1 Introduction
Among international courts, the Court of Justice of the European Union (CJEU) is generally regarded as unusually successful – a model upon which many other international courts have been based (Börzel and Risse, Reference Börzel and Risse2012). Its innovative and controversial advancements in law have encountered relatively little overt resistance by national policy-makers (Kelemen, Reference Kelemen2016). Many of its innovations have been codified in subsequent EU legislation or incorporated into the EU treaty framework. Instances where the EU legislature has explicitly reversed undesired judgments are rare (Davies, Reference Davies2014; Larsson, Reference Larsson2016; Martinsen, Reference Martinsen2015). Nonetheless, there is an impression among scholars and observers that the CJEU has in recent years ventured into more controversial terrain. In recent judgments, the redistributive social consequences of earlier doctrines, taken to their logical ends, have become clearer. Permissive judgments on the access of EU migrants to social benefits have raised concerns about ‘welfare migration’ (Blauberger and Schmidt, Reference Blauberger and Schmidt2014), which are played up in national debates, such as during the referendum on UK membership. The judgments in the so-called ‘Laval quartet’ have alerted social democrats and trade unions to the possible detrimental consequences of EU free-movement principles for national collective labour law (Joerges and Rödl, Reference Joerges and Rödl2009; Rönnmar, Reference Rönnmar2008). Academic debate about the CJEU has become more heated, with some prominent contributors calling for non-compliance with controversial rulings and curbs on the court's powers (Herzog and Gerken, Reference Herzog and Gerken2008; Scharpf, Reference Scharpf2008). While such demands have never seriously been considered by policy-makers, there is an impression that the CJEU, too, is experiencing increasing resistance (Kelemen, Reference Kelemen2016).
In this issue, Madsen, Cebulak and Wiebusch (Reference Madsen, Cebulak and Wiebusch2018) introduce a nuanced classification of resistance to international courts. They disaggregate the concept of resistance into ordinary forms of criticism, which are labelled pushback and extraordinary forms of criticism, which are labelled backlash. Pushback occurs within existing institutional structures, whereas backlash aims at fundamentally altering such structures. This contribution takes an empirical look at the reality of such resistance against the CJEU. Its focus lies on ‘commonplace’ resistance, as it has existed throughout the history of the EU, not the recent extraordinary circumstances in some EU Member States, particularly in Poland and Hungary. These are covered in more detail elsewhere (Blauberger and Kelemen, Reference Blauberger and Kelemen2017; Kelemen, Reference Kelemen2016). Debates about controversial CJEU judgments also played a role during the UK referendum on Brexit,Footnote 1 but the outcome of that referendum should be read as a backlash against the process of European integration more generally, or even against elites in a very broad sense. The CJEU hardly topped the list of villains in the ‘Leave’ camp. Some – perhaps even more – of the ire was directed at the European Court of Human Rights, and a clear distinction was not always made. It is worth noting, however, that the exit from the jurisdiction of the CJEU has since become one of the ‘red lines’ of the Brexit negotiators (Armstrong, Reference Armstrong2017, p. 113).
This contribution demonstrates that, while incidences of backlash have been rare, pushback against CJEU judgments has been more widespread than the standard account of the court's historic successes has admitted. Strategies of non-compliance and evasion, by both national authorities and national judiciaries, have often put a check on the wider ramifications of the more controversial judgments of the CJEU. This, coupled with an apparent ‘retrenchment’ from expansive interpretations of EU law on the part of the CJEU (Blauberger et al., Reference Blaubergerforthcoming; Spaventa, Reference Spaventa and Kochenov2017), will likely continue to prevent backlash by Member State authorities – at least as far as they remain committed to the rule of law.
The paper proceeds as follows. It fist examines the different forms that resistance against the CJEU may take. Subsequent sections separately review data – so far as this is available – on such resistance. They review literature on compliance with CJEU rulings by political authorities and national judiciaries. Beyond non-compliance, the paper looks at strategies Member State authorities employ to protect national policies from the wider ramifications of CJEU doctrines. A final section summarises the central insights.
2 Forms of resistance against the CJEU
In this issue, Madsen, Cebulak and Wiebusch (Reference Madsen, Cebulak and Wiebusch2018) define resistance to international courts as an attempt at blocking or reversing advancements in law triggered by those courts. At the core of the project of European integration, and thus of EU law, lies the vision of an integrated European market where labour, goods, services and capital move freely across borders. While the project's ambition has always reached beyond the economic sphere, the bulk of EU law continues to concern such free movement (Barnard, Reference Barnard2016). It is only more recently that the free movement of persons, embodied in the concept of Union citizenship, has begun to dissociate itself from economic activity, and the EU Charter of Fundamental Rights has only recently introduced an explicit fundamental rights catalogue into EU law. In pursuit of its aims, EU law replaces specific national provisions, based on either (‘positive’) reregulation by means of EU legislation or the mere (‘negative’) removal of national barriers to trade. The CJEU is the ultimate arbiter of conflicts between European rules and national rules, and it is generally regarded as favouring European rules over national ones (Larsson and Naurin, Reference Larsson and Naurin2016; Pollack, Reference Pollack, Grossmann, Cohen, Follesdal and Ulfstein2018). Resistance against the CJEU therefore generally takes the form of opposition to the replacement of national rules. Reasons for such resistance can range from narrow self-interest in the distributional outcomes of national rules to more general concerns for the integrity of legal traditions, economic institutions or national identities.
As in all international legal orders, the implementation of EU law essentially relies on the willingness of its subjects to comply, which may be questionable for the stated reasons. Compared with other international legal orders, however, the EU has a rather sophisticated set of mechanisms in place by which compliance with EU law can be enforced. National authorities have a legal obligation to implement EU law and adjust administrative practices accordingly (Article 4(3) TEU). Where they fail to do so, they can be taken to court, either by the European Commission or by affected citizens and companies. Where infringements of EU law are prosecuted by the Commission, Member States potentially face financial sanctions for continued non-compliance, or even immediate sanctions where they fail to timely transpose EU directives into national law (Article 260 TFEU). Member States can also be held financially liable for the costs incurred by citizens and companies on account of a failure to correctly implement or apply EU law. While EU institutions, in particular the Commission, show an interest in not escalating conflicts and refrain from ‘overusing’ certain enforcement tools (Falkner, Reference Falkner2015), citizens and companies frequently invoke EU law in cases before national courts, either against national authorities or against other citizens and companies. The judicial enforcement of EU law is therefore an everyday practice. There is no disagreement among Member States about the principles of this enforcement regime. In fact, they have been strengthened with almost every treaty change in the past (Kelemen, Reference Kelemen2016).
