A. Rush Portuguesa

The Court’s ruling in Rush Portuguesa concerned a Portuguese company which had entered into a subcontracting arrangement with a French company for the construction of a high-speed rail line.Footnote ³¹ Rush Portuguesa specialised in the construction of public infrastructure projects. It sent, or ‘posted’, its Portuguese employees to France to carry out the construction. French law specified that only the Office national d’immigration was permitted to recruit third-country nationals to work within France.Footnote ³² It maintained that Portuguese nationals were not entitled to the exercise of the free movement of workers until the end of the transitional period specified in the Act of Accession.Footnote ³³ It further argued that Rush could not rely on the freedom to provide services, as such employees were governed by the free movement of workers. Consequently, the office issued a decision against Rush.Footnote ³⁴ Rush argued this was contrary to Community law, per Articles 59 and 60 of the Treaty.Footnote ³⁵ The French Administrative Court, hearing the case, asked the Court of Justice, inter alia, if (A) Community law allowed Member States to preclude the company from providing services with its Portuguese workforce and (B) if Rush could be required to secure permits for its workforce.Footnote ³⁶ The Court of Justice considered that the Act of Accession’s transitional period governing the free movement of workers did not apply to Rush as the workers were not moving from Portugal to access the labour market of France. Rather, in its view, their employment relationship remained linked to their Portuguese employer, who had sent them temporarily to provide a service.Footnote ³⁷ At no time did the workers gain access to the French labour market.Footnote ³⁸ It, however, noted that Community law did not preclude Member States from extending their legislation and collective agreements to such workers.Footnote ³⁹

The Court’s ruling was controversial as it was perceived as undermining the political agreement the Act of Accession represented. Specifically, while Portugal would join the Community, there would be a moratorium on its nationals’ entitlement to work in other Member States. Carter notes that the decision represents the establishment of the legal concept of a ‘posted worker’ for the first time in EU law.Footnote ⁴⁰ Garben describes the Court’s decision as its ‘original sin’, arguing it was not reached via a deliberate and intentional interpretation of the Treaty provisions. Rather, she is of the view that the Court’s decision was motivated by a ‘concern to limit the effect of the transitional period’ to ‘enhance the application of EU law in the face of this transitional period’.Footnote ⁴¹ She highlights that the Court’s conception of a worker in Rush was not in line with its earlier case law, which identified a worker as being ‘any national of a Member State who actually and genuinely performs work in another Member State’.Footnote ⁴² There is undoubtedly a dissonance here. Garben points toward Regulation No. 1612/68 to support her position. That Regulation stated, in its preamble, that the right to free movement of workers must be enjoyed, without discrimination, by permanent, seasonal and frontier workers and by those who pursue their activities for the purpose of providing services.Footnote ⁴³ This point was also raised by the Advocate General in their opinion.Footnote ⁴⁴

Yet, as the Court of Justice noted, the Act of Accession made it clear that the relevant provisions of Regulation No. 1612/68 were not to apply to Portugal until 1 of January 1993, ie, after ruling in Rush. Footnote ⁴⁵ Therefore, the Court could not, in reaching its decision, rely on that regulation as an interpretative tool to understand EU primary law. Moreover, the provisions referred to by both Garben and the Advocate General are in the preamble of the Regulation and hence are not operative. Instead, the preamble to the Regulation may only be used as an aid for the interpretation of the provisions of the Regulation, where there is uncertainty regarding their meaning. Their use as broader interpretative tools to understand the Treaty, therefore, seems problematic.Footnote ⁴⁶ Garben’s position, therefore, overlooks that the Court’s decision in Rush was arguably a result of the political decisions of political institutions. Those political choices became law via the Act of Accession. Accession agreements have the same status as primary law, hence taking precedence over secondary law.Footnote ⁴⁷ It is not illegitimate to say the Court was required to defer to the political choices of the Union’s political institutions, given that these choices became legal norms when they were embedded within the Union’s legal system through the Act of Accession. Those legal rules reflect deliberate policy decisions made by political actors. In contrast to the view that Rush undermined social protection standards, Marshall argued in 1991 that the Court’s ruling ‘protects the labour markets of the old Member States while encouraging cost-effective construction projects which benefit both parties to a public works contract … [I]t takes the EEC one step closer to fulfilling its goal of economic integration and expansion’.Footnote ⁴⁸ Davies, reflecting on the Court’s decision ten years after its issuance, observed that perhaps the Court made a basic judicial error: it answered a question that was not necessary for it to reach its decision. By stating that Community law did not prevent Member States from extending their legislation or collective agreements to any person employed within their territory, it reassured host states that they could strengthen their laws to cover this new category of worker. In response, an alarmed Commission, concerned about the impact on cross-border trade, later proposed the Posted Workers Directive.Footnote ⁴⁹

B. The legislative response

In response to the Court’s decision in Rush, the Commission tabled a proposal in 1991 for a Directive to regulate this area of law.Footnote ⁵⁰ The purpose of the dossier was to coordinate the terms and conditions of employment for workers engaged in work within the framework of the freedom to provide services.Footnote ⁵¹ The Commission stressed that increasingly mobile cross-border workers within the Community would lead to disputes as to the applicable national labour rules.Footnote ⁵² In its view, an overreliance on conflict of law mechanisms to resolve such issues would increase the likelihood of recourse to litigation. Such litigation would likely result in barriers to the freedom to provide services due to uneven enforcement and conflicting jurisprudence between national courts.Footnote ⁵³ Based on these factors, the Commission outlined the aims, scope and content of its proposal as seeking to remove the obstacles and legal uncertainties that would impede the freedom to provide services via the creation of a framework that would establish clear rules.Footnote ⁵⁴ In particular, the ‘coordination’ of mandatory rules to be complied with in the host State to ensure worker protection.Footnote ⁵⁵ The Commission, in its proposal, therefore, placed a stronger emphasis on the economic and market regulatory logic of the proposal compared to social concerns. In addition, it emphasised the coordination of rules between jurisdictions to establish legal certainty as to whose rules apply when workers are posted between Member States.

