‘Pushing the boundaries’: a dialogical account of the evolution of European case-law on access to welfare

ABSTRACT

In the context of the 2004 Enlargement, several EU governments reformed their social legislation to restrict the access to benefits of job-seeking or inactive EU citizens. Many of these restrictions were in tension with the case-law of the European Court of Justice, but when it came to judge their compatibility with EU law, the ECJ was more lenient than many anticipated. This article analyses this shift in ECJ case-law by looking at the dialogue between the Court and national authorities against the backdrop of EU legislative reform. It demonstrates that Member States contributed to the evolution of case-law by ‘pushing the boundaries’ of EU law both domestically and before the Court. It shows in particular how closely the arguments presented before the Court by national judiciaries or governments correlate with the new interpretations adopted by the Court itself. This is illustrated with empirical evidence from the UK and Germany.

KEYWORDS:

• European Court of Justice
• case law
• member state administrations
• welfare rights

1. Introduction

EU law gives beneficiaries of free movement the right to enjoy welfare benefits without discrimination in the host State, while leaving it to each Member State to design its own welfare system. But while economically active citizens enjoy full access to those benefits, jobseekers and inactive citizens have more limited entitlements.

Starting with the 1998 Martinez Sala judgment, the European Court of Justice (ECJ) promoted an expansive interpretation of the latter entitlements (Blauberger and Schmidt 2014:2; Roos 2016, 269). However, in the run-up to the great Enlargement of 2004, fears of ‘welfare tourism’ – the idea that people emigrate to obtain welfare benefits rather than to work (Kvist 2004, 306) – led several governments to reform their social legislation in order to restrict them (Schmidt, Blauberger, and Martinsen 2018:1391; Roos 2016, 280).

Many of these restrictions were in tension with the expansive ECJ case-law just mentioned, but when it came to judge their compatibility with EU law the ECJ was more lenient than many anticipated. This more recent case-law, epitomized by the 2014 Dano judgment, is often depicted as a dramatic departure from the previous case-law: a ‘striking break’ separating a progressive-minded ‘constituent phase’, extending from the late ‘90s to the mid-2000s, from the current ‘reactionary phase’ (Spaventa 2017:207–208; Blauberger et al. 2018, 1423). Vocal criticism from parts of legal doctrine (see e.g., Shuibhne 2015; O’Brien 2016; Iliopoulou-Penot 2016; Spaventa 2017) has led to suggestions that the shift was legally ‘puzzling’ and politically motivated (Blauberger et al. 2018, 1432).

One suggestion made was that the change of Court’s jurisprudence can be explained – at least as a contributing factor – by the politicization of EU free movement and shifts in public opinion as revealed through media analysis (Blauberger et al. 2018, 1422) or more generally by a ‘wider mood of integration estrangement’ (Shuibhne 2015:916; Spaventa 2017, 209). Without necessarily disagreeing with this reading, we believe that it is capable of further refinement. While accounts centred on public opinion tend to depict the Court as a lone actor trying to plot a sustainable course for free movement law in times of euro-scepticism, we argue here that a dialogical account more accurately captures the legal and political dynamics at play.

The first dialogue that we refer to is that between the Court and the EU legislator. Through legal analysis, we demonstrate that the Court’s new course largely flows from the policy choices made with Directive 2004/38 – the act which recast free movement legislation on the eve of the enlargement. Not a ‘puzzling’ shift, then, but a rather orthodox pattern of judicial deference to the political choices of the legislature.

The second dialogue that we refer to is between the Court and national authorities. This is an aspect that, to our knowledge, the existing literature on the evolution of the Court’s free movement and welfare case-law has not thoroughly explored: the relatedness between the evolution of national welfare practices, the arguments that they bring before the Court, and the evolution of the latter’s case-law. In this sense, our theoretical proposal draws on the public policy approach because it focuses on the analysis of the policy environment (Falkner 2018:769; Matthieu et al. 2018).

In order to carry out this part of our analysis, we examine selected judgments of the ECJ against the backdrop of evolving national policies in Germany and the United Kingdom (UK). In both countries, social policy reforms purportedly adopted to implement Directive 2004/38 have generated litigation leading to the involvement of the ECJ. This article shows how national actors exploited the legal uncertainties surrounding the topical provisions of EU Law, or the Court’s reading thereof, to advocate for stricter interpretations before the Court itself, strategically signalling their policy preferences. In other words, Member States contributed to the evolution of ECJ case-law by ‘pushing the boundaries’ of EU law both domestically and before it.