Central to this enforcement regime is the role of the CJEU as the last instance for the adjudication of conflicts over the proper interpretation and application of EU law. Again, there is no conflict among Member States about this role in principle. That, of course, does not mean that all of its rulings are automatically accepted. This is especially true where the CJEU develops innovative doctrines, sometimes in clear opposition to the intent of the EU legislature (Dawson et al., Reference Dawson, De Witte and Muir2013; de Waele, Reference De Waele2010). Examples for such types of judgments include the creation of an obligation to recognise other Member States’ product regulations, the CJEU's very permissive definition of what constitutes a ‘worker’ who would enjoy unrestricted free-movement rights, its (initially) generous interpretation of migrant citizens’ access to host state welfare benefits, its very restrictive interpretation of the national labour market rules that can be applied to posted workers or its development of the concept of age discrimination in labour market rules.
The occurrence of resistance is therefore unsurprising. Resistance can take many forms and come from many sources. No longer ‘blessed with benign neglect’ (Stein, Reference Stein1981, p. 1), the CJEU today faces a broad audience for all of its judgments (Kelemen, Reference Kelemen2016). The most concentrated source of engaged criticism is the large audience of academic lawyers and legal professionals who analyse legal developments in a multitude of specialised fora. For this audience, pushback is an almost daily exercise. Academic debate is also the dominant source of strong criticism that could be classified as backlash (e.g. Herzog and Gerken, Reference Herzog and Gerken2008; Scharpf, Reference Scharpf2008). It is rarer for CJEU judgments to catch the attention of a mass audience. Media only rarely report on judgments, and even quality outlets sometimes fail to properly differentiate between the CJEU and the European Court of Human Rights. Public opinion of the CJEU varies with the public perception of the EU in general, but opinion surveys often show higher approval ratings for the CJEU than most other public institutions (Kelemen, Reference Kelemen2016, p. 132; Pollack, Reference Pollack, Grossmann, Cohen, Follesdal and Ulfstein2018).
While backlash has occasionally emanated from the academic audience, the media and the wider public, such efforts can only translate into concrete effects if it they are taken up by political or judicial authorities. National authorities can have strong incentives to oppose judgments if they disagree with their policy implications or if a ruling would result in extensive costs. National judiciaries can be faced with contradicting demands from EU law and national law and express strong preferences for the integrity of national legal traditions. So far, the CJEU has experienced little backlash, interpreted as extraordinary criticism that challenges the EU legal system as such, from either national governments or national judiciaries. The reversal of unwanted CJEU rulings by ordinary means, either through legislation or treaty change (often termed ‘override’), is also exceedingly rare (Davies, Reference Davies2014; Larsson, Reference Larsson2016; Martinsen, Reference Martinsen2015). However, the fact that the CJEU's role in principle has gone largely uncontested and its rulings are rarely reversed does not mean that it is not resisted. Pushback from core audiences is more common than the traditional ‘success story’ of the CJEU implies. The remainder of this paper will focus on forms of such pushback by national political authorities and the judiciary.
During procedures before the CJEU, Member States have the opportunity to submit so-called ‘observations’, outlining their standpoint in the legal issue at hand. Such observations have been interpreted as signals to the CJEU of the limits of an ‘acceptable’ ruling and as implicit threats of non-compliance or override. Recent research has highlighted that the CJEU is receptive to such signals on the whole (Carrubba et al., Reference Carrubba, Gabel and Hankla2012; Larsson and Naurin, Reference Larsson and Naurin2016). National governments therefore have an opportunity to pre-empt unwanted CJEU doctrine. However, it is not rare for the CJEU to rule against even vocal Member State opposition (Stone Sweet and Brunell, Reference Stone Sweet and Brunell2012). Once the CJEU has issued a contested judgment, many different political and legal institutions mediate its application, and pushback can take multiple forms. In particular, a distinction should be made between the failure by a national authority to comply with an individual judgment and attempts by Member State authorities to limit or nullify the broader impact of judgments on national policies. This follows the distinction used in the literature between relatively narrow ‘compliance’ with the outcomes of international adjudication and the broader ‘effectiveness’ of international courts (Helfer, Reference Helfer, Romano, Alter and Avgerou2013). The former, so far as it is deliberate, is a relatively explicit form of pushback. It exhibits some characteristics of backlash in that non-compliance with an individual ruling is outside the bounds of ‘ordinary’ criticism. At the same time, it does not explicitly challenge the EU legal system as such. Such non-compliance is not uncommon. The latter poses the question of the practical, ‘real-world’ effects of CJEU doctrines on national policies – a point that has been a curious blind spot in the extensive literature on the CJEU so far (Blauberger and Schmidt, Reference Blauberger and Schmidt2017b, p. 7). National authorities have various means of ‘containing’ the intermediate authority of the CJEU, limiting the effects of its rulings to the individual case (Conant, Reference Conant2002). By ignoring a judgment's wider ramifications (its ‘erga omnes’ effect), national policy-makers can force individuals to claim their EU rights again and again – a cumbersome strategy that only few can afford.
The following sections concentrate on resistance to CJEU judgments by national political and judicial authorities and follow the distinction between compliance and effectiveness. They review data on the incidence of the outlined pushback strategies.
3 Pushback through non-compliance with CJEU rulings
This section reviews the available data on the extent to which national political and judicial authorities comply with CJEU rulings. The original treaties that founded the European Economic Community did not contain a special procedure by which to enforce a court ruling. A CJEU ruling was treated as just another obligation arising out of EU law that would be enforced like any other source of EU law. In the late 1980s, concerns about non-compliance with CJEU rulings led the Commission and some Member States (not least the UK) to advocate for a more forceful enforcement mechanism (Kilbey, Reference Kilbey2010). The Maastricht Treaty, signed in 1992, introduced the possibility to apply penalties for non-compliance with a CJEU ruling. This mechanism, which was strengthened by the 2009 Lisbon Treaty, only relates to one type of CJEU ruling: those delivered as a consequence of an infringement procedure brought by the Commission against an EU Member State for failure to fulfil an obligation. Where the CJEU has issued a ruling against a Member State in such a procedure, the Commission can now bring an expedited second infringement procedure in case the ruling is not complied with, at the end of which the CJEU can apply a lump-sum fine and a penalty that recurs until the initial ruling is complied with.