The European Economic and Social Committee, in its opinion on the proposal, expressed concern regarding the aim of the Directive.Footnote ⁵⁶ It argued there was a tension between its goal of addressing both the freedom to provide services and securing the protection of workers. The Directive’s legal basis in the Treaty was viewed as not reflecting these dual goals.Footnote ⁵⁷ Parliament issued its first report on the proposal in 1992.Footnote ⁵⁸ It noted the significant impact of the Court’s case law as a source of framing for the proposal.Footnote ⁵⁹ Rush Portuguesa was explored in detail, the report stating the ruling ‘acted as a springboard for the present proposal’.Footnote ⁶⁰ It described the dossier as aiming to increase the ‘…freedom [to provide services] by removing some of the legal obstacles and uncertainties…’.Footnote ⁶¹ Parliament asks, based on Rush, if there are ‘…some restrictions [on the freedom to provide services] that can be retained?…’ and can Member States ‘…insist that those posted be treated identically to their nationals doing the same work? …’.Footnote ⁶² This indicates Parliament was contemplating the scope of the Court’s non-discrimination jurisprudence, and if it could restrict the freedom to provide services without discriminating between nationals from different Member States. A primary motivation, therefore, appears to be the regulation of the freedom to provide services and what type of restrictions may be retained by Member States. Parliament surmises that a host state may apply its legislation and collective agreements to posted workers. However, they may not, ‘… in the name of equality of treatment, apply all their national legislation where this leads to discrimination’.Footnote ⁶³

Parliament does explore the social dimension of the dossier, albeit briefly, highlighting the importance of the proposal by emphasising three consequences should it not become law. First, there would be increased movement of ‘cheap labour temporarily posted from south to north. This would be a distortion of competition … because competitive advantage would not be derived from superior productivity or the superior quality of the product but from fundamental differences in social and economic development’.Footnote ⁶⁴ This is less a concern about the protection of posted workers and more of a concern about protecting higher-earning workers from posted workers. It also establishes a very specific understanding of ‘fair competition’ based on quality and not price, which clearly emphasises a political value judgement on what forms of competition are considered ‘fair’. It also presupposes a more expensive service to be a better-quality service. Second, it highlights that it could become more common for workers to do the same work in the same place but be subject to excessive differences based on nationality, which would ‘… be against fundamental social principles’. Third, it notes that situations could emerge where posted workers were not protected by any law due to cascading subcontracting arrangements.Footnote ⁶⁵ These points show that where the labour and social dimensions of the proposal emerge, the concern appears to be about protecting high-earning workers from the impact of cheaper labour.

Parliament tabled over thirty amendments in its first report, many of which reflect the above points.Footnote ⁶⁶ For example, amendment 4 sought to introduce a recital 16(a), which specified, ‘Whereas, however, the mandatory rules for minimum protection in the home country must be observed where these provide for better terms and conditions for workers than those in the host country’.Footnote ⁶⁷ This provision can be considered in conjunction with amendment 23, which sought to introduce a new Article 3, paragraph 1(a) specifying that where the rules or collective agreements of the home State provided ‘better terms and conditions for workers than those of the host country, Member States shall ensure that the terms and conditions which apply in the home county are observed for workers posted to the territory of the host country’. [Emphasis added].Footnote ⁶⁸ In plain English, these amendments seem to have had the intention to benefit workers moving from jurisdictions with higher labour standards to those with lower standards, and not vice versa. Amendment 23 was accepted in 1996.Footnote ⁶⁹ By contrast, Amendment 28, which proposed a new Article 3(d), which would have stated that the act ‘shall be without prejudice to the right of Member States to apply or lay down laws, regulations or administrative provisions which are more advantageous to the workers covered in this Directive’ was not adopted.Footnote ⁷⁰ That amendment appears to have been intended to permit host states to go above the provisions of the Directive and apply higher standards than those under generally applicable domestic law. However, the amendment was rejected.Footnote ⁷¹ Parliament issued its final opinion on the dossier on 4 September 1996,Footnote ⁷² it was forwarded for a second reading in September and successfully passed the Parliament vote, leaving only the Council to assent to its adoption.Footnote ⁷³ In sum, Parliament’s amendments indicate an intention to only allow higher labour standards to prevail between jurisdictions.

The Working Party on Social Questions (WPSQs) was the body within the Council responsible for the dossier. WPSQ observed that the Treaty did not prevent Member States from applying their legislation and collective agreements to persons employed temporarily on their territory.Footnote ⁷⁴ It concluded that a ‘nucleus’ of mandatory rules for the ‘minimum’ protection of workers needed to be coordinated.Footnote ⁷⁵ Several Member States wanted the proposal to specify that it was not prejudicial to the home state’s application of more favourable working conditions, with the Commission indicating its openness to this. In particular, Belgium, Germany, Spain, Italy and France stated that the proposed ‘negative wording would imply that working conditions in the host country were always more favourable than in the country of origin: this was certainly not always the case’.Footnote ⁷⁶ A subsequent note from 15 July 1992Footnote ⁷⁷ is highly relevant; marked ‘N.B’,Footnote ⁷⁸ it stated, concerning Article 3(1), that ‘maintaining the law applicable to the employment relationship where it was more favourable than the law applicable in the Member State in which the work was carried out should not be excluded’Footnote ⁷⁹ This provision would later be amended and become Article 3(7) of the 1996 framework. The apparent political intention revealed by the negotiation documents is to allow a home state with higher labour standards, whose workers are posted to a host state with comparatively lower standards, to continue to apply their employment rules to the workers posted. By extension, it enables the extra-territorial application of the labour standards of the higher standard state within the jurisdiction of the lower standard state.