The next section of the article situates it within existing literature and elaborates our argument. The third section sets the legal scene for the evolution of the ECJ case-law. The fourth section analyses social policy reforms in Germany and the UK as well as the connection between the litigation generated by those reforms, the cases brought before the ECJ and its rulings. The final section concludes.

2. ‘Pushing the boundaries’ as motor for change in EU case-law

As stated, the central claim of this article is that the reorientation of the case-law referred to above is largely the outcome of ongoing dialogues between the Court and the European legislator, and between the Court and national actors. In other words, recent rulings of the Court reflect the legal changes brought about by Directive 2004/38 and the national practices in that specific policy field (Mathieu, Christian., and Hartlapp 2018, 656).

The Directive 2004/38 has been described as an attempt to ‘modify the Court’s jurisprudence in a restrictive way’ (Blauberger et al. 2018, 1423). While legally speaking it could not do that – ‘modify’ the Court’s interpretation of Treaty provisions – it compelled the Court to reconsider its position. After all, from a constitutional perspective, ‘judicial respect for legislative choices is a standard expectation’ (Shuibhne 2015, 891). This aspect is examined in greater detail in the next section.

As for the dialogue between the ECJ and national actors, the literature traditionally offers two accounts: the legal autonomy perspective, advocating that the ECJ is insulated from member states’ influence, and the political power approach arguing that the Court is concerned by approval of the member states given that the effectivity of its decision finally depends on the level of compliance (Geoffrey., Kelemen, and Schulz 1998, 149–150). We draw on this second approach and share the premise that the Court is indeed responsive to member states’ preferences. However, we argue that the literature has not exhaustively described the strategies available to national actors when in disagreement with ECJ rulings.

The concepts traditionally employed are those of threats of legislative override (Larsson and Naurin 2016, 391) and contained compliance (Conant 2002). Without dismissing the explanatory power of these concepts, we argue here that they do not fully capture the dynamics at play in our cases, and do not overlap with our concept of ‘pushing the boundaries’. Credible threats of legislative override require supranational collective and unanimous action (Carrubba, Gabel, and Hankla 2008, 438). Our focus, however, is on how national governments or other national actors attempt to influence ECJ decisions individually. ‘Contained compliance’ captures the behaviour of governments implementing EU Law only partially, with the goal of maintaining national regulatory preferences (Blauberger 2012). This can be seen only indirectly as a strategy to push for a change in the Court’s case-law. As we will see, however, national actors directly and actively challenge the current understanding of EU law and instigate change in dialogue with the ECJ (Krämer-Hoppe 2018, 806). Indeed, they use their channels of communication with the Court to stimulate its response and obtain a new case-law more in line with their preferences.

This is the strategy that we call ‘pushing the boundaries’, and as noted it is conceptually distinct from the two concepts discussed, albeit connected to containing compliance. The latter (Conant 2002) reflects a top-down perspective of Europeanization where EU law is an independent variable and national actors are either passive or have to engage actively in complex adjustment to circumvent EU law (Heindlmaier and Blauberger 2017, 1204). Its objective is to avoid further interferences of supranational actors (Conant 2002, 32), not to elicit their response. Pushing the boundaries is also conceptually different from threats because its main medium is ‘exchanging views’ with the Court (Krämer-Hoppe 2018, 807) with the aim of persuading it to change.

It is worth noting that the Court does not have agenda-setting power and only decides on the cases that are brought before it (Tridimas and Tridimas 2004, 12). Thus, a necessary precondition for the dialogue with national authorities to occur is that relevant cases are introduced, and this usually happens because of national practices that are contestable from the standpoint of EU Law.

The European Commission, as ‘guardian of the Treaties’, may open an infringement procedure. This is a protracted process, which explains why – in our case – some of the issues of interpretation that had arisen with the transposition of the Directive in the mid-2000s only reached the Court years later. Infringement procedures provide Member States with ample opportunities to argue their case. The first step of the procedure is political (Closa 2019, 709) and referrals to the Court are preceded, and often prevented, by negotiations. If the case is referred, governments may submit observations to the Court and usually do so to defend their practices. This, as we will see, is one of the places where the correlation between Member States’ efforts and the new orientations of the Court become visible.