Infringement cases, however, constitute only a minority of all cases handled by the CJEU (e.g. Court of Justice of the European Union, 2017, p. 88). The majority of the Court's docket is made up of preliminary references sent by national courts in cases initiated by private parties. The ability of citizens, groups and companies to turn to domestic courts to claim rights based in EU law is a feature that sets the EU legal order apart from most international legal regimes (Tallberg, Reference Tallberg2002). No special enforcement mechanism exists for CJEU judgments issued under the preliminary reference procedure but, since such rulings are sent back to the referring (national) court for a decision on the merits, the procedure activates domestic mechanisms of compliance with judicial rulings, which are generally presumed to be more forceful than international mechanisms. National courts also often apply EU law directly without a preliminary reference to the CJEU. Issues of pushback through non-compliance in this context primarily concern the extent to which national courts are willing to follow the CJEU's interpretation of EU law. Compliance mechanisms differ quite significantly between the types of cases the CJEU rules on. Whereas, in infringement proceedings, it is the behaviour of national political authorities that is most relevant for compliance, in the preliminary reference procedure, it is mostly national judiciaries. The paper will deal with these separately.
3.1 Political authorities’ compliance with CJEU judgments
Research on Member State compliance with CJEU judgments suffers from the same problems as compliance research more broadly, which is that non-compliance is difficult to detect and that official accounts may not reflect the real state of compliance (Hartlapp and Falkner, Reference Hartlapp and Falkner2009). Since its first ‘Annual Report on Monitoring the Application of EU Law’ in 1984, the Commission has regularly published quite detailed information on how it follows up on CJEU rulings in infringement cases. Whether or not the Commission records are a reliable indicator for overall compliance patterns has been subject to some debate (Börzel et al., Reference Börzel2010; Smith, Reference Smith, Drake and Smith2016). The primary objections are that the Commission has limited capacity to follow up on implementation and often relies on national ‘whistle-blowers’, which might not be evenly distributed, and that the Commission pursues its own agenda when choosing which cases to pursue (Hartlapp and Falkner, Reference Hartlapp and Falkner2009). CJEU case-law has given the Commission virtually unlimited leeway to take up or drop cases as it sees fit.
In any case, even if Commission data underreport actual incidences of non-compliance, its annual reports (and more recently its online database on infringement decisionsFootnote 2) indicate that compliance with CJEU rulings in infringement procedures is not a given. From 1983 to 1991, the Commission referred thirty-four cases to the CJEU for national authorities’ failure to implement a judgment,Footnote 3 often several years after the original judgment was published. This corresponds to about 12 per cent of all infringement judgments in that period (285). Before the introduction of penalties for non-compliance with CJEU rulings in infringement cases, the second judgment had the same status as the first, and the Commission has reported multiple instances where even a second judgment did not bring about compliance.Footnote 4 In only one instance did the Commission refer a case for the third time.Footnote 5
After the Maastricht Treaty introduced the possibility for penalties for non-compliance with CJEU rulings, the Commission was initially reluctant to pursue cases that would result in penalties. The true state of compliance with CJEU rulings for this period is therefore difficult to establish (Falkner, Reference Falkner2015). Only after 1997 did the Commission start referring cases of non-compliance with previous rulings to the CJEU with requests for financial penalties. Until the end of 2016, the Commission had sent eighty-six such cases to the Court, which corresponds to about 9 per cent of infringement cases ruled on by the CJEU from 1997 to 2016 (962).Footnote 6 The CJEU ruled on thirty-three of these (the rest were withdrawn before a judgment), issuing financial penalties in thirty-one.
More detailed information on the Commission's follow-up of CJEU judgments in infringement cases has been available since 2003 in the Commission's online database on infringement decisions. According to this source, the Commission has officially initiated ‘second round infringement proceedings’ (infringement proceedings regarding a previous CJEU judgment) with a letter of formal notice in 461 cases between 2003 and 2016, indicating that compliance was at least delayed in just under half of all cases ruled on by the Court in this period. This number corroborates earlier findings by Daniela Panke, who found that compliance was either absent or incomplete within two years of a CJEU judgment in about 50 per cent of German cases in the field of social and environmental policy from 1978 to 2000 (Reference Panke2007, p. 851).Footnote 7
Non-compliance with CJEU rulings varies by policy areas. Initial referrals to the CJEU in the period covered by the Commission's infringement decision database (2003–2016) were handled mostly by the Directorate Generals responsible for the environment (22 per cent), for internal market and services (18 per cent) and for mobility and transport (13 per cent). Second-round referrals see a shift in emphasis. An even higher share concern environmental cases (32 per cent). The share of internal market cases is about the same as in first-round cases (22 per cent), but a disproportionately large set of second-round cases relate to competition policy (15 per cent). While the difficulty remains of separating out Commission enforcement priorities from actual instances of non-compliance, the Commission's data indicate that CJEU judgments relating to the environment and to competition policy encounter particular implementation problems. An analysis by Gerda Falkner and her research team of all cases in which the Commission has asked the CJEU to issue a penalty for non-compliance with a previous ruling comes to the conclusion that the primary reason why Member States fail to implement CJEU rulings is the desire by national authorities to protect important constituencies and to avoid implementation measures that are particularly costly – much more so than administrative capacity problems or genuine interpretation issues (Falkner, Reference Falkner2014; Reference Falkner2015). Cases concerning environmental protection and state aid to industry clearly fall into this category.
Closer analyses of the aftermath of the cases in which the CJEU issued a penalty for non-compliance with previous judgments indicate that even financial penalties do not guarantee that the underlying implementation problem is remedied. Ian Kilbey shows that the Commission has developed creative means of ‘face-saving’ in order to close cases after some penalties have been paid and some efforts towards compliance undertaken, even though the problem persists (see also Batory, Reference Batory2016; Kilbey, Reference Kilbey2010). Gerda Falkner (Reference Falkner2015) comes to similar conclusions in her study referenced above. These cases clearly show the limits of EU institutions’ ability to enforce CJEU rulings. It is reasonable to assume that the Commission has an interest in minimising the amount of cases in which it gets ‘shown up’ in this fashion, which would suggest that ‘hopeless’ cases may not get pursued to the full extent. This is more reason to assume that Commission enforcement data likely underreport true instances of non-compliance.