In June 1993, the Council received a communiqué from the Commission.Footnote ⁸⁰ An important note is included concerning Article 3’s scope and legal operation, the Commission specifying that a new ‘Paragraph 3 has been inserted to reflect the opinion of Parliament and is designed to guarantee the application of the more favourable working conditions provided for by the law applicable, usually the law of the home country’.Footnote ⁸¹ This new paragraph 3 states, ‘Paragraph 1 shall not prevent the application of the terms and conditions of employment provided for by the law applicable, which are more favourable to workers’.Footnote ⁸² An amended version of the text was circulated by the WPSQs four months later, in November; however, no delegations submitted additional comments on this new provision.Footnote ⁸³ These drafting documents introduce some ambiguity as regards the intentions of the legislators. Did they not intend that host states could also go beyond the minimum terms and conditions of employment laid down in the dossier, so that its provisions were a floor rather than a ceiling? It is not explicitly stated that this is the intention, so clarity here is hard to find. However, a later note circulated by the Presidency of the Council in 1995, seeking the delegation’s views on a number of issues, appears to add some clarity. Based on the agreement reached in December 1994, Article 3 would indeed be ‘kept open’ (ie, was to be a non-exhaustive list, a floor rather than a ceiling). However, the proposed text tied this understanding to what was then Article 3(6), which governed the public policy exceptions to Article 3(1).Footnote ⁸⁴ Article 3(1) was, therefore, intended to be a floor provision via its interaction with Article 3(6). Article 3(1), read alone, therefore constitutes a ceiling above which standards could not rise and not a floor below which standards could not fall. Article 3(6) would become Article 3(10) in the final text of the Directive. This interpretation was confirmed in January 1996 when the Council Presidency noted that under the current text of the dossier, Member States could apply terms and conditions on matters beyond Article 3(1)’s provisions in the case of public policy via Article 3(6).Footnote ⁸⁵ This provides considerably less flexibility to host states to establish higher standards for posted workers protection compared with Article 3(7), which negotiation documents indicate was intended to allow home states to continue to apply their rules if of a higher standard. Article 3(7), therefore, disproportionality benefits wealthier states’ regulatory systems vs less well-off jurisdictions, given that it embeds an asymmetrical preference for the application of rules of a higher level of protection, which are more likely to be Northern/Western European Member States. A decision of the European Parliament dated the 28th of October 1996, notes the institution’s approval of the Council’s position.Footnote ⁸⁶ This interpretation runs contrary to that reached by many scholars in the field. Kilpatrick, for example, understands Article 3(1) to establish a ‘floor of protection for posted workers, a nucleus of mandatory rules for minimum protection’, which is non-exhaustive as Article 3(7) provides that the floor does ‘not prevent the application of terms and conditions of employment which are more favourable to workers’.Footnote ⁸⁷ This is the opposite of the conclusions reached via the analysis herein.Footnote ⁸⁸ The legislative negotiations illustrate an attempt to embed an asymmetrical mutual recognition mechanism. This asymmetry was intentionally constructed by the legislature through the deliberate introduction into the 1996 framework of rules that differentiate between home and host state control. Particularly via the interaction between Articles 3(1) and 3(7), the latter provision privileges certain labour regulatory models over others if they are ‘more favourable’ to workers. The 1996 act institutionalised a bias favouring the application of higher labour standards and wage rates over lower ones, thereby privileging the regulatory systems of older, Northern and Western Member states over newer, Eastern and/or Southern states, as such regulatory divergences often align along these lines. In this sense, if moving from east to west, Western rules will apply, and if moving from west to east, Western rules will also likely apply. This conceptual framing allows the Directive to be understood not merely as a site of legal compromise but as a mechanism that embeds the uneven consolidation of regulatory authority for rule recognition across the single market.

I argued based on the above analysis that the Posted Workers Directive established a mutual recognition mechanism in line with Janssen’s framework. Article 3(1) contains prior conditions for the establishment of mutual recognition; this list is exhaustive, and it specifies the rules that a host state may apply to posted workers, which must be recognised by the home state. While Article 3(1) does not harmonise the content of rules, such as the substantive content of the term ‘minimum rates of pay’ under Article 3(1)(c), it does establish that the national rule as to what is equivalent to a minimum rate of pay must be constituted in compliance with the Directive’s requirements. Specifically, minimum rates of pay must be established in compliance with Article 3(1) first and second indents, meaning set down by law or administrative provision, or via universally applicable collective agreements. Once established in compliance with the Directive, they can be considered rules on the terms and conditions of employment for the Directive. Hence, minimum rates of pay can then be mutually recognised between jurisdictions. The Directive, therefore, coordinates the ‘type’ of rules, such as the terms and conditions of employment, that Member States must recognise. It does not, however, attempt to determine the content of those rules recognised. There are two exceptions to the general rule of Article 3(1) that can be identified from the negotiation documents. The first is implemented via Article 3(7). The provision governs Article 3(1) in that it confers regulatory authority under the Directive upon the state whose rules are ‘more favourable’ to workers. Whose rules are ‘more favourable’ is framed by the terms and conditions of employment specified by Article 3(1). National rules of the host or home state, falling within the scope of Article 3(1), are only applicable where they are equal to the standard that is of higher protection to the worker being posted; they are ‘more favourable’. This interpretation was taken by Davies in 1997 following the passage of the proposal into law. He noted that the ‘home State employer does not escape the regulation of its State of establishment, except to the extent that the host State’s provisions are more favourable to the employee’ via Article 3(7).Footnote ⁸⁹ The second exception to the rule that can be identified from the negotiations is contained in Article 3(10) of the final act, this provision making the list non-exhaustive. It specifies that the host state may apply higher standards than those contained in Article 3(1) on grounds of public policy when applied on a basis of equality of treatment, meaning non-discrimination.