Relevant cases may also come before the ECJ through preliminary references. Problems of implementation generate litigation before national courts (Golub 1996, 361) and these may or must, as the case may be, refer the case to the ECJ so it can clarify any aspects of EU law involved. The Court depends on national courts not only for the correct referral of preliminary questions, but also for the effective implementation of EU law, so they are key interlocutors (Weiler 2001:192; Alter 1996, 481).

Beyond their appearance of neutrality and impartiality, national courts can and do use referrals to pursue their agendas e.g., to challenge national policies and legislation (Tridimas and Tridimas 2004; Golub 1996, 362) or to settle disputes between higher and lower national courts (Alter 1996, 470). This is common in a legal system, such as EU Law, characterised by ‘constitutional pluralism’, where even within the same State courts of different levels hold different views about the place of EU Law and the interpretation of its prescriptions (Stone Sweet and Stranz 2012, 93). Indeed, referrals to the ECJ may be quite outspoken as to the desired outcome (Schmidt 2014, 771), and are the mechanisms through which national courts inform the ECJ about what they consider to be problematic interpretations or developments (Golub 1996, 379).

This constant interaction between the Court, governments and national courts is a process through which all actors can signal each other with their preferences. It gives the Court the opportunity to refine its earlier rulings by accommodating concerns and criticism of national actors (Kramer 2016, 288). In Davies’ words, ‘trends in outcomes in Luxembourg may therefore be (partly) the product of trends at national level’ (Davies 2018, 1458).

Before we test the relevance of these arguments in our empirical cases, the next section sets out the legal context in which these have unfolded.

3. Setting the legal scene: free movement and access to welfare

Since the Treaty of Rome of 1957, persons qualifying as ‘workers’ have enjoyed free movement and equal access to all social benefits. When at the turn of the 1990s free movement was extended to jobseekers and inactive citizens, welfare entitlements were not. On the one hand, the ECJ established in Antonissen (C-292/89) that jobseekers were a subcategory of ‘workers’ enjoying equal access to the labour market, but not to social benefits. On the other hand, the Maastricht Treaty gave inactive EU citizens the right to free movement, but legislation made their stay contingent on ‘having sufficient resources to avoid becoming a burden on the social assistance system of the host Member State’ (see Directive 90/364, Article 1).

This tidy arrangement was shaken in 1998, when the Martinez Sala judgment (C-85/96) established that EU citizens staying legally in another Member State could invoke non-discrimination in regard of all benefits covered by EU law, including those that EU legislation itself reserved to workers only. This judgment, usually described as paradigmatic of the expansive ECJ case-law of the era, remained in fact an outlier. Its facts were highly idiosyncratic (Davies 2018, 1450). And while its implications were less radical than it seems – Member States could still expel citizens that did not have sufficient resources – the Court’s conclusion that all EU citizens were entitled to social benefits merely by staying legally in another Member State flew in the face of the legal framework outlined above (Tomuschat 2000).

Grzelczyk (C-184/99), decided two years later, is also considered a milestone of the ‘expansive’ phase. The Court established that recourse by an inactive citizen to public assistance cannot automatically be equated to lack of sufficient resources. This was a progressive reading of the legislation. However, far from giving inactive citizens full access to welfare, it only established the right to a ‘certain financial solidarity’ so long as the burden did not become ‘excessive’ for the host State. Besides, the Court based Mr. Grzelczyk’s right to equal access to welfare on the fact that he was exercising free movement rights, not on the fact that he was European citizen legally staying in the host State as in Martinez Sala. This different rationale, upheld in a number of other cases (Timmermans 2010, 346), implies that only citizens fulfilling the conditions for free movement – i.e. having sufficient resources – may claim equal access to welfare.

Subsequent case-law brought further details – and raised new uncertainties. In Bidar (C-209/03), the Court held that EU citizens are in principle entitled to student grants, but that Member States may require ‘a certain degree of integration’. In Collins (C-138/02), the Court confirmed that jobseekers were not entitled to the same benefits as workers, but held that they had an equal claim to financial benefits ‘intended to facilitate access to employment in the labour market’. The host State could however exclude those that did not have a ‘genuine link’ with the local employment market.

All in all, the case-law from this era was broadly speaking expansive, but its actual contribution to free movers’ rights was fairly ‘insubstantial’ (O’Brien 2016, 942) and – contrary to the monolithic description frequently given – it was fraught with uncertainties and contradictions.