Even taken at face value, these data show that compliance with CJEU rulings in infringement cases is at least delayed in about half of all cases and more seriously resisted in about one out of ten cases. Optimists will interpret this as a respectable compliance rate of 90 per cent, but it seems doubtful that these numbers compare favourably with those of national supreme courts with which the CJEU is often equated (Rogowski and Gawron, Reference Rogowski and Gawron2002).
3.2 National judiciaries’ compliance with CJEU judgments
Whereas there are at least some official data available on Member State compliance with rulings in infringement proceedings, research on compliance with CJEU rulings on preliminary references on the other hand is entirely reliant on case-by-case analysis. Preliminary references arise out of domestic conflicts where one of the parties (or the national judges themselves) find the case's resolution to hinge on a question of EU law. National courts can decide to refer this question to the CJEU for a preliminary ruling. The CJEU rules on the validity of the piece of EU law in question or gives an authoritative interpretation of this law, and refers the case back to the national court. Subsequently, at least in principle, the national court applies the interpretation of EU law given by the CJEU to the facts of the case at hand.
Domestic cases that involve a question of EU law typically either revolve around the validity of a piece of EU law or around the compatibility of a national rule or practice with EU law obligations. The first kind of case amounts to a form of judicial review of EU legislation. The second kind in turn is a form of judicial review of national law or administrative practices. Such cases can amount to a form of de-centralised infringement proceeding brought by private parties in national courts explicitly or implicitly against a national authority (de Witte, Reference De Witte, de Witte, Mayoral, Jaremba, Wind and Podstawa2016; Hübner, Reference Hübner2017).Footnote 8 The central difference between such de-centralised cases and an infringement procedure initiated by the Commission is that the final outcome in the former is a judgment by a national court. Should the national court find that, in light of the CJEU's case-law, a national rule or practice has to be disapplied (because it conflicts with EU law), national authorities are faced with an authoritative ruling by a national court and domestic rules of compliance and enforcement apply (Alter, Reference Alter1996). The assumption is that the ‘compliance pressure’ on national authorities is greater if the final judgment is issued by a national court rather than an international one. It is this involvement of national judges in the enforcement of EU law that sets the implementation of CJEU judgments apart from its international peers. The ‘complicity’ of national courts is generally seen as a cornerstone of the success of the EU legal regime compared with other international law regimes (Alter, Reference Alter1996).
However, a necessary condition for this system to work is that national courts actually follow the CJEU's interpretations. Such compliance cannot be taken for granted. There are two ways in which pushback against CJEU doctrines can emanate from national courts: one where the national court makes a reference but decides not to follow the interpretation of the CJEU and another where the national court chooses not to refer in the first place. Lower national courts are under no obligation to refer a question to the CJEU and parties to the case have no means to force a reference (Bernitz, Reference Bernitz2016). National courts of last instance are technically under an obligation to refer cases involving questions of EU law but, in practice, they can and do find ways to avoid this (Golub, Reference Golub1996, p. 362; Piqani, Reference Piqani, de Witte, Mayoral, Jaremba, Wind and Podstawa2016). The ‘acte claire’ doctrine, established by the CJEU itself to avoid trivial references, allows national courts to apply EU law on their own where its meaning is sufficiently clear and national courts, particularly courts of last instance, use this leeway extensively (Bernitz, Reference Bernitz2016). Other relatively open norms in EU law, such as proportionality in the assessment of barriers to free movement or the ‘sufficient seriousness’ of a breach of EU law in the assessment of state liability, also leave extensive space for interpretation that national courts can use creatively (Davies, Reference Davies2012).
Given the centrality of national courts for theories of European legal integration, there is rather little systematic and comparative empirical research done on this question. Early research on individual Member States, based on case-by-case analyses, indicated that courts comply at about the same rate as Member State authorities in infringement proceedings (Korte, Reference Korte1991; Schwarze, Reference Schwarze1988). Newer research on the preliminary reference procedure has presented a more detailed picture of how national courts can avoid complying with CJEU rulings they disagree with outside of open defiance, even where they have sent a reference to the CJEU (Davies, Reference Davies2012; Nyikos, Reference Nyikos2003). One such strategy is to refer the question again, another to interpret the facts of the case so that the CJEU's interpretation of EU law does not apply and yet another to silently apply a different interpretation without overtly stating disagreement. In her study of 300 preliminary reference cases from 1961 to 1994, Stacy Nyikos found evidence for all three strategies, although they remain exceptions; 289 CJEU judgments, or 96 per cent of her sample, were implemented.
National courts can use the dossier they refer to the CJEU to offer their point of view on how their specific question should be answered. Rather than a one-way street, where national courts simply pose a question and patiently await an answer, the preliminary reference procedure can in this way become a vehicle for strategic interaction between courts. This option is in fact frequently used and ‘pre-emptive opinions’, as they have been called (Nyikos, Reference Nyikos2006), similar to Member State observations, can act as a signal to the CJEU as to what an acceptable ruling could look like. A case in point is the recent (and first ever) case referred to the CJEU by the German Constitutional Court.Footnote 9 In phrasing its question about the proper interpretation of the European Central Bank (ECB)’s competence to purchase national government bonds on the secondary market (so-called ‘Outright Monetary Transactions’), the German court appeared to suggest that it would accept no other interpretation than the restrictive one it had suggested as part of its dossier (Jones and Kelemen, Reference Jones and Kelemen2014). In a study of 574 cases from 1961 to 1994, Stacy Nyikos found that national courts make use of pre-emptory opinions in 41 per cent of the cases. Karin Leijon found such opinions in half of her 359 randomly selected cases in the period from 1992 to 2012 (Leijon, Reference Leijon2015).
