its own power
114
The Political System of the European Union
Judicial Politics
115
or institutional interests An EU constitution has developed because the EU legal system had its own internal 'integrationist' logic. Instead of a hierarchy of norms, the governments established the goal of 'ever closer union' as the ultimate norm. This forced the CJEU to develop the doctrines of direct effect and supremacy. Furthermore, there was an effect utile in the legal workings of the EU, whereby the CJEU preferred to apply EU law in the most efficient and effective way, which compelled the CJEU to promote legal integration in order to prevent the EU political system from becoming ineffective and unworkable (Cappelletti et al., 1986).
In the same vein, legal-formalist explanations posit that national courts were eager to find ways to reconcile their previous jurisprudence with the emerging EU legal system. Through the preliminary references system, the CJEU provided national courts with the appropriate argumentation and rationale for them to absorb the new doctrines into their national legal systems (cf. Wincott. 1995). Variations in the use of the preliminary references system and the dates of acceptance of the CJEU doctrines can be explained by variations in national legal cultures and doctrines (Chalmers, 1997; de Witte, 1998; Mattli and Slaughter, 1998a, 1998b; Maher, 1998; Stone Sweet, 1998). On the cultural side, different systems of training judges, different promotion systems and different career paths had produced different patterns of behaviour and reasoning by judges. Also, each system had a different relationship between administrative, constitutional and common law courts, and different rules, traditions and powers of judicial review. On the doctrinal side, the place of fundamental rights in domestic constitutions and how the concept of sovereignty was defined affected the relationship between national legal norms and the EU constitution. Moreover, when the treaties are incomplete, it is up to the court to interpret and delimit them, and in turn instruct the EU legislature how to interpret the court's ruling, within the appropriate discretionary limits. Legislative override is, in this view, a myth, and the EU legislature is best seen as an agent of the court (Davies, 2016). Also, national judges are compelled to refer cases to it according to Article 267, and the possibility of referral encourages integrationist judges to refer cases where integrationist rulings are possible, and to rely on integrationist principles when ruling on cases in national courts. As such, judicial activism is national as well as supranational (Davies, 2012).
These legal-formalist explanations have some important shortcomings. On an empirical level, the doctrines of supremacy and direct effect are not simply logical extensions of the EU treaty: if federalization of the EU had been intended from the outset, the treaty would have contained a supremacy clause as in other federal constitutions. Also, many national courts were not immediately
convinced of the CJEU's justification of direct effect and supremacy (cf. Alter, 1998a: 230-4). From the general study of courts and judicial politics we know that the institutional interests of courts and the personal policy preferences of judges drive judges' actions. In a sense, the structural and cultural logic of the law is simply another set of constraints within which courts and judges secure these aims. Consequently on a theoretical level, explanations of the emergence of the EU constitution must also take into account the institutional and policy incentives of EU and national judges and the strategic motivations of other actors in the system. Kelemen (2016) points out that the CJEU has, in particular compared to other international courts, enjoyed a remarkable supportive external context ranging from national government to national courts and the broader European legal order. But, as the Court is increasingly involved in politically sensitive issues close to the heart of the national state, and with some national governments turning hostile towards further European integration in the wake of the Eurozone crises and the influx of refugees following the Arab Spring and the Syrian civil war, this supportive external environment should not be taken for granted.
Intergovernmentalism: The CJEU as an Agent of the Member States
From an intergovernmentalist perspective, the development of the EU constitution can be understood as a deliberate strategy by national governments (e.g. Garrett, 1992, 1995; Garrett and Weingast, 1993; Cooter and Drexl, 1994; Garrett et al., 1998; Kelemen, 2001). In this view, governments have consciously allowed the CJEU, national courts and transnational litigants to promote legal integration in the EU because it has been in the governments' political or economic interests. If the CJEU or a national court takes an action that is contrary to a government's interest, the government will simply ignore the ruling. High-profile clashes between national governments and the CJEU or national courts over EU legal issues are rare but not because the governments are powerless in the face of court activism. Instead, courts are careful not to make decisions that threaten government interests, and governments accept decisions that appear to be against them because they are in fact in their long-term interests. Courts are strategic actors. However, they are constrained by the possibility of government threats, such as reform of the treaty or the passing of new legislation. This explains why the CJEU has refused to establish that directives have horizontal direct effect, despite the opinions of several advocates-general and numerous academic lawyers.