Barely one month after Collins, on the eve of the Enlargement, the Council and European Parliament recast completely the existing free movement legislation by adopting Regulation 883/2004 on the coordination of social security and Directive 2004/38 on free movement rights. Both acts largely restated the law as it stood, with the Directive codifying the mildly progressive elements of Grzelczyk (see e.g., Article 14(3)). However, in matters of access to welfare, the Directive reflected the Member States’ fears of welfare tourism: it restated the sufficient resources condition for inactive citizens, emphasizing its importance for preserving the public finances of host States, and it introduced a new Article 24(2) authorizing the host State to exclude from social assistance inactive EU citizens in their first months of stay, as well as jobseekers. Article 24(2) also allowed Member States to exclude inactive EU citizens, during their first five years of stay, from ‘maintenance aid for studies […] consisting in student grants or student loans’.

This new legal framework left a number of legal questions unsolved and raised new ones (Thym 2015, 26). First of all: how could the categorical exclusions of Article 24(2) be reconciled with judgments such as Bidar and Collins? In Förster (C-158/07), dealing with maintenance aid for foreign students, the Court confirmed Bidar but accepted in light of new Article 24(2) that a five-year waiting period was not disproportionate to make sure that the student was sufficiently integrated. So, while reaffirming its old case-law, the Court displayed deference to the legislator – in keeping with its general orientation of the period relating to social legislation (Verschueren 2012, 177).

In Vatsouras (C-22/08 and C-23/08), concerning jobseekers, the Court did the inverse: it maintained its Collins case-law and held that – in order not to clash with it – the terms ‘social assistance’ in Article 24(2) of the Directive should be interpreted as not covering benefits intended to facilitate access to the labor market. Vatsouras was less deferential than Förster, likely because what was at stake was one of the core freedoms of the internal market (Spaventa 2017, 208). Still, it sought reconciliation between the Court’s own reading of the Treaty and the Directive, rather than simply letting the former prevail over the latter.

There were other unsettled legal issues. Article 24(1) and (2) contained provisions strongly in tension with Martinez Sala. Furthermore, the exact test to be applied under the sufficient resources conditions, and under the Directive’s terms ‘unreasonable burden’, was still unclear even after Grzelczyk. Finally, there was an apparent contradiction between Regulation 883/2004, which prima facie gave EU citizens who were resident and socially insured in the host State equal access to ‘special non-contributory cash benefits’, and Directive 2004/38, which authorized the exclusion from all ‘social assistance’ in the cases contemplated by Article 24(2). These are the issues at the heart of our case-studies.

4. Empirical cases

4.1 Germany: a dialogue of judges

In March 2006, the first government of Angela Merkel, a coalition between the Christian Democratic Union (CDU) and the Social Democratic Party (SPD), introduced amendments to Social Code II to transpose the Directive 2004/38. This new legislation excluded EU nationals ‘who are in Germany solely for the purpose of finding a job from the basic insurance for unemployed’.¹

Purportedly, this implemented Article 24(2) of the Directive. However, as we have seen, it was unclear at the time whether this provision was valid, or how it related to the non-discrimination principle as affirmed in Martinez Sala and Regulation 883. As a result of this uncertainty, legal conflicts emerged around the types of social benefits that had to be available for EU nationals, and the residence conditions imposed to access them (Heindlmaier and Blauberger 2017, 1212). In turn, this ‘generated a fair amount of divergent national jurisprudence’ (Devetzi 2019, 341).

In 2010, the German Federal Social Court (FSC) ruled that the newly-introduced exclusions conflicted with the European Convention of Social and Medical Assistance of 1953. The government responded by introducing an ad hoc reservation thereto. But in 2013, the FSC used an alternative legal reasoning to uphold the claim to social benefits of a Bulgarian citizen, holding that she was habitually resident in Germany and that it could not be proven that her residency ‘result[ed] solely from the purpose of seeking employment’ (FSC, 2013).