National judiciaries’ pushback against CJEU rulings is much more difficult to assess in cases where national courts apply (or choose not to apply) EU law without a reference to the CJEU. Such cases make up the bulk of all cases in which a point of EU law is at issue (Conant, Reference Conant2002; Davies, Reference Davies2012), but reliable numbers are largely lacking. Denise Hübner's recent study of national court cases involving an interpretation of EU directives, one of the few such attempts, finds that no reference was made in about 80 per cent of the 1,310 cases in her sample (Hübner, Reference Hübner2017). A viable pushback strategy for national courts is therefore to apply their own interpretation of EU law or to ignore EU law altogether. Reasons for national courts to do so are multiple. It could be hostility towards judicial review of legislation more generally, which has been attested to judges in majoritarian democracies, such as the UK and Scandinavia (Wind et al., Reference Wind, Martinsen and Rotger2009). Marlene Wind has moreover reported that Danish government officials discourage Danish judges from sending references to the CJEU (Wind, Reference Wind2010, p. 1051). It could also be judges’ personal preferences, either for the integrity of national legal systems or for individual policies, such as residence rights or welfare provisions (Davies, Reference Davies2012, pp. 86–87; Golub, Reference Golub1996; Leijon and Karlsson, Reference Leijon and Karlsson2013; Wind, Reference Wind2010, pp. 1052–1053) or more mundane reasons such as lack of knowledge (Mayoral et al., Reference Mayoral, Jaremba and Nowak2014) or managing case-load (Jaremba, Reference Jaremba, de Witte, Mayoral, Jaremba, Wind and Podstawa2016).
National judges’ historic resistance to accept innovative doctrines, first and foremost EU law's primacy and direct effect, is well documented (Alter, Reference Alter1996; Slaughter et al., Reference Slaughter, Stone Sweet and Weiler1998). Overt pushback against CJEU judgments is rare, but the few instances in recent years have been registered with some alarm (Kelemen, Reference Kelemen2016). In January 2012, the Czech Constitutional Court decided not to apply a judgment of the CJEU on the interpretation of EU rules regarding the calculation of old-age pensions for workers from the former Czechoslovakia. The original case had been referred by the Czech Supreme Administrative Court and the CJEU had issued its judgment in June 2011.Footnote 10 The Czech Constitutional Court, to which the case was subsequently appealed, held that the calculation of pension rights relating to the former Czechoslovakia was not a cross-border issue, that the CJEU was wrong to declare itself competent to address this issue at all and that the CJEU had therefore overstepped its boundaries. This was interpreted as the very first incidence in which a national high court explicitly defied a ruling by the CJEU (Dyevre, Reference Dyevre2016, pp. 107–108). The perceived mood of defiance by high courts seemed to be confirmed by the German Constitutional Court's barbed reference regarding the ECB's bond buying programme, which was described above (Dyevre, Reference Dyevre2016, p. 109). In this case, the CJEU disagreed with the German court's interpretation and accorded the ECB broad discretion to conduct monetary policy. There was some initial doubt whether the German Constitutional Court would take a confrontational stance and dismiss the CJEU's ruling, but its 2016 judgment largely followed the CJEU's line. In August 2017, the German Constitutional Court again referred a question about the ECB's competences in monetary policy, this time regarding large-scale purchases of assets, as part of its strategy of ‘quantitative easing’. Again, the German court has expressed doubts about the legality of this strategy, although its tone this time around has been interpreted as being more open to judicial dialogue (Goldmann, Reference Goldmann2017). The Supreme Court of Denmark, too, has recently decided not to follow a CJEU ruling. This case is noteworthy because, unlike in the Czech case, it was the Supreme Court itself that sent the original reference.Footnote 11 While the facts of the case were unremarkable, the Supreme Court decided to interpret the applicability of CJEU doctrines in Denmark through the lens of the Danish Accession Act (Madsen et al., Reference Madsen, Olsen and Sadl2017). It held that all judicial innovations that took place after the most recent amendments to the Accession Act would not be legally binding on Danish courts. The doctrine that the CJEU drew on to suggest the incompatibility of the Danish law in question, the prohibition of discrimination on the grounds of age, had been developed by the CJEU in a 2005 rulingFootnote 12 and was handed down after the latest amendment to the Accession Act. The Supreme Court, in direct contradiction to the CJEU's ruling on its own reference, declared that it would overstep its constitutional boundaries were it to disapply national law on account of a novel CJEU doctrine (Madsen et al., Reference Madsen, Olsen and Sadl2017).
It is as yet unclear what consequences these cases of explicit pushback against the CJEU by national high courts will have for the wider relationship between the EU court and national judiciaries, but they are indicative of a growing willingness on the part of national judges to draw red lines. Such overt pushback takes place in the public sphere and is generally subject to intense debate, at least among a specialist audience. In contrast, empirical work on continuing patterns of covert pushback, for instance in cases where national courts avoid references, is largely lacking.
Technically, the failure of a national court to apply the CJEU's interpretation of EU law can be interpreted as an infringement of EU law, for which the Commission could initiate infringement proceedings. The Commission, however, has been very cautious in using the infringement procedure against national judiciaries, not least since such procedures are always officially directed at Member State governments, who in turn have little control over the judiciary (Andersen, Reference Andersen2012, pp. 61–65; Taborowski, Reference Taborowski2012, pp. 1902–1911). Only very few cases have reached the CJEU in which the Commission found a national judiciary's decisions to infringe EU law,Footnote 13 but these cases were not based on a misapplication of previous CJEU case-law. The Commission carefully directed its challenges to the failure of governments and legislatures to remedy underlying legislation, not to the court judgment in question. Only in one case has the Commission directly addressed what it found to be a legal error by a national court, but this case was never referred to the CJEU.Footnote 14 Similarly, ‘judicial error’ in interpreting EU law has been the subject of several preliminary references to the CJEU.Footnote 15 The CJEU used these cases to establish in principle the possibility for individuals to seek damages for the ‘faulty’ application of EU law by a national court, including the failure by a court of last instance to send a preliminary reference, albeit only under very strict conditions (Leczykiewicz, Reference Leczykiewicz, Chalmers and Arnull2015, pp. 230–231). However, none of these cases so far concerned the application by national judiciaries of previous CJEU case-law.
The Commission's only direct challenge to a national judiciary's interaction with the CJEU to date was an infringement procedure launched against Sweden in 2003 on account of the Swedish Supreme Court's low number of references and its practice of not giving reasons for a refusal to send a reference. The Commission terminated the procedure in 2006 after the Swedish parliament amended procedural rules requiring courts of last instance to justify their decision not to make a referral (Bernitz, Reference Bernitz2016).
