The Implementation in Sweden of the General Anti-Avoidance Rule of the Anti-Tax Avoidance Directive
Article Category: Article
Published Online: Apr 22, 2024
Page range: 50 - 59
Received: Jun 14, 2022
Accepted: Nov 24, 2022
DOI: https://doi.org/10.2478/ntaxj-2023-0002
Keywords
© 2023 David Kleist, published by Sciendo
This work is licensed under the Creative Commons Attribution 4.0 International License.
The Anti-Tax Avoidance Directive (ATAD)1 was adopted on 12 July 2016 and requires member states to implement legally binding anti-avoidance measures in several areas to prevent certain forms of corporate tax planning. The measures include a General Anti-Avoidance Rule (GAAR) and four specific anti-avoidance rules. The ATAD applies to taxpayers that are subject to corporate tax2 and shall have been given effect in the member states by 1 January 2019.3 The focus of this article is the GAAR of ATAD and its implementation in Sweden, which raises a number of interesting questions.
According to the Swedish Government, the Swedish GAAR that was already in place when the ATAD was adopted sufficiently implements the GAAR of the ATAD. That assessment is by no means self-evident, as there are clearly differences between them. In addition to questions of European Union (EU) law compatibility raised by these differences, the “passive” implementation in Sweden of the GAAR of the ATAD raises questions concerning the extent to which EU law will affect its interpretation. Historically, the interpretation of the Swedish GAAR has developed over time in an entirely Swedish legal context, but that might change as a result of the Swedish GAAR’s new function as an implementation of the GAAR of the ATAD.4 Thus, the purpose of this article is to discuss the Swedish GAAR’s new function as an implementation of the GAAR of the ATAD, and what it means for the influence of EU law on its interpretation, specifically with regard to the differences that exist between the GAARs.
This article includes a doctrinal analysis of legislation and case law, meaning a systematic exposition of the rules and precedents governing the subject matter, and an analysis of their meaning and of the relationship between them according to accepted discipline standards and norms. The relevant legislation in this case is mainly the Treaty on the Functioning of the European Union (TFEU)5 and the ATAD. There are also numerous cases on the interpretation of the TFEU, which are important for the analysis presented in this article.
There is much to be said in regard to the interpretation of the TFEU and other EU law, but in this context, I limit myself to making a few comments. First, there is no provision in the TFEU that gives precedence to any particular method of interpretation. Consequently, literal interpretation, contextual interpretation, and teleological interpretation are all available to the Court of Justice of the EU (CJEU). Although the principle of legal certainty requires the CJEU to not depart from the clear and precise wording of an EU law provision, a literal interpretation is often not sufficient to determine the meaning of a provision, especially in light of the fact that the TFEU is drafted in broad terms and is characterized by a purpose-driven functionalism that limits the possibilities of a textualist approach. Second, where linguistic divergences arise as a consequence of the fact that the legislation is adopted in several equally authentic language versions, the CJEU may not give priority to one linguistic version over the others but must interpret the EU law provision in question in light of the normative context in which it is placed and the objectives it pursues (Lenaerts and Gutiérrez-Fons, 2013). Thus, contextual and teleological interpretation is bound to play a more prominent role and literal interpretation a correspondingly less significant role than when, for instance, Swedish national legislation is being interpreted.
Consequently, the methods for interpretation of national legislation may differ from the methods used for interpreting EU law. However, this difference is mitigated by the fact that national legislation, as emphasized later, must, to a considerable extent, be interpreted in conformity with EU law, meaning that the methods employed for interpreting EU law could, indirectly, become relevant for interpreting national legislation, too.
Following this short introduction, an account is given of the relation between EU law and legislation on corporate taxation in general. I then present and compare the GAAR of the ATAD and the Swedish GAAR. I then discuss the Swedish GAAR’s compatibility with the ATAD, including the Swedish Government’s assessment in this regard. Finally, I analyze the impact of EU law on the application of the Swedish GAAR.
There is no separate legal basis in the TFEU for harmonization in the area of direct taxation. Directives on income tax are, therefore, adopted with reference to Article 115 of the TFEU, which states that “the Council shall issue directives for the approximation of such laws, regulations or administrative provisions of the Member States as directly affect the establishment or functioning of the internal market.” Consequently, the ATAD was adopted on this basis.
