Article 1(2) PSD and the EU Concept of Abuse: Clarification, Continuity, and Causation in Nordcurrent group

2.1 Different General Anti-Avoidance Measures in EU Tax Law

EU tax law contains a range of legal mechanisms that are grounded in the concept of abuse. For the sake of simplicity, these mechanisms may be referred to collectively as general anti-avoidance measures, as they articulate abuse in broad, principle-based terms and empower state authorities either to deny the granting of tax benefits in abusive situations or to disregard abusive arrangements when determining tax liability. Although they differ in origin and legal form, all such mechanisms serve the common purpose of preventing taxpayers from improperly exploiting EU law.

In the EU legal order, general anti-avoidance measures exist in two principal forms. The first is the general principle prohibiting abuse of rights, a judge-made doctrine developed in the Court of Justice’s case law.3 The second consists of statutory general anti-avoidance rules (GAARs) laid down in the directives on direct taxation.4

The coexistence of these measures has given rise to significant conceptual and practical complications. The general principle of prohibition of abuse of rights serves, on the one hand, as a ground of justification that allows Member States to enforce domestic anti-avoidance rules even where these restrict the fundamental freedoms. On the other hand, the principle also operates directly within the scope of EU secondary law, enabling Member States or courts to deny benefits conferred by directives when the underlying arrangement is abusive. In this latter capacity, the principle applies to abuse of the VAT Directive but also overlaps in scope with the statutory GAARs contained in the directives on direct taxation. This coexistence raises questions concerning hierarchy, coordination, and the extent to which the Court’s case-law-based abuse test should inform, supplement, or be constrained by the statutory definitions of abuse laid down by the EU legislature.

Over the last decades, both the Court of Justice of the European Union and the European Council have significantly developed and refined these general anti-avoidance measures. These measures appear to be governed by a set of increasingly detailed conditions formulated by the two institutions.

Their scope has likewise expanded, extending the reach of EU anti-abuse doctrine. Contrary to arguments advanced in earlier academic commentary and by several Member States, the Court has made clear that the principle of prohibition of abuse of rights operates not only as a justification for national restrictions on the fundamental freedoms and in the field of VAT, but also as a basis for denying benefits conferred by secondary EU law within the sphere of direct taxation.5

The EU legislature has followed a similarly expansive path. With the adoption of Article 6 ATAD, the European Council introduced a statutory GAAR that applies not only to benefits grounded in EU law but also to those arising under domestic corporate tax systems. The provision thereby extends the EU’s general anti-abuse framework into the internal legal orders of the Member States, regulating situations that, in an earlier period, would have been regarded as purely domestic.

A further development can be seen in the fact that both institutions have moved from permitting Member States to counter abusive conduct to requiring them to do so. The Court’s case law now establishes that Member States must disallow abusive use of EU law, and Article 6 ATAD obliges them to apply a general anti-avoidance rule meeting the minimum standard laid down in the directive.6 Together, these evolutions reflect a decisive shift: general anti-avoidance measures are no longer optional or supplementary tools but are instead integral and mandatory components of the EU legal order’s approach to combating tax avoidance.

2.2 Doctrinal and Terminological Uncertainty

The notion of abuse in EU tax law has long been marked by conceptual uncertainty, particularly concerning the precise conditions that must be fulfilled for an arrangement to be classified as abusive. Over time, the Court of Justice has elaborated different formulations of abuse, giving rise to what can broadly be described as two court-made tests within the framework of the general principle prohibiting abuse of rights.

Originally in the context of cases involving potential abuse of secondary law, the Court developed what is often referred to as the abuse test, which consists of two cumulative elements: an objective and a subjective component. The objective element requires that the purported use of a tax benefit runs counter to the aim or purpose of the secondary legislation that grants that benefit. The subjective element demands that the intention to obtain a tax advantage plays a sufficiently significant role in the design or execution of the arrangement. The test thus combines purposive legislative interpretation with an inquiry into the taxpayer’s underlying motivations.7

Within the context of the fundamental freedoms, the Court adopted a different but related standard: the wholly artificial arrangement test. While this test similarly incorporates a form of subjective assessment relating to the taxpayer’s intention to obtain a tax benefit, it also features an additional requirement—namely, that the arrangement must lack economic or commercial substance to such an extent that it can be considered artificial. This artificiality criterion has been central to delineating the boundary between legitimate tax planning and abusive conduct.

When viewed together, these court-developed tests—and the statutory GAARs introduced by the European Council—exhibit a number of common conceptual elements. In varying formulations, they tend to refer to:

1. a distinction between genuine or real arrangements and those considered artificial or non-genuine;

2. a requirement that the taxpayer has, to a meaningful degree, the intention of obtaining a tax advantage; and

3. a condition that the granting of the benefit in question would defeat the aim or purpose of the legislation at issue.