In sum, studies confirm the notion that national judiciaries’ compliance with CJEU case-law is uneven. Since systematic comparisons are lacking, little is known about potential cross-national patterns of resistance. Where national courts refer questions to the CJEU, they tend to comply with the resulting judgment, although here, too, compliance is not complete. Earlier studies reported non-compliance in the 10 per cent range, whereas the only recent detailed study detects non-compliance in 4 per cent of cases. National judges’ compliance with CJEU doctrine is much more difficult to trace when courts choose not to refer and instead to apply EU law themselves. We know that this happens in the vast majority of cases where a point of EU law is at issue. It is even more difficult to systematically trace resistance in cases where national courts summarily ignore EU law. Whereas overt conflict between national high courts and the CJEU is amply documented, patterns of less evident pushback, in particular by lower courts, receive very little attention. Since lower national courts are the ordinary courts of EU law, and the odds of any individual conflict reaching a court of last instance are low, this leaves vast holes in the understanding of the everyday administration of justice in subject matters covered by EU law.
4 Pushback through containment
While non-compliance with CJEU judgments is a fairly strong form of pushback, national political authorities often have the possibility to contain wider ramifications of judge-made doctrines even when they comply with a CJEU judgment in a narrow sense. While this form of resistance sends less obvious signals to the EU court, the national room for manoeuvre in neutralising controversial judgments may be the central mechanism that has so far prevented full-blown backlash against the CJEU (Werner, Reference Werner2016). Member State authorities have frequently found means to limit the overall effectiveness (Helfer, Reference Helfer, Romano, Alter and Avgerou2013) of CJEU decisions. Rather than merely tracing developments in doctrine and recording compliance, it is therefore important to ask how far CJEU rulings have practical effects ‘on the ground’. Do citizens or companies have greater opportunities after a ruling than before? Research on the practical effects of court rulings has deeper roots in the US, where initial optimism about the Supreme Court's impact on civil rights turned into scepticism about the ability of courts to bring about wider social change (Rosenberg, Reference Rosenberg1991; Scheingold, Reference Scheingold1974). The reference point for similar studies in the EU is Lisa Conant's Reference Conant2002 study on the strategies by which national authorities manage to limit the impact of CJEU judgments on policy outcomes (Conant, Reference Conant2002). Conant, like Rosenberg, argued that, in the absence of supportive political pressure following the development of innovative legal doctrines, national authorities will respond by isolating the effects of single judgments, applying them only to the case at hand while ignoring their wider ramifications – that is, denying them intermediate authority and ‘erga omnes’ effect, forcing individuals to claim their rights in court every time these rights are infringed. Conant regarded this type of response as a baseline: ‘In the absence of mobilization by organizational or institutional actors, governments can usually pursue a strategy of contained compliance’ (2002, p. 38).
Despite the extremely large body of work on innovative CJEU doctrines by both academic lawyers and political scientists, work similar to Conant's has been rare until quite recently. It is only in the last few years that the question of the wider effectiveness of CJEU judgments has received renewed attention (Freedland and Prassl, Reference Freedland, Prassl, Freedland and Prassl2015a; Treib, Reference Treib2014). Empirical studies cluster most heavily around the controversial CJEU doctrine developed in the so-called ‘Laval quartet’Footnote 16 (Freedland and Prassl, Reference Freedland and Prassl2015b) and the field of social policy, with a particular emphasis on access to welfare benefits (Blauberger and Schmidt, Reference Blauberger and Schmidt2014; Heindlmaier and Blauberger, Reference Heindlmaier and Blauberger2017) and cross-border health care (Martinsen, Reference Martinsen2005; Martinsen and Mayoral, Reference Martinsen and Mayoral2017; Obermaier, Reference Obermaier2008). In all of these issues, the CJEU intruded on what Ran Hirschl has called ‘mega-politics’ (Hirschl, Reference Hirschl2008), addressing deeply entrenched national conceptions of labour market and welfare regimes, and has met with pushback from an audience beyond specialists. At the same time, only a small section of the literature devoted to these doctrines has asked what their ‘real-world’ effects happen to be. This section will review the main finding of this literature.
The CJEU's judgments in the four cases that have collectively become known as the ‘Laval quartet’, handed down between 2007 and 2008, created one of the most intense controversies the CJEU has faced in recent years. The cases Viking and Laval both concerned trade union industrial action against companies that either sought to relocate to a country with lower labour costs or that posted workers from low-wage countries to higher-wage countries while refusing to sign a collective agreement with host-country unions. The CJEU found that trade union's right to strike, while in principle a fundamental right in EU law, was not unconditional but needed to be balanced against EU free-movement principles, in particular the freedom of establishment and the freedom to provide services. In Laval, in particular, the CJEU left the referring Swedish court little leeway but to declare the strike in question to be disproportionate and therefore illegal – a far-reaching step, given that the right to strike in Sweden had previously not been subject to proportionality tests. The cases Rüffert and Commission v. Luxembourg both concerned the extent to which national authorities could apply national labour law regulations to posted workers. Here, too, the CJEU found in favour of the free movement of services over national labour regulations. Many commentators saw these judgments as a serious threat to national systems of industrial relations with a clear anti-labour, liberalising bent (e.g. Joerges and Rödl, Reference Joerges and Rödl2009). As such, they created a paradigmatic environment to study how controversial innovative CJEU doctrines would play out ‘on the ground’. With a few years’ hindsight, several studies have explored the ‘real-world’ impact of the judgments in different national settings (e.g. Bücker and Warneck, Reference Bücker and Warneck2010; Freedland and Prassl, Reference Freedland and Prassl2015b). The main result of these efforts is that the impact of these judgments is heavily conditioned by the preferences of political parties and important interest groups, in particular employer associations. The judgments in Viking and Laval, for example, were shown to have had little impact outside Sweden, where Laval had originated, the peak employer organisation had supported the legal case against the unions and where a conservative-liberal government was broadly in favour of labour market reform, and the UK, where no statutory right to strike existed in the first place, unions were already faced with financial liabilities and now put even further on the defensive after the CJEU rulings (Bruun and Jonsson, Reference Bruun, Jonsson, Bücker and Warneck2010; Novitz and Syrpis, Reference Novitz, Syrpis, Freedland and Prassl2015).