It follows from Article 3 of the ATAD that the GAAR and the other anti-avoidance rules in the ATAD set a minimum standard and that the ATAD “shall not preclude the application of domestic rules aimed at safeguarding a higher level of protection for domestic corporate tax bases.” In other words, the GAAR of the ATAD can be seen as a “floor” for the implementation of national GAARs—a more lenient approach would not fulfill the requirements of the ATAD—but it does not preclude member states from implementing stricter GAARs.
On the other hand, primary EU law sets a “ceiling” for national GAARs, as it follows from the case law of the CJEU that in situations where a freedom of the TFEU might be triggered, national anti-tax avoidance measures shall be allowed effect only if they specifically relate to wholly artificial arrangements aimed at escaping the national tax normally payable.6 As pointed out by Schön (2020), the jurisprudence of the CJEU in this regard can be interpreted in two different ways. First, artificiality can be perceived as relevant for the justification of a restriction. Under this interpretation, national anti-avoidance measures restricting any of the four freedoms of the TFEU may be justified only if they specifically relate to wholly artificial arrangements aimed at escaping the national tax normally payable. Accordingly, such tax measures must not be applied where the taxpayer, regardless of the existence of tax motives, carries on genuine economic activities. Second, artificiality can be seen as relevant for determining the scope of the fundamental freedoms. According to this interpretation of case law, the setting up of a wholly artificial arrangement falls outside the scope of the four freedoms altogether, so that the taxpayer is not protected by them. In this case, too, anti-tax avoidance measures must not be applied where the taxpayer, regardless of the existence of tax motives, carries on genuine economic activities.
Either way, it is clear that primary EU law can require the nonapplication of a national anti-tax avoidance measure to an arrangement that is not wholly artificial, regardless of whether that measure has been adopted by the national legislator without involvement of EU institutions or implements EU secondary law. However, provided that a national measure is applied in a nondiscriminatory manner and does not impede free movement or restrict access to the market of another member state, it would not be considered in breach of the four freedoms.
The jurisprudence of the CJEU dealing with the compatibility of national income tax rules with primary EU law has for the most part dealt with the four freedoms of the TFEU. However, if national anti-tax avoidance measures are implementing EU law or if the situation is otherwise governed by EU law, such measures must also comply with the fundamental rights codified in the Charter of Fundamental Rights.7 Furthermore, if the case falls within the scope of EU law, national law must comply with general principles of EU law, which are also part of primary EU law.8 The general principles of EU law include such fundamental rights as are guaranteed by the European Convention for the Protection of Human Rights and Fundamental Freedoms or result from the constitutional traditions common to member states.9 Among those general principles of EU law are, according to the CJEU, proportionality, legal certainty, legitimate expectations, equality, and procedural justice (Craig and de Búrca, 2020).
Directives are binding as to the result to be achieved but leave it to the national authorities to choose the form and methods to achieve that result.10 Thus, there is no obligation to implement legislation that contains the same wording as the ATAD or to introduce new legislation at all, as long as the objectives of the ATAD are achieved.
In this regard it is important to keep in mind that there is, according to the CJEU, a far-reaching obligation on the part of member states (and thus on the part of the courts and authorities within them) to ensure that the result sought by a directive can be achieved. Domestic law must, therefore, be interpreted in light of the directive’s wording and purpose.11 This obligation also exists, according to the CJEU, when the national legislation predates the directive, as is the case with the Swedish GAAR.12 Although directives might be invoked more frequently than other forms of EU legislation, the obligation to interpret national legislation in conformity with EU law concerns all EU law, primary as well as secondary. Thus, regardless of the implementation strategy chosen by Sweden, potentially such “harmonius interpretation” of the Swedish GAAR could ensure that the implementation does not fall short of the minimum requirements of the ATAD and that it does not conflict with primary EU law.
After this brief account of some central aspects of the relation between EU law and national legislation on corporate taxation, we are now ready to move on to the GAAR of the ATAD and the Swedish GAAR.
According to Article 6.1 of the ATAD, a member state shall, for the purposes of calculating corporate tax liability, “ignore 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 the applicable tax law, are not genuine having regard to all relevant facts and circumstances.” The paragraph also points out that an arrangement may comprise more than one step or part.