However, there are considerable variations between general anti-avoidance measures and tests in regards to which of these textual elements appears. Further, the Court has used different adjectives in its formulation of these textual elements.

The diverging formulations found in both the Court-developed abuse tests and the statutory GAARs have led many commentators to assume that these linguistic differences must carry substantive legal significance—namely, that they establish distinct thresholds or requirements for finding abuse. This textualist reading has been particularly prominent in the academic literature on the concept of abuse in EU tax law, where differences in wording are often treated as intentional signals of different legal standards. According to this view, the threshold of abuse is lower when it comes to more harmonized fields of tax law compared to the concept of abuse concerning the freedoms of movement.8

A contrasting strand of scholarship argues that the Court has not treated the various formulations as separate or autonomous tests. On the contrary, the Court has often linked differently worded standards to one another, drawing connections across lines that some commentators have regarded as doctrinally separate. In this view, the variations in wording reflect the evolution of the case law rather than the existence of rigidly distinct legal tests.9

The statutory GAAR in Article 1(2) PSD, which formed the basis of the Nordcurrent group judgment, provides:

Member States shall not grant the benefits of this Directive to an arrangement or a series of arrangements which, having been put into place for the main purpose or one of the main purposes of obtaining a tax advantage that defeats the object or purpose of this Directive, are not genuine having regard to all relevant facts and circumstances.

An arrangement may comprise more than one step or part.

Article 1(3) then states the following:

For the purposes of paragraph 2, an arrangement or a series of arrangements shall be regarded as not genuine to the extent that they are not put into place for valid commercial reasons which reflect economic reality.

When the provision was adopted—and similarly, when Article 6 ATAD was introduced—it was argued that these rules did not conform to the standard used by the Court when assessing whether anti-avoidance measures could be justified under the general principle prohibiting abuse of rights. The criticism centred on the view that the statutory GAARs appeared to introduce a lower threshold of intention, allowing a finding of abuse where the tax advantage was merely one of the main purposes, rather than relying on the stricter standard associated with the wholly artificial arrangement test developed under the fundamental freedoms.10

The meaning and importance of the various textual elements—“non-genuine”, “artificial”, “main purpose”, the function of intention, or the requirement that a benefit “defeats the purpose” of the legislation—have always been uncertain. The Court has offered limited guidance on how these elements relate to specific factual indicators, and its case law has been inconsistent, often mixing the vocabulary of different tests within a single judgment. As a result, the academic literature is itself divided: interpretations of these concepts differ widely, and there is no consensus on how central notions such as artificiality, genuineness, the taxpayer’s aim, and the purpose of the legislation interact with each other.11

Because the conditions for establishing abuse remain conceptually unclear, the academic literature has often approached the notion of abuse by categorising it in terms of specific manifestations previously recognised in the case law. As a result, scholars—and at times even the Court itself—have tended to equate “abuse” with particular forms of abusive behaviour rather than with a general legal standard. Consequently, the assessment of whether the presence of a company in a given Member State is abusive has, in some contributions, been reduced to an inquiry into its physical presence in that state.12 Likewise, the determination of whether a transaction constitutes abuse has at times been framed as depending solely on whether it was concluded at arm’s length, thereby narrowing the analysis to a single economic indicator rather than engaging with the broader purposive assessment required by the abuse doctrine.13

A further question that has attracted considerable attention concerns the extent to which it is necessary to establish a clear and specific link between an arrangement deemed to satisfy the conditions for abuse and the particular tax benefit obtained. This issue became especially prominent in the wake of the Court’s judgments in the so-called Danish Beneficial Ownership Cases, which addressed alleged abuse of the Parent-Subsidiary Directive and the Interest and Royalty Directive.14 In those cases, the Court was confronted with the question whether potential abuse could be excluded where the economic reality of the arrangement would, in any event, have entitled the taxpayer to the very benefit that was allegedly being abused. The Court’s reasoning on this point was markedly opaque and has been read by some commentators as suggesting that such factual circumstances are irrelevant to the abuse assessment, thereby raising further questions as to how closely the benefit must be causally connected to the abusive arrangement.15

Another view, however, is that the judgments should primarily be understood to mean that such factual circumstances may indeed be relevant and capable of discounting the possibility of abuse, but that it is ultimately for the taxpayer to substantiate this by providing adequate evidence. On this interpretation, the judgments in the Danish cases do not render the economic outcome of the arrangement irrelevant but rather emphasise that the burden of demonstrating its significance rests firmly with the taxpayer.16