This ruling was hotly contested. The welfare benefits at issue are funded by the Länder, which in turn claim that the Federation should bear the costs. The German local authorities responsible for the management of the benefits decided not to implement the FSC decision, asked for increased Federal allocations and opted to wait for the government to enact new legislation (Heindlmaier and Blauberger 2017, 1213). Lower courts also resisted the ruling (Blauberger and Schmidt 2014, 5) and decided to either deviate from it (Absenger and Blank 2018, 164) or to refer the matter to the ECJ for clarification on the ‘possibilities to restrict union citizens’ access to national welfare’ (Blauberger et al. 2018:1437; Devetzi 2019, 342). This is how the Dano case (C-333/13) came before the Court in 2013.

The case was distinctive in at least two respects. First, on its facts, it was a textbook case of ‘welfare tourism’ (Davies 2018, 1454): the claimant had no work, did not look for one, and had no resources of her own. Her stay was not illegal under German law, but she clearly had no right to stay under EU Law. Yet, she invoked EU law to claim additional welfare benefits. Secondly, the referring court – the Leipzig Social Court – asked the ECJ to answer comprehensively on the question of what, if any, EU-derived welfare entitlements a person in such a situation had. In its Order for a preliminary reference (Leipzig Social Court 2013), it mentioned every non-discrimination rule in the book: Article 4 of Regulation 883, Article 18 TFEU, i.e. the basis of Martinez Sala, and Article 24 of the Directive. Had it intended to draw the ECJ attention on the need for a clearer case-law on the matter, and to favour a restrictive interpretation, it would not have formulated the questions differently, nor picked as different case. Indeed, the Leipzig Court was explicit about the risks of welfare tourism, the controversy it generated in Germany, and the need for the ECJ to clarify its position (see para. 3 c) and d) of the order). These concerns were echoed in the Written observations submitted on 30 September 2013 by the German Government to the Court (German Government 2013, paras. 73 and 130 ff).

The judgment of the ECJ – in Grand Chamber formation – reads like a direct response to the points raised in these documents. First, it rules that specific legislative provisions such as Article 24 of the Directive or Article 4 of Regulation 883 take precedence over the general prohibition of discrimination of Article 18 TFEU. This new doctrine – legally debatable but still within the bounds of acceptable interpretation (Spaventa 2017, 221) – takes Förster one step further: instead of holding that Treaty and legislation are both applicable, and ‘forcing’ the Treaty provision to agree with the legislation, it disables the latter as a concurrent and directly applicable rule. This is in line both with a deferential approach towards the legislator and with the demand for maximum legal certainty made by the national judge. Having taken Article 18 TFEU out of the picture, the Court relies on the letter of Article 24(1) of the Directive to finally lay Martinez Sala to rest and fully confirm the alternative line of reasoning implied in Grzelczyk: only those who fulfil the conditions to stay laid down by EU legislation, and thus exert free movement rights, have an equal entitlement to welfare benefits. This right is further subject to the categorical exclusions of Article 24(2) and to the proviso that inactive citizens, whose stay is predicated on having ‘sufficient resources’, may only claim welfare benefits so long as this does not impose an ‘unreasonable burden’ on the host State. As for Regulation 883, the Court relies on the Brey judgment – given one year earlier and discussed below – to state that it does not oppose national rule reserving ‘special non-contributory benefits’ to persons having a right to stay under EU law.

While it is true that Dano employs unprecedented language to stress the right of Member States to fight ‘benefit tourism’ (Thym 2015, 25), it legally builds on foundations laid down in previous case-law. More to our point, it is explicitly deferential to the legislator’s policy choices (see Dano paras. 67–82) and openly seeks to solve the problems represented by the national judge. Furthermore, while the ECJ follows its own argumentative path, the two key legal conclusions made by the Court, and summarized above, correspond to some of the arguments submitted by the Leipzig Court or by the German Government: that special legislative provisions override Article 18 TFEU; that legislation must be interpreted as a whole in light of the aim to avoid ‘excessive burdens’ for host States; that persons having no right to stay under EU Law may not invoke EU-derived welfare entitlements (German Government 2013, paras. 60, 90 and 120; Leipzig Social Court 2013, para. 3 b) in fine).