In regard to the assessment of whether an arrangement is genuine or not, it follows from Article 6.2 that “an arrangement or a series thereof shall be regarded as non-genuine to the extent that they are not put into place for valid commercial reasons which reflect economic reality.”
Finally, Article 6.3 provides that the consequence of an arrangement being ignored in accordance with Article 6.1 is that the tax liability shall be calculated in accordance with national law.
Clearly, the anti-abuse doctrine found in primary EU law and developed by the CJEU—most famously in the previously mentioned
The scope of the GAAR of the ATAD is not limited to cross-border situations. It is, according to the eleventh recital of the ATAD, important “to ensure that the GAARs apply in domestic situations, within the Union and
The current Swedish GAAR—
According to the Swedish GAAR, a transaction shall be disregarded for tax purposes if (1) it, on its own or in combination with other transactions, is part of an arrangement that results in a substantial tax benefit for the taxpayer; (2) the taxpayer has, directly or indirectly, participated in the transaction, (3) the obtaining of the tax benefit, taking into account the circumstances, can be assumed to have been the main purpose of the arrangement, and (4) a tax assessment on the basis of the arrangement would be contrary to the purpose of the tax legislation, determined on the basis of the formulation of the tax legislation as a whole and the provisions that are applicable or have been circumvented in the particular case.17
The question of whether the Swedish GAAR is applicable in a particular case has to be decided by an administrative court following an application by the Swedish Tax Agency.18 Thus, the Tax Agency is not competent to decide on its own whether the Swedish GAAR applies.
If the Swedish GAAR applies, the tax liability shall, as a main rule, be determined as if the arrangement had not been entered into or carried out. Thus, typically the same consequence applies as under the GAAR of the ATAD. However, the Swedish GAAR also provides for alternative consequences. If the arrangement, taking into account the economic result excluding the tax benefit, appears to be a detour in comparison with the transactions that are closest at hand, the tax assessment shall instead be made as if the taxpayer had chosen those transactions. If the mentioned grounds for tax assessment cannot be applied or would lead to an unreasonable result, the tax liability shall be determined based on what is reasonable.19
It is not the purpose of this article to analyze in detail the differences between the GAARs. However, some important differences will be pointed out as they highlight the relevance of analyzing the influence of EU law on the Swedish GAAR.
The Swedish GAAR and the GAAR of the ATAD share some common traits, but there are also significant differences. For instance, the fact that the Swedish GAAR applies to the income tax assessment for municipal and state tax purposes means that it applies to income tax in general and that its scope, in contrast with that of the ATAD, is not limited to taxpayers that are subject to corporate tax. Thus, in this regard, the Swedish GAAR is wider in scope and applies to individuals, too.
It also follows that there is a procedural requirement for the Swedish GAAR that has no equivalent under the ATAD—the Swedish Tax Agency may submit an application to an administrative court concerning the application of the GAAR, but it is the court that decides whether to apply it. In other words, it is not sufficient that the Swedish Tax Agency considers the requirements of the Swedish GAAR to have been met.
Another difference relates to the consequences of applying the GAAR. The GAAR of the ATAD provides that the arrangement in question shall be ignored, whereas the Swedish GAAR also provides for other outcomes, such as taxation based on alternative transactions or on what is reasonable.
Based on this account, we can also note a couple of differences in the formulation of the prerequisites that could result in differences in the application. It follows from the GAAR of the ATAD that a member state shall ignore arrangements that have been put into place “for the main purpose or one of the main purposes of obtaining a tax advantage.” The Swedish GAAR requires that “the obtaining of the tax benefit … can be assumed to have been the main purpose of the arrangement.” This difference in wording gives the impression that the threshold for applying the Swedish GAAR is higher.
Another notable difference is that the GAAR of the ATAD applies to arrangements that are “not genuine,” whereas no such requirement follows from the wording of the Swedish GAAR. In this regard, the threshold for applying the Swedish GAAR seems to be lower.
