Parent-Subsidiary Directive and abuse: beneficial ownership may not be sufficient where the overall arrangement is not genuine

In the Opinion delivered on 21 May 2026 in Case C-203/25, Neo Group (“Opinion”), Advocate General Kokott addresses a particularly sensitive issue in the application of the Parent-Subsidiary Directive, namely whether exemption from withholding tax may be denied even where the dividends are distributed to an EU parent company which, from both a formal and an economic perspective, is the actual recipient of those dividends. The solution proposed is of particular systemic relevance. While stating that, as a rule, the parent company’s status as beneficial owner excludes abuse, the Advocate General takes the view that this is not necessarily the case where the intragroup distribution forms part of an overall non-genuine arrangement designed to shield the final beneficiary from taxation. The Opinion therefore shifts the focus from the position of the immediate recipient alone to an assessment of the entire distribution structure and of its overall tax outcome.

The case

The Lithuanian company Neo Group UAB (“Neo”) distributed dividends in 2016 and 2017 to its Cypriot parent company (“Retal”), applying the exemption from withholding tax provided for under the domestic legislation transposing Directive 2011/96/EU (“PSD”). According to the Lithuanian tax authority, however, those flows formed part of a broader series of transactions, at the end of which the profits were transferred to the natural person at the top of the group (“PG”), who was not resident in Lithuania, through Retal, which actually carried on a business activity, and a second Cypriot top holding company (“PPH”), the parent company of Retal and itself wholly owned by PG[1].

In the light of the foregoing, the referring court raised the following six questions for a preliminary ruling:

1. whether the exemption from withholding tax provided for by the PSD may be denied, on the basis of the anti-abuse rule laid down in that Directive, even where the dividends are paid to a parent company resident in an EU Member State which actually carries on an economic activity and is the beneficial owner of the flows, but is nevertheless regarded as part of a chain of non-genuine transactions through which the dividends are transferred to the beneficial owner of the entire group; 
2. whether the substantial correspondence, or minimal difference, between the amounts of dividends transferred along the chain may constitute an indication of their abusive transfer; 
3. whether the domestic anti-abuse clause may be applied by the source State even where the transactions which led to the classification of the chain as non-genuine were carried out in another Member State; 
4. whether withholding tax on dividends, applied on the basis of the anti-abuse rule, may be justified irrespective of the segment of the chain in which the tax advantage arises and, therefore, even where that advantage arises only at the level of the final beneficiary; 
5. whether abuse may be inferred from the use made of the dividends received, where the subsequent distribution of profits develops continuously over time; 
6. finally, whether the anti-abuse clause may be invoked against the entity distributing the dividends even without proving that that entity was, or could have been, aware of the existence of the chain of fictitious transactions.

The relevant EU provisions

Under EU law, the objectives and rationale of the PSD are clarified in recitals 3, 4 and 6, namely:

• to eliminate economic double taxation on income distributed in the form of dividends, already taxed at the level of the subsidiary as business profits, by exempting such income from withholding tax (recital 3); 
• to create within the EU, with respect to groupings of companies of different Member States, conditions analogous to those of an internal market, in order to ensure the effective functioning of that market. According to the PSD, such operations should not be hampered by specific restrictions, disadvantages or distortions arising from domestic provisions. It is therefore necessary to provide for tax rules which are neutral from the point of view of competition, in order to allow enterprises to adapt to the requirements of the internal market, to increase their productivity and to improve their competitive strength at international level (recital 4); 
• to foster cooperation between companies of different Member States, which would otherwise be disadvantaged in comparison with cooperation between companies of the same Member State (recital 6). 

However, the PSD itself sets a limit to the application of the exemption by providing, in Article 1(2) and (3), for a general anti-abuse clause, which reads as follows:

“2. 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. 

3. 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.”.