Therefore, under the 1996 Directive, Member States were under an obligation to ensure companies operating within their jurisdiction (as defined by Article 1), providing cross border workers within the context of the freedom to provide services (per Article 2), apply the rules of Member States that fall within the coordinated matters (via Article 3(1)(a)–(g)) and are established under national law in compliance with the Directive (per Article 3(1), first and second indents). Not only does the Directive impose a positive obligation upon the Member State to which workers are posted (as defined by Article 2) to ensure companies falling with the Directive (per Article 1) guarantee the terms and conditions of employment (set out by Article 3(1)(a)–(g)), but furthermore this positive obligation implies a negative obligation upon the country of origin to recognise the host state’s regulatory control, unless this presumed right of regulatory control is superseded by Article 3(7). Hence, establishing a mutual recognition mechanism. This was the same interpretation of the provision taken by the Court of Justice in its subsequent case law in both Laval and later Rüffert.Footnote ⁹⁰ This amounts to a mutual recognition mechanism under Janssen’s framework, given that Member States must determine whose labour rules apply in cross-border situations. At the same time, the framework provides a preference for the application of ‘more favourable’ rules under Article 3(7), meaning that regulatory jurisdiction oscillates depending on the Member States in question – unless a justification for divergence from this general framework can be invoked under Article 3(10). In effect, the Directive establishes a requirement for Member States to permit labour rules lawful in one Member State to produce external legal effects in another Member State’s territory: an output-oriented understanding of mutual recognition in line with Janssen’s framework.

Daubler, writing in 1998, noted that the ‘hard core’ of labour law rules established via Article 3 was a ‘quite impressive’ catalogue of provisions.Footnote ⁹¹ Yet Daubler also observed a series of issues presented by the dossier, particularly for workers moving from lower-wage jurisdictions to higher-wage ones. Such employees are usually poorly informed on a range of issues, from knowing their rights, understanding how to pursue judicial action, and fearing being perceived as ‘disloyal’ for asking to receive wages higher than those in their employment contracts based on the standards or pay rates within the host state.Footnote ⁹² In sum, they are at a disadvantage. Evju notes that the Directive, via Article 3(1), took a conflict of laws approach to the management of cross-border issues of posting, but is of the view that the Directive’s overarching objective is to promote the cross-border provision of services.Footnote ⁹³ Davies, reflecting on the passage of the dossier into law, argued that the choice of legal basis suggested the adoption of an act with the aim of facilitating the cross-border provision of services rather than being a social instrument designed to protect workers. He suggested that the Directive does indeed advance the interest of home state employers, but that the primary beneficiaries under the Directive are the regulatory systems of the host state.Footnote ⁹⁴ The negotiations on the dossier reveal that indeed the social and labour protection concerns appear to be secondary to the central concerns of ensuring the freedom to provide services. Regulating the freedom to provide services, in the context of posted workers, allowed the Union to address all other issues.

C. The Laval Quartet

The Laval Quartet refers to a series of rulings of the Court of Justice which proved so controversial that they perhaps do not require the following review. The first ruling was Viking. Viking was a Finnish ship operator which sought to ‘reflag’ its ship in Estonia. ^{Footnote 95} By doing so, it could operate under Estonian labour law and pay lower wages to its employees.Footnote ⁹⁶ Its staff were, via a national trade union, members of the International Transporters Workers’ Federation. The Federation had a policy of opposing reflagging, where the motivation was to lower labour costs.Footnote ⁹⁷ Consequently, the national union planned industrial action. Viking claimed such action would infringe its right to establishment under Article 43 of the Treaty.Footnote ⁹⁸ A preliminary reference was made, and the Court of Justice determined that, though ultimately for the national court to decide,Footnote ⁹⁹ such actions could make it less attractive to exercise the right of establishment.Footnote ¹⁰⁰ The Court simultaneously noted that the right to strike forms an integral part of the general principles of Union law. Yet, that right must be balanced against the rights of others and may, therefore, be subject to restrictions.Footnote ¹⁰¹ While the right to strike is, therefore, a right under Union law, its objective of protecting workers would no longer be tenable if those jobs were ‘not jeopardised or under serious threat’.Footnote ¹⁰² Ultimately, the consequences of the decision in Viking were never conclusively resolved beyond the preliminary reference request, as the case was settled out of court.Footnote ¹⁰³

Laval was decided a week after Viking. ^{Footnote 104} Whereas Viking related to the right of establishment, the Laval case pivoted around the freedom to provide services.Footnote ¹⁰⁵ In Laval, a Latvian company secured a construction contract in Sweden. Under Directive 96/71/EC on the posting of workers, Laval was required to pay its workers the minimum rates of pay applicable under Swedish law per Article 3(1). A Swedish trade union sought to have Laval sign its collective agreement, the terms of which were more favourable to the posted workers from Latvia than the rules the company was legally required to comply with under Article 3(1). Furthermore, the nature of the agreement prevented Laval from determining the wages to be paid in advance.Footnote ¹⁰⁶ Laval refused to sign, and the trade union called industrial action, resulting in Laval being unable to conduct business in Sweden.Footnote ¹⁰⁷ Laval claimed the industrial action infringed its right to freely provide services.Footnote ¹⁰⁸ The Court of Justice, following a preliminary reference, ruled that compliance with Article 49Footnote ¹⁰⁹ applied not only to rules which are public but also to measures designed to regulate, collectively, the provision of services.Footnote ¹¹⁰ It noted that in light of its decision in Viking, the right to take collective action for the protection of workers may constitute an overriding reason in the public interest, such as to justify restrictions on the freedom to provide services.Footnote ¹¹¹ However, such actions may not be justified where negotiations on pay are characterised by a lack of sufficiently precise and accessible provisions that make it excessively difficult in practice for undertakings to determine their obligations concerning minimum pay rules applicable under Article 3(1).Footnote ¹¹² The Court found the collective agreement could not be described as universally applicable for Article 3(1), second indent, read in conjunction with Article 3(8), because it was not the case that the collective agreement must be observed by all undertakings within Sweden. Nor was it observed by the whole industry.Footnote ¹¹³ Furthermore, as Sweden had no system for declaring collective agreements to be of universal application, it had to apply for such recognition.Footnote ¹¹⁴ Sweden utilised a decentralised approach to collective bargaining that determined wages on a case-by-case basis.Footnote ¹¹⁵ The collective agreement, therefore, did not form part of the notion of minimum rates of pay under Swedish law. As such, there were no legal grounds through which Sweden or the union could impose the collective agreement on Laval.