On 30 August 2018, the Swedish Government presented a bill to the Swedish parliament concerning the implementation of the controlled foreign company (CFC) rules of the ATAD into Swedish law. In this bill, which exceeds 100 pages, the Government also presents its view on the implementation of the GAAR of the ATAD in an assessment that covers less than four pages.20 As mentioned earlier, the Government concludes that the existing Swedish GAAR implements the GAAR of the ATAD.
According to the ATAD, a prerequisite for application of the GAAR is that the arrangement in question is not genuine; that is, that the arrangement has not been put into place for valid commercial reasons that reflect economic reality. No such requirement is included in the Swedish GAAR, which according to the Swedish Government means that the Swedish GAAR is wider in scope.21 Although the difference is pointed out by the Swedish Government, it is not commented on in any way. Presumably, the wider scope of the Swedish GAAR is not seen as a problem when it comes to the implementation of the GAAR of the ATAD, as provisions aimed at safeguarding a higher level of protection for domestic corporate tax bases are allowed under the ATAD.
Further, the Swedish Government points out that the Swedish GAAR requires participation by the taxpayer. Although the GAAR of the ATAD does not include a similar requirement, it is, according to the Government, implicit in the GAAR of the ATAD that the taxpayer has participated.22 No further comment is given by the Swedish Government in this regard.
The Swedish GAAR requires that an arrangement results in a “substantial” tax benefit and not just in any tax benefit. No corresponding express requirement applies under the GAAR of the ATAD. However, according to the Swedish Government, it can be questioned whether it would be proportional to apply a GAAR if the tax benefit falls below the threshold set by the Swedish GAAR. Against this background, the Government concludes that in this regard, too, the GAAR of the ATAD is covered by the Swedish GAAR, presumably meaning that the scope of the GAAR of the ATAD is narrower and fits within the scope of the Swedish GAAR.23
Moreover, the Swedish Government refers to Article 6.1 of the ATAD, which requires that an arrangement has been put into place “for the main purpose or one of the main purposes of obtaining a tax advantage” for the GAAR to apply, whereas the corresponding requirement under the Swedish GAAR is that the tax benefit is “the main purpose” of the arrangement, meaning that the tax benefit outweighs all other reasons for the actions undertaken by the taxpayer.24 Without commenting on this deviation, and despite the fact that the threshold for application of the Swedish GAAR can be perceived as being higher than under the GAAR of the ATAD, the Swedish Government goes on to conclude that in this regard, too, the GAAR of the ATAD is covered by the Swedish GAAR.
Finally, a prerequisite for application of the GAAR of the ATAD is that the tax advantage that the taxpayer seeks to obtain “defeats the object or purpose of the applicable tax law,” whereas the Swedish GAAR requires for its application that “a tax assessment on the basis of the arrangement would be contrary to the purpose of the tax legislation, determined on the basis of the formulation of the tax legislation as a whole and the provisions that are applicable or have been circumvented in the particular case.” The condition set forth in the Swedish GAAR conforms, according to the Swedish Government, with the condition stated in the GAAR of the ATAD.25
The Swedish Government also pays attention to the differences in regard to the tax consequences of applying the GAARs. As mentioned earlier, the consequence according to the GAAR of the ATAD is that the transactions in question shall be disregarded. Typically, that is the case under the Swedish GAAR, too, but it also provides for alternative consequences. The Swedish Government laconically concludes that the Swedish GAAR covers the GAAR of the ATAD.26
In some respects, the Swedish Government points out differences between the GAARs and then jumps to the conclusion that the Swedish GAAR “covers” the GAAR of the ATAD and that no new legislation is needed, without really explaining why. The Government might be right in its conclusion, but the arguments for that conclusion are not presented in a clear manner.
If a new Swedish GAAR that followed the wording of the GAAR of the ATAD had replaced the existing Swedish GAAR, the legal precedents on the Swedish GAAR would have become obsolete. It will take time for new case law relating to the GAAR of the ATAD to develop, particularly in light of the fact that the CJEU has the final say. The absence of applicable precedents would have increased legal uncertainty in a field that is already notoriously difficult for practicing lawyers to handle. It is possible that the Swedish Government was reluctant to change the existing GAAR for this reason. This argument seems to have been important, at least, for the Finnish Government’s decision to rely on the existing Finnish GAAR (Scherleitner, 2019). However, as the Swedish Government did not explain the reasoning behind its choice to keep the existing Swedish GAAR, it is hard to know.