The Political System of the European Union
Judicial Politics
111
Garrett (1995) proposed a simple model to explain why governments often accept CJEU rulings against them. The model posits that governments take two main factors into account: the domestic political clout of the industry that is harmed by the CJEU decision and the potential gains to the national economy as a whole. If the industry is domestically weak and the general economic gains will be large the government will accept the CJEU ruling and put up with complaints from the domestic industry For example, with regard to the Cassis de Dijon judgement Garrett argues that the German government accepted a ruling that would damage its (relatively small) spirits industry because the rest of the German economy stood to benefit from the trade liberalization that would result from the principle of mutual recognition. Conversely, if the industry in question is domestically powerful and the general economic gains will be small, the government will engage in 'overt evasion' of the CJEU's decision However, this rarely occurs because the CJEU is careful to avoid such a showdown. The implication is that in the Cassis de Dijon case the CJEU waited for the right case to come along in order to establish the principle of mutual recognition-for a similar model of CJEU behaviour on international trade disputes, see Kelemen (2001).
By focusing on the centrality of national governments in the EU system and conceptualizing their actions as highly rational, these explanations have some of the same limitations as the general intergovernmentalist approach (see Chapter 1). At an empirical level, there is substantial evidence that the CJEU and national courts have often taken decisions that governments have opposed, and which have had negative effects on the competitiveness of national economies in the single market (Mattli and Slaughter, 1995). At a theoretical level, meanwhile, this can be explained by the fact that governments do not have perfect information about the likely outcome of delegating adjudication to the CJEU and national courts (see Alter, 2001: 182-208: cf. Pierson, I996). For example, when the Treaty of Rome was signed, few governments realized that the CJEU would establish the doctrines of direct effect and supremacy or could predict the significance of the preliminary reference procedure (Alter, 1998b).
In later work, Garrett, Kelemen and Schulz (1998) accept that governments are not completely free to ignore adverse rulings. For example, in cases where the EU treaties are clear and the legal precedent is strong, the costs to a government of ignoring an adverse ruling (in terms of threatening the very foundations of the EU) will be high. In other words, although national governments behave strategically, there are long-term constraints on them as a result of their allowing the CJEU to develop its own legal precedents and norms. However, the main thrust of Garrett, Kelemen and Schulz's argument remains: the CJEU is heavily constrained if the potential costs to a powerful domestic constituency are high
or if a large number of governments are likely to be adversely affected by an CJEU ruling. For example, the Barber judgement on equal pension rights for men and women imposed substantial costs on all governments. In response, the governments added a protocol to the treaty that prevented the retroactive application of the judgement, and subsequently the CJEU moderated its activism in this area - although it extended its activities in other areas of pension rights (cf. Pollack, 2003: 360-72).