Laval has been widely criticised, in particular for elevating economic freedoms above social rights.Footnote ¹¹⁶ Evju argues that the Court’s interpretation of the 1996 Directive, considered against its drafting background, is problematic given the Court’s decisions impinge upon important features of collective labour law and industrial relations dynamics.Footnote ¹¹⁷ The review of the drafting background conducted herein does not support Evju’s position. He notes that the Court’s restrictive approach to the kinds of terms and conditions of employment included within the scope of Article 3(1) of the Directive effectively disallows collective bargaining models typical in Scandinavian states. However, this view overlooks the express wording of Article 3(1), which emphasises that the Directive would include only universally applicable rules within its scope. For the Court to allow non-universally applicable agreements to come within its scope would have been contra legem unless they could be included via Article 3(8). Evju further argues that the Court’s case law sidelines the role of trade unions.Footnote ¹¹⁸ Yet Deignan notes that while some might frame the Court’s approach to Swedish labour law as ‘harsh’, if the Court had allowed Swedish law to stand uncorrected, it would have denied European law of its effectiveness and uniform application.Footnote ¹¹⁹ Relatedly, Refslund emphasises the conflict of legislative norms between the vertical and horizontal systems. The Nordic approach of self-regulation of labour markets is underpinned by voluntary agreements, secured via collective bargaining with limited state interference, and a more limited role for national law.Footnote ¹²⁰

Louis, in 2018, examined the impact of the European Trade Union Confederation’s (ETUC) judicial activities in the area of posted workers. The ETUC established a task force for coordinating with national legal teams to develop a coherent response, and additionally, established a working group to determine ‘how to win these cases’.Footnote ¹²¹ This resulted in the creation of a litigation network with the aim of defending trade union rights following Viking and Laval. In particular, to limit, or even reverse, the implications of those decisions.Footnote ¹²² The network’s approach was centred around the use of international social law to influence the perceived primacy given to economic freedoms within the EU legal order.Footnote ¹²³

In 2022, Louis explored the legal and political impact of the rulings from a social constructivist perspective, illustrating how trade unions and labour academics interpreted the rulings as detrimental to social Europe. Yet, he argues that these conclusions are not, stricto sensu, legal consequences of the rulings themselves, rather the judgments were construed as a defeat of social Europe because they were read as such by these actors.Footnote ¹²⁴ Building on this, Louis examines the impact of the Court’s decisions by considering its interaction with other institutions. By doing so, his approach adopts a ‘law in context’ view of the rulings, emphasising the role of the protagonists in the subsequent interpretation and impact of case law on social and political reality.Footnote ¹²⁵ He notes that Viking and Laval initially emerged from relatively normal conflicts in the private sector between trade unions and employers. At first, there was a lack of concern regarding the relationship between economic freedoms and social rights under EU law. Discussions initially centred around the interpretation of Swedish law in conformity with the 1996 Directive.Footnote ¹²⁶ This changed with the involvement of the ETUC, which, he argues, led to the subsequent establishment of a task force, which mobilised what he terms ‘judicial lobby’ activities. The task force’s teams produced documents presenting Laval and Viking as cases where the notion of social Europe was on the menu.Footnote ¹²⁷ Louis observes a consensus developed among labour law scholars – before the decisions had been issued – that the cases were of crucial importance to the future of a social Europe.Footnote ¹²⁸ In this regard, Louis notes that while the initial reactions to the rulings by trade unions were ‘cautiously positive’, this transformed into harsh criticism. Initially, the recognition of the right to take collective action was welcomed; later, this evolved into a perception of cementing the importance of market freedoms over social rights.Footnote ¹²⁹ Once issued, the decisions were viewed as a major defeat, although the Court for the first time found collective actions to represent a fundamental right under European law (without any explicit acknowledgement of such a right in primary law at the time).

Ericsson’s work explores Laval by considering the many and various readings, and possible misreadings, of the ruling. She emphasises that the law is a matter of interpretation, your interpretation of the law flows from your own vantage point in the broader political landscape. She noted that the narrative that the Court undervalued collective bargaining regimes was prominent in the literature, highlighting a number of examples.Footnote ¹³⁰ Rosas described the judgment as consistent with the notion of social Europe because it protects all workers by limiting the scope for Member States to use labour standard-setting regimes to shield their domestic labour markets.Footnote ¹³¹ Nic Shuibhne maintains that the Court missed an opportunity to mould a more nuanced approach to freedom of movement issues, in particular given its strict application of the justification and proportionality test.Footnote ¹³² Relatedly, Davies argues that the Court’s use of the proportionality assessment was inappropriate, as it assessed the interaction of the trade union’s right to strike with the employer’s right to the freedom to provide services without a sufficient understanding of the industrial relations context in which the industrial action was actually occurring.Footnote ¹³³ Ericsson furthermore examines the arguments made by Barnard, Novitz and Dorssemont, who contend that the Court’s rulings demonstrated both its economic bias and deregulatory approach to integration.Footnote ¹³⁴ Others, such as Deakin, are critical of the mere fact that the Court engaged with the question of judicial balancing at all, believing that the application of the proportionality test by the Court was inappropriate, echoing the views of Davies, Robin-Olivier and Evju.Footnote ¹³⁵

Against these critiques, Carter notes that competition in terms of labour costs and wages is a normal part of a market economy. In that sense, posted workers are, in his view, no more a source of unfair competition or social dumping than is the case within national markets. For him, social dumping can only be understood from the perspective of those who are negatively affected by the trend.Footnote ¹³⁶ Carter also notes that a variety of factors, such as negative integration dynamics and mutual recognition, can create structural biases which result in a process of deregulation. Mutual recognition often means that a host state’s rules are incompatible with internal market law, with few possibilities to derogate. This affects economies differently. Scharpf highlighted that liberal market economies are less affected by Laval and related rulings compared to coordinated market economies. Drawing on the framework of Hall and Soskice, Scharpf characterises liberal market economies as systems in which the state plays a minimal role in the economy and labour market intervention. Instead, the state focuses on creating the conditions for functional markets to prosper via fair competition and the provision of means-tested social assistance. In coordinated market economies, the state plays a more direct role in the economy via, for example, higher levels of labour market regulation, collective bargaining systems and the provision of universal social security systems. As a result, the Court’s jurisprudence is likely to impact coordinated market economies more directly.Footnote ¹³⁷ Cremers notes that, concerning the impact of posted workers on domestic wages and on social dumping dynamics, the identification of the macro-economic impact of such activities is, in reality, challenging. At the time of publication of that paper, Cremers observed that few Member States maintained systematic records of workers posted to their jurisdiction.Footnote ¹³⁸