Alternatively, it would have been possible, as a middle course, to keep the existing Swedish GAAR and make minor adjustments to it where significant differences between it and the GAAR of the ATAD exist. That way, most of the precedents on the Swedish GAAR would have remained relevant. However, in defense of the position taken by the Swedish Government, it should be pointed out that there is—as mentioned previously—a far-reaching obligation to interpret the Swedish GAAR in light of the ATAD’s wording and purpose. Thus, it might be possible to apply the Swedish GAAR in conformity with the GAAR of the ATAD, in spite of the differences. This is discussed in the following section.
As described previously, national laws shall as far as possible be interpreted in a manner that is consistent with EU law. Thus, legislation that implements a directive must be interpreted in light of the directive’s wording and purpose to ensure that the result sought by the directive can be achieved. This obligation also exists when a national legislative act predates the directive and has no specific connection to it. Thus, there is an obligation to interpret the Swedish GAAR in light of the ATAD’s wording and purpose, also taking into account primary EU law. This obligation to interpret national legislation in conformity with EU directives and other EU law is sometimes referred to as
However, there are limits to this obligation. National laws do not have to be interpreted in a way that violates general principles of law (particularly those of legal certainty and nonretroactivity), is contrary to the wording of the provisions in question, or goes beyond the limits of interpretative methods recognized by domestic law.27
In this section I discuss briefly what this means when it comes to the interpretation of the Swedish GAAR, and whether such harmonious interpretation can bridge the differences between the GAAR of the ATAD and the Swedish GAAR.
As pointed out earlier, the GAAR of the ATAD can be considered a floor for the application of the Swedish GAAR, whereas primary EU law sets a ceiling for the application. In other words, if the threshold for application of the Swedish GAAR is lower than that of the GAAR of the ATAD, so that it applies to situations that are not covered by the GAAR of the ATAD, this will not conflict with the ATAD, as the ATAD does not preclude a higher level of protection for domestic corporate tax bases, but it could amount to a breach of primary EU law.
One example of a situation that could infringe on primary EU law as interpreted by the CJEU is if the Swedish GAAR was to be applied in a way that constitutes a restriction on the freedoms of the TFEU in a situation where a taxpayer carries on genuine economic activities, which could be the case even if there are tax motives for the activities. In this regard, it is important to note that, although the wording of the Swedish GAAR does not distinguish between domestic and cross-border situations, a restriction on the fundamental freedoms might still be present if the Swedish GAAR is applied in situations that, de facto, only arise in a cross-border context (Scherleitner, 2019) or if it covers resident taxpayers that are engaged in cross-border arrangements more often than resident taxpayers involved in purely domestic arrangements (Kuźniacki, 2020). However, as long as the Swedish GAAR is applied in a nondiscriminatory manner and does not impede free movement or restrict access to the market of another member state, it would not be considered in breach of the four freedoms and no artificiality test would be required28 (but, possibly, the application of the Swedish GAAR could still be in breach of other primary EU law).
On the other hand, if the Swedish GAAR is applied in way that makes its threshold for application higher than that according to the GAAR of the ATAD, this would amount to an under-implementation of the GAAR of the ATAD. Similarly, if the tax consequences under the Swedish GAAR are more lenient than under the GAAR of the ATAD, then it fails to achieve the minimum standard (Haslehner, 2020). As emphasized in this article, according to the CJEU, there is a far-reaching obligation on the part of member states (and thus on the part of the courts and authorities within them) to ensure that the result sought by a directive can be achieved, meaning that domestic law must be interpreted in light of the directive’s wording and purpose.29 Liability for tax in Sweden, however, cannot be based directly on the GAAR of the ATAD, because under EU law directives cannot have direct effect to the detriment of individuals and to the advantage of a state that has failed to implement the directive properly. Inadequate implementation cannot be redressed, then, by disregarding domestic provisions and applying the underlying directive.
However, as the Swedish GAAR (like any GAAR) contains vague criteria and is inherently subjective, it leaves considerable room for interpretation. In my view, it would be possible in most cases to interpret it in a way that is consistent with EU law, or, in other words, that falls within the boundaries set by primary EU law and the ATAD. Of course, this presupposes that Swedish courts are open to arguments based on EU law and not too focused on the established Swedish case law concerning the Swedish GAAR.