Denmark is an interesting case in point for resistance through containment. The Danish industrial relations regime is very similar to the Swedish, and industrial action against foreign employers frequent. Yet, the CJEU's ruling in Laval had little impact (Bruun and Jonsson, Reference Bruun, Jonsson, Bücker and Warneck2010). Before the Laval judgment, the Danish judiciary had shielded national industrial relations from interference by EU law. In 2005, in a case that bore strong similarity to the facts in Laval, the Danish Labour Court had decided against the plaintiff employer association and in favour of the unions in finding no fault with a sympathy strike against a foreign company employing posted workers in the building sector. Rather, it held that the Danish trade unions had an ‘evident and strong’ interest in ensuring that work in Denmark is done under the conditions of Danish collective agreements, even where that work is carried out by foreign companies employing foreign workers.Footnote 17 The Danish court also saw no necessity to refer the question to the CJEU, since it found the relevant EU law sufficiently clear (Neergaard and Nielsen, Reference Neergaard and Nielsen2010, pp. 458–459). After the CJEU's Laval judgment, which came to very different conclusions, the conservative-liberal Danish government set up a tripartite commission to work out a legislative response. The resulting law explicitly allowed collective action by Danish unions against foreign employers with the aim of concluding a collective agreement, albeit under conditions that were formulated specifically to inoculate the law against a court challenge based on the CJEU's Laval ruling. With the support of the social partners, the Danish legislator used the leeway offered by the CJEU's case-law to ‘court-proof’ its legislative solution. It prevented as far as possible the negative consequences to its national labour law regime and preserved its original regulatory goals (Bruun and Jonsson, Reference Bruun, Jonsson, Bücker and Warneck2010; Seikel, Reference Seikel2015).
A similar dynamic was at play in the legislative response by German Länder to the CJEU's judgment in the Rüffert case. That case related to the question whether public procurement tenders, which are of significant economic importance, can demand foreign bidders to adhere to domestic collective-bargaining agreements. In its judgment, the CJEU had explicitly struck down such a clause in a public procurement contract issued by the German Land of Lower Saxony. Since they do not refer to minimum rates of pay laid down in general legislation, the CJEU found collective agreement clauses to be a disproportionate restriction on the freedom to provide services. In the aftermath of this judgment, many Länder governed by coalitions involving social democrats adopted minimum-wage legislation specifically related to public tenders – a construction that they saw compatible with CJEU case-law (Blauberger, Reference Blauberger2012). As a sign of an apparent retrenchment in its free-movement case-law, the CJEU recently confirmed the compatibility of this solution with free-movement principles.Footnote 18
A different set of studies has set out to analyse the domestic impact of CJEU doctrines on EU citizenship as it relates to access to welfare benefits.Footnote 19 At the time of its inclusion into the Treaty of Maastricht, Member State governments had not intended for EU citizenship rights to exceed existing free-movement rights codified in secondary legislation (Hofmann, Reference Hofmann2013, pp. 238–239). The CJEU, however, held in a long series of judgments that the treaty provisions on EU citizenship directly confer rights upon citizens and that EU citizenship can entitle migrants to social benefits as long as they could demonstrate a ‘real link’ to their host society.Footnote 20 It proclaimed that EU citizenship was ‘destined to be the fundamental status of nationals of the Member States’ and that EU citizenship established ‘a certain degree of financial solidarity between nationals of a host Member State and nationals of other Member States’.Footnote 21 For national authorities, the combined result of this line of case-law raised the spectre of ‘welfare migration’ by economically inactive EU citizens (Blauberger and Schmidt, Reference Blauberger and Schmidt2014). Authorities in countries offering relatively generous social benefits broadly moved to restrict access to such benefits by the unemployed, the under-employed and students. While the UK initially imposed relatively few restrictions, the British government started to introduce and reinforce a series of eligibility tests in 2013. These formulated thresholds, such as the minimum weekly earnings required in order qualify for access. Such generally applicable thresholds ran counter to the original CJEU doctrine, which had required individual assessments rather than mass administrative procedures (Blauberger and Schmidt, Reference Blauberger and Schmidt2017a, p. 445). German social administration, in contrast, had from the outset rejected applications for social benefits by EU migrants who were not clearly employed. While there was considerable doubt whether this approach would not be discriminatory under EU law, German authorities waited for case-law to clarify such questions (Blauberger and Schmidt, Reference Blauberger and Schmidt2017a, p. 446; Heindlmaier and Blauberger, Reference Heindlmaier and Blauberger2017, p. 1212). Austrian authorities in turn require migrant EU citizens to prove their right to reside, which is tied to financial self-sufficiency, and threaten to withdraw this right if EU migrants apply for social benefits, leaving them in a situation of considerable legal uncertainty (Heindlmaier and Blauberger, Reference Heindlmaier and Blauberger2017, p. 1208). These approaches push back against the CJEU's initially generous interpretation of EU citizenship rights. Restrictive positions were eventually vindicated when the CJEU started to withdraw from its expansive doctrine on access to social benefits and entered was has been interpreted as a ‘reactionary’ phase in its citizenship case-law in the context of the European economic crisis in the late 2000s (Spaventa, Reference Spaventa and Kochenov2017).Footnote 22
A similar scenario unfolds in relation to the CJEU's series of case-law regarding EU citizens’ access to cross-border health care. Member State governments never intended to subject the health-care sector to EU law's free-movement regime but, in a series of cases starting in 1998, the CJEU decided to the contrary.Footnote 23 It found that health care constituted a service covered by EU law even if patients did not have to pay for it, and that patients would have the right to cross borders to receive this service and be reimbursed by their health insurance mostly without requiring prior authorisation (Martinsen, Reference Martinsen2005).