In Rüffert, the Court built on its Viking and Laval jurisprudence, finding the 1996 Directive – interpreted in light of Article 56 TFEU – prevented Germany from requiring companies, subcontracted by public procurement contractors, to comply with specific domestic provisions on remuneration.Footnote ¹³⁹ This was because Member States could not require a company to go beyond the minimum terms and conditions of employment applicable by law.Footnote ¹⁴⁰ The Court concluded that the collective agreement in question was not of universal scope, and therefore did not apply via Article 3(8) as Germany had a system for declaring such agreements to be of universal application and had not done so.Footnote ¹⁴¹ As such, the collective agreement could not be enforced. In Commission v Luxembourg, it was argued by the Luxembourgish government that a domestic legal requirement for wages generally to be linked to the cost of living complied with Article 3(1)(C). This was rejected by the Court as that provision related to minimum rates of pay only and not wages generally.Footnote ¹⁴² Luxembourg claimed that the rule aimed to maintain good labour relations and therefore constituted a public policy exception per Article 3(10).Footnote ¹⁴³ The Court noted that Article 3(10) did establish grounds for derogation.Footnote ¹⁴⁴ However, such derogations must be interpreted strictly, so that their scope cannot be determined unilaterally by each Member State.Footnote ¹⁴⁵ The public policy exception may only be relied upon as a derogation if there is a ‘genuine and sufficiently serious threat to a fundamental interest of society’.Footnote ¹⁴⁶ For a Member State to avail of Article 3(10), appropriate evidence or analysis as to the expediency and proportionality of the restriction on the freedom affected must be supplied, along with ‘precise evidence enabling its arguments to be substantiated’.Footnote ¹⁴⁷ The Luxembourgish government only cited general objectives of protecting workers’ purchasing power and maintaining positive labour relations.Footnote ¹⁴⁸ Luxembourg, therefore, could not rely on the public policy exception.Footnote ¹⁴⁹

The decisions of Ruffert and Commission v Luxembourg can be linked back to Laval. The Court interpreted the 1996 Directive as meaning that when a host state applies terms and conditions of employment that are more favourable than those established by its national law or by universally applicable agreements, this would create situations in which posting workers from Member States with lower labour protection standards would become less attractive or more difficult. This, therefore, constituted a restriction under the Court’s case law, via its interpretation of Article 3(1) as representing an exhaustive list of rules which can be applied to workers. This reading appears to be in line with the general intention of the EU legislature as seen from the legislative negotiations on the 1996 framework. An interpretation which was further embedded in Rüffert and Commission v Luxembourg by the Court.Footnote ¹⁵⁰ In Deignan’s view, the by-product of the Court’s interpretation of the 1996 framework in Laval and Viking is a system akin to that envisaged by what was known as the ‘Bolkestein’ Directive (ie, the proposed measure which later evolved into the Services Directive) and its strict country-of-origin principle or mutual recognition mechanism.Footnote ¹⁵¹ In summary, the Laval Quartet generated a ‘tsunami’ of academic commentary, most of which viewed the Court’s approach negatively.Footnote ¹⁵²

D. The recast Directive

Despite the centrality of the Laval Quartet rulings to the Commission proposal for a Recast Posted Workers Directive, references to these cases are notably absent from the written records of the institutions’ negotiation positions on the dossier. By contrast, later rulings such as Elektrobudowa are referred to frequently during negotiations. In Elektrobudowa, the Court of Justice somewhat clarified the scope and meaning of the term ‘minimum rates of pay’ under Article 3(1)(C). It outlined, for example, that host states may require posting companies to include holiday allowances, daily flat-rate allowances, and compensation for travel time within the notion of minimum rates of pay. Additionally, where a collective agreement sets different rates of pay for workers based on the categorisation of employees into pay groups, they are valid once universally applicable and transparent.Footnote ¹⁵³ The Commission tabled the proposal for the Recast Directive in 2016.Footnote ¹⁵⁴ Its purpose was the ‘targeted revision of the Posting of Workers Directive to address unfair practices and promote the principle that the same work at the same place should be remunerated in the same manner’.Footnote ¹⁵⁵ The Commission outlined that the act would specifically address issues not addressed via the 1996 act.Footnote ¹⁵⁶ The attached Staff Working document cites a significant report commissioned by Parliament examining the Laval Quartet.Footnote ¹⁵⁷ That report outlines the main achievement of the 1996 act as establishing obligations for host states to ensure posted workers have a basic level of minimum labour protection applied.Footnote ¹⁵⁸ The report characterises the 1996 Act as a ‘minimum labour law’,Footnote ¹⁵⁹ a claim not supported by the overview of the negotiations concluded herein. The report presented the consequence of the Court’s rulings in Viking, Laval, Rüffert and Commission v Luxembourg as clarifying that: (1) collective actions, which include industrial actions, may amount to a restriction of economic freedoms under the Treaties;Footnote ¹⁶⁰ (2) collective actions are a fundamental right under Union law, but must be balanced against free movement rights;Footnote ¹⁶¹ (3) the 1996 Directive had been interpreted by the Court of Justice as a ceiling rather than a floor and as exhaustive coordination of national measures.Footnote ¹⁶² The Commission’s proposed revision of the 1996 framework sought to remove the reference to ‘minimum rates of pay’ and instead refer to the notion of ‘remuneration’.Footnote ¹⁶³ The move from the term ‘minimum rates of pay’ toward ‘remuneration’ is an attempt to address the Court’s interpretation of what was included in the notion of minimum rates of pay per Article 3(1). The Commission, while observing the Court’s case law, contributed to the progressive clarification of the law, and also noted that case law is unpredictable, necessitating legislative intervention.Footnote ¹⁶⁴ By moving toward the term remuneration, the Commission appears to have intentionally sought to widen the scope of Article 3(1) to include a broader array of rules within the notion of wages or minimum rates of pay.