For instance, the fact that the Swedish GAAR requires that the obtaining of a tax benefit is the main purpose of the arrangement as opposed to one of the main purposes under the ATAD implies that the Swedish GAAR is not sufficiently strict in this regard. A situation could be conceived where slightly less than 50% of the purpose relates to the obtaining of a tax benefit, which would be enough according to the GAAR of the ATAD, but not according to the Swedish GAAR. In practice, however, it is it often unclear how the purposes should be quantified, meaning that the weighing of the purposes against each other becomes more or less subjective. In my view, there is room for interpreting the Swedish GAAR in a way that avoids a conflict with the ATAD of the GAAR. Even in a case where the nonapplicability of the Swedish GAAR is due to the existence of purposes for an arrangement other than the obtaining of a tax benefit, it might be difficult to attribute the nonapplication of the Swedish GAAR to an incorrect implementation of the GAAR of the ATAD.
Other differences might be more difficult to bridge by means of harmonious interpretation, such as the previously mentioned procedural requirement under Swedish law; that is, that the Swedish Tax Agency is not entitled to apply the Swedish GAAR without involving a court. On the one hand, it can be argued that rules regulating judicial proceedings fall outside the scope of EU law. On the other hand, one could argue that Article 6.3 of the ATAD requires that tax authorities shall be entitled to apply and determine the consequences of the GAAR, or that the effectiveness of EU law requires that tax authorities are able to apply the GAAR (Croneberg, 2021). However, even if the procedural requirement in Swedish law is considered as limiting the effectiveness of EU law, harmonious interpretation can hardly provide the Swedish Tax Agency with a legal basis for applying the Swedish GAAR without involvement of a court.
On a similar note, Fritz (2020) argued that the Swedish GAAR reflects the anti-abuse doctrine in EU law and that it is strange that the Swedish Tax Agency is not entitled to apply the Swedish GAAR when there is an obligation, according to the case law of the CJEU, to deny a benefit that constitutes abuse of EU law. In my opinion, this is not so strange.
First, although the anti-abuse doctrine has clearly played a role in the formulation of the GAAR of the ATAD, the anti-abuse doctrine and the GAAR of the ATAD cannot be equated. For instance, the GAAR of the ATAD might require a member state to ignore certain arrangements, thereby denying tax advantages granted under the corporate tax law of that member state, regardless of whether or not those benefits derive from EU law or not. The anti-abuse doctrine, on the other hand, might require a member state to deny benefits grounded in EU law, but does not concern itself with other tax benefits (Schön, 2020).
Second, and as pointed out by Fritz (2020), the anti-abuse doctrine is a general principle of EU law, which applies regardless of whether the principle has been codified in the national legislation. Thus, regardless of the procedural requirement under the Swedish GAAR, the Swedish Tax Agency may refuse to grant tax benefits on the basis of the anti-abuse principle if the criteria for applying that principle are met and if the tax benefits in question derive from EU law.
In this final section, I discuss to what extent the Swedish GAAR’s new function as an implementation of the GAAR of the ATAD can be expected to affect its interpretation and application in the future. As has already been pointed out, national courts and authorities have an obligation to interpret national legislation such as the Swedish GAAR in conformity with the underlying directive, also taking into account primary EU law. Only if such an interpretation would violate general principles of law, be contrary to the wording of the Swedish GAAR, or go beyond the limits of interpretative methods recognized by domestic law would there be no obligation to let EU law influence the interpretation of the Swedish GAAR.
Thus, if the CJEU clarifies the interpretation of the GAAR of the ATAD, this could influence the interpretation of the Swedish GAAR. However, insofar as the Swedish GAAR is already as strict or stricter than the GAAR of the ATAD, the interpretation of the Swedish GAAR would not be affected by the GAAR of the ATAD. Primary EU law, on the other hand, could affect the interpretation of the Swedish GAAR in a way that limits its previous application. If, for instance, the Swedish GAAR is applied in a way that constitutes a restriction on any of the four freedoms provided in the TFEU, it would be in breach of primary EU law to apply it in a situation where the taxpayer carries on genuine economic activities. Thus, a court would be required to disapply it in such a situation, in spite of the fact that the Swedish GAAR does not contain an artificiality test.