Supranational Politics: The CJEU as an Independent but Constrained Actor
An alternative explanation depicts the CJEU as an explicitly powerful political actor capable of using its discretion to forward its interests by catering to the interests of national courts and private interests against the interests of the member states. Although the CJEU, unlike, for example, the US Supreme Court, does not publish dissenting opinion, scholars have recently started to tap into the sources of information of the revealed preferences of the judges. As mentioned above, Malecki (2012) draws on chamber composition in an attempt to tease out differences in the preferences of judges, while Hermansen (2020) argues that the president of the CJEU prevents those judges with the most politicizing opinions from serving as a case-rapporteur on cases with politicizing potential. In an attempt to locate the CJEU relative to the Commission, the European Parliament and the member states along the EU integration dimension, Ovadek (2021) finds that the court is more supranational than all of the member states, but less so than the Commission and the Parliament. Moreover, he also finds systematic evidence in line with the notion that the court systematically adjusts its rulings in response to member state preferences. Finally, inviting legal experts to rate judges on the General Court of the EU (formerly the Court of First Instance), Wijtvliet and Dyevre (2021) find that the pro-business attitudes, rather than pro-integration, matter for the General courts decisions on state-aid and competition cases
Karen Alter (2008) highlights (our steps that are crucial for understanding the development of judicial politics in the EU. First, since the CJEU can only rule on specific cases, it relies on other actors to provide cases it can rule on. The main non-state actors that have sufficient resources to bring cases to the CJEU are private firms and interest groups. Stone Sweet and Brunei (1998a, 1998b) argue that firms involved in the import and export of goods are the dominant private litigants in the EU legal system. These interests have a particular incentive to secure effective application of the free movement of goods and services, and have sufficient resources to take actions all the way
118
The Political System of the European Union
Judicial Politics
119
through to the CJEU. Regarding interest groups, Lisa Conant (2002) argues that variations in the incentives of groups explain why EU law has developed in areas other than those of direct interest to the CJEU or national governments. Concentrated interests (who potentially face large costs/benefits from EU law) tend to be better organized than diffuse interests (who potentially face small costs/benefits from EU law) (see Chapter 6). But, the relatively low cost of gaining access to the CJEU means that EU law has been a vehicle for the promotion of some interests that are underrepresented in several domestic systems of interest representation (Pollack, 1997).
Second, the CJEU can only ruie on cases that get referred to it. Most cases involving European law are not referred. Unlike the CJEU, national courts are not interested in the emergence of an EU constitution to promote the goal of European integration. Instead, national courts use the EU legal system to secure their interests and policy preferences within their national contexts (e.g. Weiier, 1991, 1994), In many domestic political systems judicial review is weak, parliaments are sovereign and governments have substantial administrative and political resources at their disposal. Consequently, national courts welcome the direct effect and supremacy of EU law and actively use the preliminary references system to strengthen their hand in the national policy process. However, Alter (2001) and Golub (1996) argue that national judges are selective in the cases they bring to the CJEU.
Also, lower and higher courts have different incentives vis-á-vis the EU legal system. Alter (1996, 1998a) contends that lower courts can use the preliminary references procedure they are able to play higher courts and the CJEU off against each other to influence legal developments in their preferred direction. As a result, lower courts have made more use of the preliminary references procedure than higher courts (Stone Sweet and Brunell, 1998b).
Third, however, activism by the CJEU has not been linear (Chalmers, 1997). Rather, CJEU activism has responded to the pace of the integration process, and has been sensitive to anti-CJEU feelings amongst certain national governments. Carrubba et al. (2008) show that the CJEU takes member state preferences into account in their judicial rulings. Specifically, as the number of governments contributing observations in support of an accused government increases, the probability of the CJEU ruling in favour of the accused government also increases. This effect is substantively larger when a government is the litigant rather than the Commission or a private individual. This suggests that the CJEU responds to the strength of feeling amongst the member states on a case-by-case basis. This helps explain why the CJEU seems inconsistent in its attempts to promote European integration (for an extended version
of this argument, see Carrubba and Gabel, 2015). The extent of this constraint member states imposed on the behaviour of the court remains contested. To some authors, the constraint hardly binds (Stone Sweet and Brunell, 2012; Davies, 2014). To others, the risk of override makes cautious judges sensitive to information suggestive of potential override, causing them to be careful not to rule in a manner that may be interpreted by governments as going beyond the legal merit (Larsson and Naurin, 2016). As such, it is not the override, but the risk of override that causes judges to behave in a way that makes override less likely, in particular in the months leading up to major treaty revisions (Castro-Monero et al., 2018).