The European Economic and Social Committee argued for the need to address the ‘false interpretation of the Directive as a maximum standard’ which had arisen following the Laval Quartet and ‘re-establish it as a minimum standard’.Footnote ¹⁶⁵ This move appears to be a clear repudiation of the Court’s interpretation of the Directive. The Committee believed collective agreements should be understood as benchmarks for remuneration.Footnote ¹⁶⁶ The inclusion of the concept of remuneration was viewed as a codification of the Court’s Elektrobudowa decision, which was seen as clarifying and widening the meaning of ‘minimum rates of pay’. In Elektrobudowa the Court of Justice expanded the interpretation of Article 3(1)(C)’s reference to the notion of minimum rates of pay by equating it with the notion of a ‘minimum wage’. It thereby interpreted the meaning of Article 3(1) through Article 3(7), second paragraph, which contained a reference to the notion of a ‘minimum wage’. That provision specifies that allowances paid to posted workers could be considered to form part of the notion of a minimum wage under Article 3(7), unless paid in reimbursement of expenditure actually incurred in relation to the posting. However, this broader notion of minimum wage can be relied upon, according to the Court, ‘only in so far as that definition, as it results from the relevant national law or collective agreements or from the interpretation thereof by the national courts, does not have the effect of impeding the freedom to provide services’.Footnote ¹⁶⁷ The Court’s broadening of the scope of Article 3(1)(C), via its reading of Article 3(7), was picked up by the legislature via the proposed shift towards the concept of remuneration. The term remuneration was also viewed favourably by the Committee of the Regions as it would strike a balance between ensuring cross-border service provision while also protecting workers and preventing unfair competition.Footnote ¹⁶⁸ A report issued by the European Parliament’s Employment and Social Affairs Committee on the issue of social dumping from 2016 provides important context.Footnote ¹⁶⁹ It makes several references to case law, noting in particular that in Laval, the right to take collective action, including industrial action, was found to form part of Union law, but such actions must be proportionate and not unjustly restrict the freedom to provide services.Footnote ¹⁷⁰ It notes that, in relation to the ruling in Elektrobudowa, the 1996 framework pursues a dual aim: (i) ensuring fair competition, and (ii) ensuring a nucleus of mandatory rules for the minimum protection of workers, supporting my previous conclusions.Footnote ¹⁷¹

Parliament’s Committee on the Internal Market and Consumer Protection proposed thirty-nine amendments. Amendment 19, altering Recital 12 of the Commission’s original proposal, implicitly adopts the Court’s language, observing that it is within the Member States’ competence to set rules on remuneration, but that these rules must be justified by reference to the protection of workers and not disproportionately restrict service provision.Footnote ¹⁷² Amendment 15 sought to alter the same provision so as to specify that wage setting is a matter for the Member States and their social partners, and care should be taken not to undermine national approaches.Footnote ¹⁷³ Amendment 15, which did not implicitly endorse the Court’s doctrinal reasoning as proposed by amendment 19, was successfully incorporated into the Recast Directive via Recital 17. Amendment 21 proposed the introduction of a new Recital 12(b) specifying that, in line with the Court’s findings, collective agreements ‘…cannot per se constitute a public policy exception within the meaning of Article 3(10)…’ of the original Directive, with Commission v Luxembourg cited as precedent for this; however, the amendment was ultimately rejected.Footnote ¹⁷⁴ Following this, informal contacts between Parliament, the Council, and the Commission resulted in an agreement on a compromise text in a vote held on 29 May 2018.Footnote ¹⁷⁵ The text was then forwarded to the Council for approval on 22 June.Footnote ¹⁷⁶

In the Council, the dossier was assigned to the Employment, Social Policy, Health and Consumer Affairs Committee in 2016.Footnote ¹⁷⁷ From the outset, the dossier was subject to tensions, the Social Questions Working Party (SQWP) observing that several delegations opposed negotiations, arguing the Commission’s work on the dossier was insufficient.Footnote ¹⁷⁸ Concerns were raised regarding the introduction of remuneration to the revised act.Footnote ¹⁷⁹ The Commission observed that it was a broad concept that would not require an explicit definition.Footnote ¹⁸⁰ The Polish Senate underlined that general interest restrictions should not go beyond what is necessary and that distinctions between posted and local workers relating to wage differences are not, stricto sensu, discriminatory.Footnote ¹⁸¹ Similarly, Latvia argued that the principle of equal pay for equal work needed to reflect that posted workers are not in the same position as local workers.Footnote ¹⁸² Estonia argued that such differences were a legitimate element of single-market competition.Footnote ¹⁸³ Croatia argued that the proposal contravened freedom of movement principles.Footnote ¹⁸⁴ There was, therefore, a clear pushback from Eastern Member States against the incorporation of the term remuneration.