Alimanovic (C-67/14), the second case considered here, came before the ECJ at the initiative of the FSC while Dano was pending. It concerned a person who, after losing her status as worker under Directive 2004/38, was staying as a jobseeker and was therefore excluded from Social Book II benefits according to national legislation. Was this allowed under EU Law? In para. 30 ff of its reference order, the FSC underscored both the need for clarification in light of the uncertainties prevailing at national level, and the fact that a ‘no’ would have demolished the March 2006 reforms. On the key point – whether the benefits at issue were ‘Collins benefits’ that jobseekers were in principle entitled to – the FSC suggested that indeed, that was the case and the exclusion could not be justified. In a continuation of the confrontation already begun years earlier at national level, the German government argued instead that Social Book II benefits had the main goal of covering basic subsistence needs, not of facilitating access to the employment market (see Conclusions AG Wathelet, para. 57 ff). Furthermore, it stressed the need for more legal certainty and ‘workable solutions’ (Iliopoulou-Penot 2016:1011 and 1025 f; see also Written observations under Dano, para. 73ff).

Taking at face value the description of the benefits made by the German government, and also adopted by the Advocate General, the Court accepted this last argument and deemed Mrs. Alimanovic excludable under the Directive (Alimanovic, para. 42–46). This is another good example of the Court taking its previous case-law – Collins and Vatsouras – and revisiting it or perhaps quietly dismantling it (O’Brien 2016, 947) in response to a new national representation of the facts. Indeed, on the front of ‘workability’ Alimanovic completed the work undertaken in Dano by establishing that the exclusionary rule of Article 24(2) required no individualized assessment which, the Court explicitly stated, favored legal certainty (para. 61).

Remarkably, the ECJ judgment did not put an end to the controversy in Germany. In the final verdict on Alimanovic, the FSC ruled that – regardless of the exclusionary possibilities opened by EU Law, and in compliance with German constitutional standards of human dignity (Seeleib-Kaiser 2018) – EU nationals having a ‘solidified residence’ in Germany could not be excluded of benefits included in Social Code XII aimed at providing minimum subsistence benefits (Munta 2018, 13). Reliance on national constitutional law short-circuited the limitations built in Directive 2004/38 as interpreted by the ECJ. However, with the support of the new ECJ case-law, the coalition government of the CDU/SPD reformed the Migrant Workers Act (Fernandes 2016) to ban the access to social benefits for EU citizens for five years (Fernandes 2016, 19). Further changes to social assistance benefits for EU nationals were also introduced on 6 June 2019 when the government approved the Law against illegal employment and abuse of benefits (2019) that provides that EU citizens of other EU member states are now able to claim child benefit payments only if they are employed in the country. Needless to say, this ‘pushes’ the boundaries of what is acceptable under EU Law even further and may in time again give raise to litigation before the ECJ.

4.2 The UK: the ‘right to reside test’ in question

The UK has a long story of introducing residence requirements of dubious EU legality for access to its welfare system. The first Habitual Residence Test (HRT) was introduced by the conservative government of John Major in the UK in 1994 to fight benefits tourism (O’Neill 2011:228; Kennedy 2011, 5). It applied to everyone arriving in the UK, including nationals having resided abroad. Nationals were included precisely to avoid conflict with EU non-discrimination rules (Kennedy 2011, 5). The HRT established an automatic relationship between residence conditions and welfare entitlements (O’Neill 2011, 228) and had to be passed by anyone claiming Income support, Housing benefits and Council Tax benefits (Cracknell 1995, 1). Persons failing the test were considered ‘person[s] from abroad’ and their claims rejected (Kennedy 2011, 3).

In 1997, members of the government criticised the test because it affected UK nationals rather than migrants (Kennedy 2011, 5). After much debate, in 2000, UK nationals having returned more than two years prior were exempted (Kennedy 2011, 7). Furthermore, the legislation enacting the HRT failed to provide a definition of ‘habitual residence’² giving raise to legal uncertainties (Cracknell 1995:1; Kennedy 2011). The absence of precise definitions gave a certain margin of manoeuvre to national authorities in deciding which conditions had to be fulfilled to access to social rights (Bruzelius 2019, 73). As noted above, the ECJ established in Collins that UK such tests could be acceptable so long as they were proportionate.

Before the 2004 enlargement, the welfare entitlements of EU migrants became further politicized and welfare tourism was used to justify policy changes (Luhman 2015, 3). Under pressure from the Conservatives, the Labour government introduced new regulations³ to avoid welfare abuse (Kvist 2004, 313). Thereunder, EU nationals had to pass a ‘right to reside’ test in addition to the HRT. This meant that inactive EU citizens could only access benefits if they fulfilled i.e. the ‘sufficient resources’ conditions. Making use of Article 24(2) of the Directive, the government further specified⁴ that inactive citizens making use of the unconditional right to stay for an initial three months, could not pass the ‘right to reside’ test (Kennedy 2011, 5).