Fourth, rulings in the CJEU must be followed through by other political actors. Different national political and institutional settings affect the way in which national courts respond to EU (e.g. Golub, 1996; Mattli and Slaughter, 1998a, 1998b; Alter, 2000, 2001). There are different levels of public support for European integration, awareness of the CJEU, satisfaction with the CJEU and general satisfaction with courts and judges (Caldeira and Gibson, 1995; Gibson and Caldeira, 1995, 1998). If courts ignore these mass sentiments, they risk provoking parliamentary challenges to their judicial autonomy and undermining public acceptance of courts and the judicial system. In addition, each national system has a different structure of legal institutions, court procedures, powers of judicial review, cost of access for litigants and legal training of judges (Alter, 2000). When following up rulings in the CJEU. both national actors and the Commission have some leeway. For example, Peritz (2018) demonstrates that the effect of trade-liberalization rulings on intra-European trade is highly contingent on the actions of domestic actors to obstruct of actively enforce the rulings. Finally, Martinsen (2015a, 2015b) shows that the extent to which the rulings of the CJEU are codified through subsequent legislation varies with the support that the consequences of the ruling enjoy amongst the EU legislators and the member states. This is particularly evident in the area of social policy (see also Hatzopoulos and Hervey, 2013).
Nevertheless, this supranationalist explanation of the CJEU suffers from some of the same weaknesses as the general supranational politics approach. In particular, this approach overemphasizes the autonomy of supranational institutions and transnational interests in the promotion of EU legal integration. These scholars argue that once transnational activities and supranational institutions have been unleashed, there is little that national governments can do to stop them (Pierson, 1996; Blauberger and Schmidt, 2017). However, national governments are the signatories of the treaties, and if provoked they can restrict the powers of the CJEU and redefine the nature of the EU judicial system.
The Political System of the European Union
Conclusion: A European Constitution?
The EU has a legal-constitutional framework that contains two of the basic doctrines of a federal legal system: the direct effect of EU law on individual citizens throughout the EU, and the supremacy of EU law over domestic law. Also, in the CJEU the EU has a powerful constitutional and administrative body to oversee the implementation of EU law and keep the EU institutions in check.
How this came about is a matter of contention. Courts have more discretion under certain institutional designs than others. The CJEU has substantial room for manoeuvre because there is only a small probability that the EU treaty will be reformed to reduce the CJEU's powers or that new legislation will be passed to overturn one of its decisions. Because there are many veto players in the EU system, at least one member state, the Commission, the European Parliament or a group of powerful transnational economic actors is likely to be able to block a reduction of the CJEU's powers or the overturning of one of its decisions.
However, the CJEU has imperfect information on how other actors will react to its decisions. Governments have shorter time horizons than courts because they face general elections every few years. This means that they are less interested in the long-term implications of delegating powers to the CJEU than in the immediate political salience of a decision. But. imperfect information also means that the CJEU is uncertain about what issues will become politically salient in which member state.
This judicial politics game has produced an incomplete constitution. The precise constitutional architecture is not fixed, particularly following the rejection of the draft Constitutional Treaty. In a sense, as Weiler (1993) argues, where the fundamental constitutional nature of the EU is concerned, the EU still has an 'unknown destination' (cf. Shonfield, 1973).
However, the current quasi-constitutional architecture of the EU is in a sense a relatively stable equilibrium: a balance between the discretion of the CJEU/ national courts on the one hand, and the conscious decisions by national governments to construct a rule of law to enable economic integration on the other. This goes hand in hand with the emerging equilibrium in the vertical allocation of competences (discussed in Chapter 1). Put this way. the constitutional settlement relating to the allocation of market regulation competences to the European level relies on a stable structure for the enforcement of contracts in these policy areas.
Nevertheless, this equilibrium could be upset by changes in public opinion, party competition and ideology, interest group politics and so on, which could push the EU towards a full federal constitutional arrangement or even result in a constitutional step backwards. It is to the political context of institutional politics that we turn in Part II of this book.