The key theme tying these points together is the question of the degree of national autonomy in wage setting (via the concept of remuneration) that should be afforded to Member States. The Presidency and the Commission were convinced that the incorporation of remuneration into Article 3(1) would not affect national competences.Footnote ¹⁸⁵ A report issued by the Permanent Representatives Committee of the Council on the proposed Directive notes that one of the main issues raised during negotiations was whether the proposal adequately guaranteed the exercise of fundamental rights, in particular the freedom to strike, negotiate, conclude, and enforce collective agreements. This observation clearly points toward the impact of Viking and Laval.Footnote ¹⁸⁶ In late October, the General Secretariat released an amended text, with Recitals 7a–8 incorporating the Court of Justice’s jurisprudence,Footnote ¹⁸⁷ and paragraph 7 of Article 3 replaced by provision (ca), which reflects the ruling in Elektrobudowa. Specifically, the revised provision outlines that allowances specific to posting form part of the concept of remuneration unless those allowances are intended to reimburse expenditure related costs incurred related to the posting. This amendment successfully made its way into the final act.Footnote ¹⁸⁸ Recital 12a states that the ‘concept of “remuneration” should include, but should not be limited to, all the elements of minimum rates of pay developed by the Court of Justice …’.Footnote ¹⁸⁹ This approach is perhaps unsurprising. Pecinovsky, for example, considers the evolution of the Court of Justice’s Laval Quartet jurisprudence in the post-2009 era.Footnote ¹⁹⁰ Examining the Court’s rulings in Elektrobudowa and RegioPost, he argues that the Court took a more ‘socially friendly’ approach in its case law. In Elektrobudowa, for example, the Court of Justice noted that while Article 3(1) of the 1996 Directive allows host states to prescribe minimum wages, it leaves the components of this notion undefined; thus, the concept of minimum wages should be interpreted in accordance with host state law unless the national definition would impede the freedom to provide services.Footnote ¹⁹¹ Relatedly, in June 2017, changes to Article 3(10) were proposed to specify that the Directive would not prevent Member States from applying ‘terms and conditions of employment on matters other than those referred to in the first subparagraph of paragraph 1 in the case of public policy exceptions’ once done in a non-discriminatory way. Significantly, in 2018, the Council noted Article 3(8) had been amended to read, ‘[i]n the absence of, or in addition to, a system for declaring collective agreements or arbitration awards to be of universal application within the meaning of the first subparagraph, Member States may, if they so decide, base themselves on’ ... [emphasis added].Footnote ¹⁹² This changes the functioning of the 1996 framework considerably, representing a significant legislative amendment in response to the Laval ruling. It allows Member States to apply rules which fall outside the scope of the original Directive, as they did not form part of universally applicable domestic rules on minimum rates of pay. On 22 June, the Council approved the text by 22 votes. (Croatia, Latvia, Lithuania, and the United Kingdom abstaining, Poland and Hungary voting no).Footnote ¹⁹³

This overview of negotiations demonstrates that the Recast Directive reinforces the asymmetry of the 1996 framework, favouring higher-wage Member States by broadening mutual recognition via the shift toward remuneration in Article 3(1)(C). Eastern Member States repeatedly contest this, reflecting the persistent east–west divide noted by Kyriazi. She examined the political dynamics behind the revision of the 1996 framework and asked what the drivers are of this shift towards the ‘social’.Footnote ¹⁹⁴ Her work draws on media coverage and elite interviews to reconstruct the policy process. She concludes that transnational organisation at the vertical level, and relatedly to this, between horizontal levels, was an important part of the policy-building process that ultimately concluded with the Recast Directive becoming law. In particular, intergovernmental negotiation tactics between Western/old and Eastern/new European Member States governments triggered dynamic feedback loops where coalitions for and against the proposal developed along these geographical and historical lines.Footnote ¹⁹⁵ She shows how conflict over revising the dossier versus maintaining the existing framework is structured predominantly along territorial lines, between East and West, or between high-wage and low-wage economies.Footnote ¹⁹⁶ Bogoeski, however, views the Recast Directive as the result of a countermovement, led by the European trade union movement, against the Court of Justice’s case law. This reflects the findings of Louis. The revision of the Court’s interpretation of the 1996 framework can be seen, in Bogoeski’s view, as a form of symbolic rescue of the European Union’s liberal market economic approach and a rejection of market access as the Union’s understanding of social justice.Footnote ¹⁹⁷ Bogoeski argues that the Court’s interpretation of the law creates a fictitious commodification of work that artificially seeks to subject labour to the rationality of the market. This, by extension, undermines the underlying inter-dependent and organic dynamics of ‘the social whole’, by dressing these arrangements in the cloak of contractual interactions which effectively ignore the underlying social relations.Footnote ¹⁹⁸

From a longitudinal perspective, Zahn argues that the posted workers Directives are examples of failed attempts to Europeanise applicable national labour rules. Both Directives attempt to promote cross-border service provision, while also seeking to ensure fair competition and respect for workers’ rights within the internal market.Footnote ¹⁹⁹ She notes that the Commission’s proposal for the Recast Directive was a step in the right direction, in particular as the inclusion of the term ‘remuneration’ allows ‘the inclusion of a variety of different elements as part of a pay package and give social partners some discretion in bargaining over pay and its constituent parts’.Footnote ²⁰⁰ However, in the sectoral pay context, the framework continues to only recognise collective agreements of a universally applicable nature. The result is that countries such as Sweden, where such agreements are not widely used, will continue to face issues related to wage equality.Footnote ²⁰¹ Rizzi has observed the longitudinal effects of Viking and Laval as a provocation for further positive integration in this area. A consequence of the Court’s rulings is a highlighting of underlying issues within the Union’s legal and constitutional system regarding the balance between fundamental rights and economic freedoms. This prompted both national and supranational responses to address these asymmetries at both the horizontal and vertical levels. Horizontal level responses, including Member State adaptation of domestic labour rules in response to case law, leading to the emergence of more coherent social policies at the European level. Equally, case law triggered distinctively European responses via the institutions as shown herein.Footnote ²⁰² Taken together, the empirical analysis demonstrates that the Recast Directive preserves the structural asymmetries of rule application via Article 3(1)’s mutual recognition mechanism. It reflects a more socially friendly turn in that the inclusion of the notion of remuneration under Article 3(1)(C) allows a wider array of domestic rules to be applied to posted workers. Yet the framework continues to favour the application of labour rules from higher protection jurisdictions. Moreover, it continues to rely on procedural mechanisms which facilitate the mutual recognition of certain ‘types’ of rules between Member States without resolving the deeper divergences in national labour systems regarding the substantive content of labour rules that led to so much of the tension in the context of the cross-border posting of workers.