The rationale of the general anti-abuse clause is clearly outlined in recital 8 of Directive 2015/121/EU, which amended the PSD:

“While Member States should use the anti-abuse clause to tackle arrangements which are, in their entirety, not genuine, there may also be cases where single steps or parts of an arrangement are, on a stand-alone basis, not genuine. Member States should be able to use the anti-abuse clause also to tackle those specific steps or parts, without prejudice to the remaining genuine steps or parts of the arrangement”.

The principles set out by the Advocate General

Against that factual and legal background, the legal analysis of the case begins at paragraph 34 of the Opinion and, in essence, focuses on one key issue: whether, in the case at hand, the conditions for abuse under Article 1(2) of the PSD are met.

The analysis is structured around four main points:

• the possibility that abuse may also exist where the recipient parent company is the beneficial owner of the dividends (first question referred for a preliminary ruling); 
• the relevance, for the purposes of establishing abuse of the PSD, of a non-genuine arrangement carried out in a Member State other than that of the subsidiary, namely in a Member State other than that in which the exemption is claimed (third and fourth questions referred for a preliminary ruling); 
• the significance to be attached to the correspondence between the amounts distributed and to the close timing between the receipt and subsequent onward distribution of the dividends (second and fifth questions referred for a preliminary ruling); 
• the weight to be given to the awareness, on the part of the subsidiary distributing the dividends, of the existence of the abusive arrangement (sixth question referred for a preliminary ruling). 

With regard to the first question, Advocate General Kokott starts from the wording of Article 1(2) of the PSD and recalls that the refusal of the benefit requires two cumulative requirements: (i) the existence of a non-genuine arrangement, identifiable to the extent that the arrangement was not put into place for “valid commercial reasons” which reflect economic reality (paragraph 3); and (ii) the pursuit of a tax advantage that defeats the object or purpose of the PSD.

According to the Advocate General, the decisive element is precisely the second requirement. The anti-abuse clause is not intended to tackle every tax-efficient, or even merely artificial, arrangement, but only arrangements aimed at obtaining a tax advantage that is contrary to the rationale of the Directive. Accordingly, the Opinion states that:

“49. In the scheme of the Parent-Subsidiary Directive, corporation tax is therefore levied on the profits of a group of companies, in principle, only at the point where they are generated at the level of the subsidiary, and income tax is levied only on the ‘final’ distribution by the group’s parent company to its shareholder, in the form of a natural person, as his or her capital gains.

50. The intended tax neutrality of distributions within a group of companies is therefore the basis for the assumption that the State (or States) in which the ‘final’ distribution by the group’s parent company to its shareholder takes place will tax that distribution in the exercise of its fiscal autonomy. The Parent-Subsidiary Directive does not, however, prescribe whether and how that ‘final’ distribution of profits is taxed.”.

On that basis, the Advocate General clarifies that abuse typically arises where the constituent elements of the situation to which the PSD attaches the exemption are fulfilled only pro forma, that is, where the person benefiting from the tax advantage is not, in reality, the person whom the PSD is intended to protect. This is the case, for example, of a merely interposed parent company, which formally receives the dividends but is not their actual economic recipient.

Conversely, where the constituent elements are also present from an economic perspective (i.e. where the recipient of the dividends is their beneficial owner) the absence of abuse must, as a rule, be presumed.

Furthermore, in such a situation, a non-genuine arrangement cannot be inferred merely from the fact that the parent company, which is the beneficial owner, in turn, redistributes its own profits, including the dividends received (paragraph 61 of the Opinion). In particular:

“62. Transmission of dividends within a group of companies to the controlling parent company, which ultimately distributes its profits to one or more natural persons, is in fact consistent with a properly operating corporation. It is specifically the purpose of a corporation to generate profits and to distribute them to its shareholders”[2].

According to Advocate General Kokott, abuse may exceptionally exist even where the parent company is the beneficial owner of the dividends, if the distribution from the subsidiary to the parent company, although reflecting economic reality when considered in isolation, forms an integral part of an abusive overall arrangement.