The new regulations purportedly transposed Directive 2004/38. Still, there were two points of friction. First, Article 14(2) of the Directive provides that Member States may not verify systematically if the conditions for residence are fulfilled. Secondly, the ‘right to reside’ test was much contested in relation to its potential discriminatory effects (Martinsen 2011, 951). EU nationals sent numerous complaints to the UK authorities and the European Commission. In spite of the existence of these potential incompatibilities with EU law (Roos 2016, 280), British courts did not request preliminary rulings from the ECJ. The historical lack of preliminary references of British courts might be explained by their preference for close textual rather than the interpretative techniques of the Court (Golub 1996, 376). In fact, the rulings of British courts displayed a ‘narrow application of habitual residence’ regarding the right to reside test that did not question its legality (O’Neill 2011, 239). This attitude can also be explained by the tradition of the UK courts of respecting the government in the decisions regarding social policy (Schmidt 2017), and considering the ECJ as a threat for national sovereignty (George 2000, 20).

However, in 2008 the European Commission initiated a protracted and high-profile infringement procedure (European Commission 2008). The Commission raised the issue, evoked above, of the relation between Directive 2004/38 and Regulation 883: in its view, a ‘right to reside’ test rooted in Directive 2004/38 could not be applied to any of the benefits falling under Regulation 883. Such benefits were by definition excluded from the notion of ‘social assistance’ in Directive 2004/38, and access thereto had to be open to any socially insured person residing in the UK within the meaning of Regulation 883. Thus a ‘right to reside’ test could not be applied as it unlawfully added to the conditions foreseen by the Regulation and was indirectly discriminatory (European Commission 2011).

Before the infringement procedure reached its judicial stage, an Austrian tribunal raised exactly the same question before the Court in Brey (C-140/12). The Commission promptly reiterated its views which, it should be noted, reflected the orthodox reading of Regulation 883 at the time (Verschueren 2014, 160). The UK also intervened and argued – along with other Member States – the opposite position that Member States could indeed apply a ‘right to reside’ test even to benefits falling under Regulation 883, if these could be understood as ‘social assistance’ under Directive 2004/38. All the intervening governments stressed that this interpretation better preserved the objective of the Directive – i.e. to protect the public purse from the ‘unreasonable burden’ potentially placed on them by free movers – and both the Advocate General and the Court followed this view (see Brey, paras. 50–61 and AG Conclusions, paras. 21–23 and 58–68). The Court also decided – of its own motion – to address another point: contrary to the Commission’s views, it held that Austrian rules similar to the UK right to reside test were compatible with Regulation 883, so long as the residency conditions were themselves in line with Directive 2004/38 (Brey, paras. 33–45).

The Commission maintained its infringement action, reducing its scope to benefits different from those concerned by Brey. What was at stake was the applicability of a ‘right to reside test’ to pure social security benefits (as opposed to ‘social assistance’ or mixed benefits as in Brey). Meanwhile, the politicization of welfare tourism increased in the run-up to the Brexit referendum. Worries about this phenomenon became central to the political debate and advocates of Brexit used such worries to criticise the constraints that EU law exercised on national welfare policies (Luhman 2015:1; Blauberger and Schmidt 2014:4;). As a result, the conservative government of David Cameron made the habitual residence test more stringent by introducing new ‘individually tailored questions’ and raising evidentiary requirements (Blauberger and Schmidt 2014, 4). Also, EU nationals arriving in the UK to look for a job needed to live in the country for three months in order to claim Child Benefits (Kennedy 2015, 31).

The case eventually came before the Court (C-308/14), which decided on it a mere nine days before the referendum in favour of the UK. This was not an obvious outcome from a legal standpoint, and the judgment’s reasoning is far from irreproachable (O’Brien 2017). As O’Brien writes, ‘it is impossible to know the intersection of legal and political in the minds of judges’ in such politically charged contexts (O’Brien 2017, 209). It is undeniable, however, that the Court’s finding closely matched the arguments put forward by the UK Government (see paras. 38–43, 49–53, 68–69 and 75–84). Thus, while the Court may certainly have had an eye on the UK electorate in ruling as it did, it most certainly heard – and gave weight to – the UK government’s plea to adopt the interpretation of EU Law best attuned to the need of ‘protect[ing] the finances of the host Member State’