4.1 The Court’s Reasoning and the Conditions of Abuse under Article 1(2) PSD

The Court’s reasoning in Nordcurrent group continues a long-running tendency in the CJEU’s case law to link together different tests for abuse that have emerged in various contexts of EU tax law. Over the past two decades, both the Court and the European Council have articulated conditions of abuse in multiple formulations—within the general principle prohibiting abuse of rights, in the wholly artificial arrangement test under the fundamental freedoms, and in statutory GAARs. Taken at face value, these formulations differ in structure, terminology, and emphasis. Yet, in practice, the Court has persistently treated them as manifestations of a single, overarching concept of abuse, drawing connections across doctrinal lines that a purely textual reading would keep separate.22

This pattern is clearly visible in Nordcurrent group. Although Article 1(2) PSD is framed in terms that differ from the two-step abuse test developed under the general principle—and from the wholly artificial arrangement test established in the freedoms jurisprudence—the Court firmly situates the statutory GAAR within that broader conceptual framework. When interpreting the conditions of Article 1(2) PSD, the Court expressly referred back to its reasoning in the Danish Beneficial Ownership Cases, where it had already linked the concept of abuse in the directive to the abuse concept developed in the context of the freedoms of movement.23 In those judgments, the Court had stated that the notion of abuse under the directives had “the same scope” as the notion of abuse underlying the wholly artificial arrangement test.24 In Nordcurrent group, the Court followed the same interpretative trajectory, citing its case law on the concept of abuse in the context of the freedoms of movement when clarifying the meaning and structure of Article 1(2) PSD.25

This cross-referencing confirms that a textualist approach to the different formulations of abuse has given rise to an unnecessarily fragmented picture of the EU concept of abuse. Despite differences in wording, the Court has consistently sought to fuse together the various tests and statutory provisions, implicitly treating them as doctrinally compatible expressions of a common principle. Nordcurrent group reinforces this interpretative method and demonstrates, once again, the Court’s reluctance to read textual divergences as signalling substantive differences in legal standards.

The process of harmonising these formulations has never been linear or elegant in the Court’s case law.26Nordcurrent group illustrates this methodological messiness. Early in its reasoning, the Court describes Article 1(2) PSD as containing two cumulative conditions:

1. the existence of a non-genuine arrangement; and

2. that the arrangement has been put into place with the main purpose or one of the main purposes of obtaining a tax advantage that defeats the objective of the directive.27

When compared to the two-step abuse test developed under the general principle, this structure appears at first to merge the traditional objective and subjective elements into one combined purposive requirement, while expressing non-genuineness as an independent condition. A reader adopting a strictly textual interpretation might therefore be tempted to infer that Article 1(2) PSD embodies a narrower or differently structured concept of abuse than the concept developed in the Court’s general case law. The separation of “non-genuine” and “purpose of obtaining a tax advantage” could be construed as introducing a distinct evaluative sequence, or as indicating that the factual indicators relevant to non-genuineness carry a different legal meaning than the indicators used under the general principle.

However, this is not the direction in which the Court ultimately proceeds. Instead, Nordcurrent group exhibits the same interpretative pattern that characterises the Danish cases: after setting out the statutory wording, the Court links Article 1(2) PSD back to its earlier case law, treating differently formulated tests as part of a unified doctrinal framework. It explicitly states that the assessment of abuse under Article 1(2) PSD “corresponds” to its previous jurisprudence and directly cites the abuse test applied in the Danish Beneficial Ownership Cases. It also draws a connection to its guidance in Cadbury Schweppes regarding the identification of abusive structures in the context of the fundamental freedoms.28 By doing so, the Court signals that the statutory GAAR cannot be understood in isolation or interpreted solely on the basis of its textual formulation.

Seen in this light, Nordcurrent group confirms the Court’s long-standing practice of treating the conditions of abuse—irrespective of the terminology employed in a given legal instrument—as variations of the same underlying concept. Textual differences are not interpreted literally but absorbed into a broader jurisprudential framework in which the general principle of prohibition of abuse of rights continues to act as the doctrinal anchor. It therefore appears safe to conclude that the concept of abuse under Article 1(2) PSD is to be evaluated in substantially the same manner as abuse under the general principle.

This interpretative approach inevitably perpetuates the terminological confusion surrounding the precise conditions that must be met for an arrangement to be regarded as abusive. Yet, as Nordcurrent group illustrates, such confusion arises not because the Court intends to introduce separate legal standards, but because it consistently subordinates textual formulations to its overarching—and increasingly harmonised—understanding of abuse in EU tax law.

4.2 The Court’s Broad Approach to the Scope of the Abuse Evaluation

A notable feature of the Court’s reasoning in Nordcurrent group is its refusal to restrict the scope of the abuse assessment to specific categories of factual circumstances or to any fixed temporal point. Instead, the Court emphasises that the evaluation under Article 1(2) PSD must remain open, comprehensive, and adaptable. This reflects a broader interpretative mindset: Member State authorities should not exclude any type of factual element that may either support or undermine a finding of abuse.