As pointed out by Croneberg (2021), the CJEU has taken into account circumstances for the purpose of determining whether an arrangement is genuine, other than the Swedish Supreme Administrative Court has done so far for the purpose of determining whether the Swedish GAAR applies. Consequently, this is an area where EU law could, potentially, undermine the existing case law on the Swedish GAAR.
The CJEU has held that the obligation to interpret domestic law in conformity with EU law requires national courts to change established case law and even disapply, on its own authority, an interpretation given to a national provision by the national Supreme Court in an interpretative judgment, if it is based on an interpretation of domestic law that is incompatible with EU law.30 Consequently, the obligation to interpret national legislation in conformity with EU law could render the existing case law obsolete. In other words, and as pointed out by Croneberg (2021), the obligation to interpret national legislation in light of a directive’s wording and purpose means that an individual cannot take for granted that a national provision that falls within the scope of the directive will be interpreted in accordance with the case law that was established prior to the adoption of the directive. Thus, the established case law might become partly obsolete.
Another interesting question is whether the impact of EU law on the Swedish GAAR will be limited to certain situations or whether it will apply to the Swedish GAAR in general. In principle, there is no obligation under EU law to interpret the Swedish GAAR consistently with EU law in a purely internal situation that falls outside the direct scope of the ATAD; for example, when the Swedish GAAR is applied to an individual without any cross-border element being present. Despite this, EU law could have an impact on the Swedish GAAR in such cases, too, as the courts would otherwise have to apply the same legislation differently depending on the situation and who the taxpayer is.
Applying the same legislation (i.e., the Swedish GAAR) differently depending on whether or not the situation falls within the scope of EU law or not would give rise to constitutional concerns. Chapter 1, §9 of the Swedish Instrument of Government requires courts and authorities to recognize that everyone is equal before the law. Thus, no one is above the law and the law should apply equally to everyone. The Swedish legislature can differentiate between persons, but that is another matter (Påhlsson, 2007).
Naturally, the principle that everyone is equal before the law does not mean that the law should have the same consequences for everyone. It is implicit in this principle that the equal treatment applies to persons who are in comparable situations. Different treatment under the same legal provision can be justified if there are objective differences. In my view, such objective differences must be grounded in the facts of the case, not in the legal qualification of the facts, as that would render the principle of equality before the law meaningless. Whether a situation falls within or outside the scope of EU law is thus, in my opinion, not in itself relevant for justifying a different treatment under the Swedish GAAR. Consequently, a ruling by the CJEU that is relevant for the interpretation of the Swedish GAAR in a situation that falls within the scope of EU law could also have repercussions for the interpretation of the Swedish GAAR in other situations.
In this context, it is worth noting that the CJEU may give preliminary rulings in situations that fall outside the direct scope of EU law, when the same legislation (in this case the Swedish GAAR) applies in situations that are within the direct scope of EU law. In the
However, in the
In other words, the CJEU may give a preliminary ruling on the Swedish GAAR even when the application concerns a purely internal situation, provided that the referring court explains to the CJEU that the Swedish GAAR also applies in situations that fall within the scope of EU law, and that the principle of equality before the law might require the court to apply the Swedish GAAR equally to situations within and outside the scope of EU law.
It is possible that Swedish national courts, for the purpose of interpreting the Swedish GAAR when it is applied outside the direct scope of the ATAD, will be reluctant to take the jurisprudence of the CJEU into account and that they will rely on established Swedish case law without taking the previously mentioned arguments into consideration. That could lead to an inconsistent interpretation of the Swedish GAAR. Sooner or later, though, there will be cases where the Swedish GAAR is applied within the scope of the ATAD or otherwise within the scope of EU law, where the influence of EU law is more obvious and does not require reasoning based on equality before the law. The interpretation of the Swedish GAAR in these cases could then spread to areas that are not directly within the scope of EU law, so that the interpretation of the Swedish GAAR in cases within and outside the scope of EU law would converge. It would be better, though, if Swedish courts take EU law into account from the start, so that inconsistent interpretation of the Swedish GAAR can be avoided.