5. Conclusions

The tension between national welfare systems and EU free movement existed since the beginning of European integration and, as this article shows, it is still unfolding. The conflict between the logics of ‘opening’ and ‘closure’, in Ferrera’s words (Ferrera 2009, 219), was highlighted during the 2004 Enlargement leading to Member States’ attempt to shield their social policies. The topical provisions of the Citizenship Directive were part of this move, but in order to fully solve the contradictions between EU law, which places the non-discrimination principle as the centre of EU free movement law, and national welfare policies that aim to protect the public purse, the ECJ also had to be persuaded to revise its case-law.

Without directly overruling the Court’s interpretation of the Treaties – a result constitutionally beyond its grasp – the legislator gave the Court an impulse and opportunity to reconsider its doctrines insofar as they seemed to conflict with the new provisions. As noted above, ‘judicial respect for legislative choices is a standard expectation’ (Shuibhne 2015, 891), and this applies in the EU as well. Indeed, the adoption of the Directive set in motion a process of gradual change that – case by case, ‘stone-by-stone’ (Lenaerts 2015) – led to the newer orientations eventually epitomized by Dano.

However, as we have shown, in order to fully account for the interpretive choices made by the ECJ it is necessary to also keep in view its constant dialogue with national authorities. As our analysis suggests, once EU law is enacted – in our case, Directive 2004/38 – existing tensions continue to evolve in the national sphere. They are fuelled by legal uncertainties due to the use of vague compromise language, or as the case may be to tension between new provisions and old case-law that is in principle still valid. The involved, iterative processes leading to the resolution of these uncertainties offer domestic actors an opportunity to push the boundaries of EU law and to promote, in dialogue with the Court, a change of jurisprudence that reflects their agendas and preferences: national governments use written observations and national judicatures use preliminary references to present the ECJ with new factual or argumentative elements capable of persuading it to complement or modify its position. In the cases we have examined, there is a strong correspondence between the arguments developed by national actors and the new doctrines established in the ECJ judgments. And in fine, the new case-law – whether good or bad is not the argument here – allows the Member States to shield their welfare system more confidently than they could have expected when the Directive was first enacted. Thus, to reiterate our initial claim, it may certainly be the case that the judges of the European Court are sensitive to the moods of the electorate, but if so it is certain that – in the case we have studied at least – the most proximate source of inspiration for the new case-law were the arguments and representations developed by domestic actors.

This article additionally shows that national actors follow different pathways to pursue this objective depending on their legal cultures and their views on European integration. In Germany, domestic actors had different preferences regarding the access of EU nationals to social benefits and used national litigation processes to signal the ECJ with their respective positions. The FSC opted for an expansive position that coincided with the past ECJ jurisprudence – and aligned with longstanding principles of German Constitutional Law – to guarantee the access of EU nationals to minimum benefits. The German legislator, in agreement with other domestic actors, instead attempted to exploit to their maximum the opportunities afforded by Directive 2004/38 to restrict legal conditions and reduce costs. In the UK, there was a consensus between national courts and the legislators about the exclusion of EU nationals from social policy schemes. The dispute thus reached the Court via centralized enforcement initiated by the Commission, and the parties seized the opportunity of an unrelated procedure that came earlier before the Court – Brey – to argue their case. As noted, after this first ‘round’, the UK government pushed the boundaries even further, its strategy eventually succeeding before the Court in the politically loaded context of the Brexit referendum.

But while the specific processes may change, the general dynamics – with States pushing the boundaries before a responsive Court, even more so out of deference to the EU legislator – are common to all the key ECJ pronouncements examined.

Acknowledgments

We would like to thank Constantin Hruschka and Giuliano Bonoli for their input and comments on an earlier draft of the article. We are also very grateful to the anonymous peer-reviewers for their useful and constructive feedback.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Swiss National Science Foundation (SNSF) [51NF40-182897].

Notes

1. Act amending the Second book of the Social Code (2006).

2. The Income Related Benefits Schemes (Miscellaneous Amendments) (No 3) Regulations 1994.

3. The Social Security (Habitual Residence) Amendment Regulations 2004 (the ‘social security regulations’).

4. The Social Security (Persons from Abroad) Regulation 2006 (SI 2006/1026).