This interpretative openness manifests itself in two key aspects of the judgment.

First, the Court declined to confine the notion of abuse to arrangements involving companies lacking physical presence that receive income and promptly pass it on within a group—the classic “conduit company” scenario. While such situations had figured prominently in the Danish Beneficial Ownership Cases, the Court rejected the view that Article 1(2) PSD is limited to these particular forms of artificiality. By doing so, the Court avoided reducing the assessment of abuse under the directive to the specific manifestations it had previously considered. Instead, it clarified that the framework of abuse under the PSD encompasses a wider range of potentially artificial arrangements, thereby preventing a restrictive or categorical reading of its earlier case law.29

Secondly, the Court resisted limiting the abuse evaluation to the moment an arrangement is put into place. Although the wording of Article 1(2) PSD might suggest a focus on the purpose underlying the arrangement at the time of its implementation, the Court rejected such a temporal constraint. It held that an arrangement initially adopted for legitimate commercial reasons may become non-genuine at a later stage if maintained despite a change in circumstances. Likewise, factual circumstances arising at a time other than the moment of the relevant transaction may be relevant to determining whether an arrangement is abusive. The Court thus made clear that the abuse assessment must account for the evolution of economic and organisational realities over time.30

These two aspects of the reasoning point towards a general principle: the Court is unwilling to exclude any factual circumstances in advance from the scope of the abuse evaluation. This approach aligns with the intrinsic nature of general anti-avoidance measures, which are designed to be broad, flexible, and capable of addressing forms of avoidance that could not have been foreseen or exhaustively identified by the legislature. At the same time, this openness comes with a drawback. By declining to specify which types of facts should be determinative in the assessment, the Court provides limited guidance to tax authorities and courts on how the statutory GAAR is to be applied in future cases. The result is a framework that is adaptable, but in some respects indeterminate, leaving substantial discretion at the national level when interpreting and applying Article 1(2) PSD.

4.3 The Required Link Between Non-Genuineness and the Attainment of a Tax Advantage

The Court’s reasoning in Nordcurrent group confirms that a finding of non-genuineness, understood as the presence of artificial or commercially unmotivated elements within an arrangement, is not in itself sufficient to establish abuse under Article 1(2) PSD. Rather, the Court underscores that there must be a demonstrable link between the non-genuine steps or aspects of the arrangement and the attainment of a tax advantage. In other words, the artificial elements must serve, in a meaningful way, the purpose of securing the tax benefit alleged to be abusive.

Importantly, the Court frames this required nexus as part of the subjective element of the abuse test: it is not enough that the arrangement features elements lacking genuine commercial rationale; those elements must be connected to a tax-motivated purpose that defeats the object or purpose of the directive.31 In this respect, the Court’s approach aligns with the underlying logic of the general principle prohibiting abuse of rights, where the taxpayer’s aim to obtain a tax advantage inconsistent with the purpose of the relevant rules is central to the determination of abuse.

The judgment also aligns with observations previously made in the academic literature. The requirement of a link between the non-genuine elements of an arrangement and the tax advantage pursued has been recognised as a necessary component of the abuse analysis in several strands of the Court’s case law. This is perhaps most evident in the VAT context, where the Court has consistently held that artificial elements must be directed towards obtaining a tax benefit in order to constitute abuse. A similar understanding underlies the case law on the freedoms of movement and on the application of secondary legislation.32

As discussed in Section 2, the Court’s reasoning in the Danish Beneficial Ownership Cases prompted considerable academic debate as to whether it is sufficient for tax authorities to demonstrate that the arrangement is non-genuine and that it gives rise to the benefit stipulated in the directive, or whether the authorities must also consider whether that benefit would nevertheless arise if the arrangement had been structured in accordance with economic reality. The Court’s reasoning in Nordcurrent group provides clarity in this respect. It makes clear that a broader perspective is required: national authorities must assess not only whether the arrangement triggers the benefit laid down in the directive, but also the overall tax effect of the structure in the Member State concerned.

This broader approach is illustrated by the Court’s reliance on the fact that the profits of the UK subsidiary were subject to a higher corporate tax rate in the United Kingdom than would have applied in Lithuania. The Court treated this as a relevant factor indicating that the arrangement may not have been designed for the purpose of obtaining a tax advantage. The implication is that the abuse analysis cannot be confined to the benefit conferred by the directive in isolation; rather, it must encompass the arrangement’s full tax consequences. If the overall effect undermines the notion that the arrangement was designed to secure a tax advantage, the subjective element of abuse may not be satisfied.

Taken together, Nordcurrent group reinforces the principle that the artificial aspects of the arrangement must plausibly serve the purpose of securing a tax advantage. This requirement ensures that the abuse test does not collapse into a mere inquiry into non-genuineness but remains rooted in the purposive rationale that underlies the EU concept of abuse.