A number of recent studies have demonstrated that national authorities, while complying with individual CJEU judgments, have found various means of limiting the impact of this line of case-law on their health-care regimes. In response to the original CJEU judgments, the Danish government, for instance, had limited patient mobility to health-care services that were not delivered ‘in kind’ – that is, those that involved an element of private pay by patients. This solution was maintained despite vocal doubts within the responsible administration about its conformity with CJEU doctrine. As in the case of industrial relations, Danish courts shielded the domestic system, making almost no mention of EU law in their health-care case-law. The law was only changed in line with CJEU case-law after a prolonged intervention by the Danish Ombudsman (Martinsen and Mayoral, Reference Martinsen and Mayoral2017, pp. 421–423). In contrast, Spanish courts have more actively engaged with CJEU case-law in their health-care jurisprudence. The Spanish legislature, however, has largely ignored even the case-law of national courts and maintained legislation limiting patient mobility, denying CJEU doctrine ‘erga omnes’ effect (Martinsen and Mayoral, Reference Martinsen and Mayoral2017, pp. 425–426). The Polish legislature, on the other hand, largely adapted legislation but retained restrictive administrative practices, particularly regarding prior authorisation, that in effect limited access to cross-border health care. In this case, too, the national judiciary broadly shielded national practices (Vasev et al., Reference Vasev, Vrangbæk and Křepelka2017, p. 470).
As these studies demonstrate, pushback against CJEU doctrine is possible even where national authorities formally comply with individual CJEU judgments. National administrations have found various means of containing the wider effects of the CJEU's expansive free-movement case-law in such diverse areas as the right to strike and the application of national collective agreements to posted workers, migrant EU citizens’ access to welfare benefits and patient mobility. Containment strategies have included a lack of legislative adaptation or retention of restrictive administrative practices, in effect forcing affected citizens to repeatedly claim their rights through courts and other arbitration bodies. Resistance can also take on the form of legislative pre-emption, when national legislatures implement narrow reform but impose new restrictions in an attempt to ‘define rather than be defined’ and to avoid future interference (Martinsen, Reference Martinsen2005, p. 1049). In some cases, national authorities can also rely on what have been termed ‘protective equivalents’ – that is, national fall-back options that allow authorities to ‘give up’ pieces of legislation or administrative practices because alternatives exist that have (so far) not been addressed by CJEU case-law (Werner, Reference Werner2016). National judiciaries also play a role in shielding national solutions from continuous challenge. This is not to say that the CJEU has little effect on national policies, but that the effect of controversial doctrines can be kept to what national policy-makers would find an acceptable level, so long as basic regulatory policy goals can be protected. While pushback through containment does not send as strong a signal to the CJEU as outright non-compliance, it is unlikely that the Luxembourg judges are oblivious of or uninterested in the ‘afterlife’ of its doctrines in the Member States. The recent adjustments in the CJEU's doctrine on free-movement rights can therefore also be read as an acknowledgement of Member State resistance below the threshold of non-compliance.
5 Conclusions
Madsen, Cebulak and Wiebusch (Reference Madsen, Cebulak and Wiebusch2018) in this issue asked for patterns or specific forms of resistance against international courts. This contribution has reviewed the available data on different forms of resistance against judgments of the CJEU. While concrete backlash has occasionally emanated from academic observers and the media, Member State authorities and the judiciary have so far broadly supported the integrity of the EU legal system, with the CJEU at its apex. Nonetheless, various forms of pushback are more common than accounts of the CJEU's apparent success suggest. It is not uncommon that national policy-makers, administrations and the judiciary fail to comply with individual rulings. Moreover, Member State authorities have developed multiple strategies to limit the practical effect of controversial lines of CJEU case-law.
Non-compliance with individual judgments is a relatively strong form of pushback against the CJEU. Despite the fact that the EU has in place a fairly sophisticated system of sanctions, compliance with CJEU rulings is at least delayed in about half of all cases after the conclusion of an infringement procedure, and the Commission refers about 10 per cent of cases back to the CJEU for continued non-compliance. Even where financial penalties have been applied, compliance remains incomplete where national interests are entrenched, and the Commission refrains from pursuing ‘hopeless’ cases in full. The extent to which national judiciaries comply with CJEU doctrine is less well known. Empirical studies suggest that between about 5–10 per cent of CJEU judgments in the preliminary reference procedure are not fully complied with by the referring court. Even fewer data are available on the extent to which national courts follow CJEU doctrine when they do not refer a question to the CJEU but rather apply EU law themselves. Prominent cases of non-compliance, usually involving national courts of last instance, are widely debated in academic literature and occasionally find the attention of the media, but little is known about more covert patterns of pushback by national judiciaries, particularly in lower instances. While sending less obvious signals to the CJEU, pushback can also take the form of national efforts to contain the wider ramifications of CJEU doctrine. In particular in cases where the CJEU doctrine has touched on salient issues of ‘mega-politics’, such as labour market or welfare policy, Member States have proven adept at keeping its practical effects within the realms of the politically acceptable. The CJEU, in turn, has responded to these forms of pushback by moderating its jurisprudence in free-movement cases.
This, of course, does not mean that the CJEU is an ineffective international court, certainly not in comparison with some of the regional courts that were designed in its image. Rather, despite its widely heralded success story, the CJEU's impact on Member State policies should not be taken for granted. Many current policies of EU Member States are impossible to explain without the influence of CJEU case-law, but the ‘malleability’ of the court's impact makes even controversial judgments acceptable to national authorities. The availability of ‘work-arounds’ that national authorities can live with shields the CJEU against significant backlash (Werner, Reference Werner2016). At the same time, the multiple processes of pushback in the Member States lead to an outcome of considerable heterogeneity. Citizens and companies can rely on EU law to very differing degrees depending on their location on the European continent. This heterogeneity may be the price to pay to avert serious backlash against an international court that has time and again shown its willingness to reach controversial conclusions and develop law beyond the express intent of Member States and the EU legislature.
Direct backlash against the CJEU remains unlikely in the future as long as Member States remain fundamentally committed to co-operation and continue to feel bound by rule-of-law norms. Recent developments in Member States such as Poland and Hungary demonstrate that this cannot be taken for granted (Kelemen, Reference Kelemen2016). The British newspaper The Daily Mail’s depiction of British High Court judges as ‘enemies of the people’ in the wake of that court's decision to grant the final say over Brexit to the UK parliament (Slack, Reference Slack2016) clearly demonstrates that populist invective against elites does not stop shy of judges, international or domestic. The EU's sophisticated enforcement regime is weakest precisely when it comes to protecting fundamental norms. As this paper has demonstrated, the CJEU itself has limited means to affect entrenched Member State policies, and relying on the CJEU to counteract these developments may not be the best strategy (Blauberger and Kelemen, Reference Blauberger and Kelemen2017). The solution to a rule-of-law crisis is unlikely to be judicial.