This follows from the fact that, under Article 1(2) and (3) of the PSD, an arrangement may comprise more than one step or part, and even individual components of that arrangement may be not genuine, provided that there is a purposive connection between that component and the tax advantage pursued in breach of the PSD:

“72. […] In other words, there can be abuse of the Parent-Subsidiary Directive where the distribution by the subsidiary to the parent company – which, in isolation, reflects economic reality – is part of an abusive overall plan”.

From that perspective, the abuse may consist in the fact that the PSD is exploited in order to generate, for the benefit of the final beneficiary who is a natural person, income which is removed from final taxation through further transactions. The example referred to by the Advocate General is precisely that of a structure which, although involving a genuinely existing and economically active parent company, is used to prevent the competent State from taxing the final distribution of profits. According to the Opinion:

“74. Such an arrangement undermines the assumption underlying the Parent-Subsidiary Directive. As has already been stated, the Parent-Subsidiary Directive assumes that the State in which the ‘final’ distribution takes place can tax that distribution in the exercise of its fiscal autonomy (see above, point 50). If it is deprived of that possibility and taxation of the ‘final’ distribution, which is otherwise provided for, is frustrated, it would therefore be inappropriate to grant the tax advantage under the Parent-Subsidiary Directive”.

On that basis, Advocate General Kokott then addresses the third and fourth questions together. In essence, the issue is whether a non-genuine arrangement carried out in another Member State (i.e. a Member State other than the one in which the exemption is claimed) may be relevant for the purposes of establishing abuse of the PSD, and whether withholding tax may be justified even where the tax advantage arises only at the final stage of the chain, at the level of the final beneficiary.

The Advocate General’s answer is clear in terms of method: what matters is not the place where the individual artificial step is located, but its purposive connection with the tax advantage obtained through the PSD. Precisely because the anti-abuse clause requires all relevant facts and circumstances to be taken into account, it cannot be ruled out in the abstract that transactions carried out in another Member State may also be relevant for the purposes of the abuse assessment.

However, Advocate General Kokott introduces an important qualification in this respect: where the alleged non-genuine arrangement concerns the tax law of another Member State, it must be assessed with particular care whether that arrangement is indeed not put into place for valid commercial reasons and whether the tax advantage pursued actually defeats the object or purpose of the PSD. From this perspective, the Opinion notes that the transactions carried out in the case at hand do not appear to lack “valid commercial reasons”, while leaving that assessment to the referring Tax Disputes Commission.

The second and fifth questions, by contrast, are answered in the negative. The domestic Tax Court asked whether the substantial correspondence between the amounts distributed along the chain and the close timing between the receipt and subsequent onward distribution of the dividends could constitute indications, or even requirements, of the existence of abuse.

In the case at hand, since the mere onward transfer of dividends does not, in itself, constitute a non-genuine arrangement, neither the correspondence between the amounts nor their temporal proximity can have decisive value or constitute autonomous requirements for abuse. At most, they are factual circumstances to be assessed in the overall context of the transaction, but not elements which, on their own, make it possible to characterise the structure as abusive.

Finally, with regard to the sixth question, the Advocate General addresses the subjective element of abuse. Since Article 1(2) of the PSD requires that the arrangement be put into place for the main purpose, or one of the main purposes, of obtaining a tax advantage, the finding of abuse also entails an assessment of the purposive element. From this perspective, Advocate General Kokott observes that what is primarily relevant is the knowledge of the person who made the decision to implement the arrangement regarded as abusive. The point, therefore, is not so much the position of the dividend payer as such, but rather the possibility of linking the awareness and purpose of the contested arrangement to a specific decision-making centre.

Conclusions

The Opinion is particularly significant because it clarifies that, for the purposes of applying the anti-abuse clause under the PSD, the parent company’s status as beneficial owner does not, in itself, exhaust the analysis. The decisive point remains whether the intragroup distribution forms part of an overall non-genuine arrangement capable of frustrating the purpose of the PSD, which takes into account and preserves taxation at the first and final links of the participation chain.

What emerges is a broader and more substantive approach, requiring consideration not only of the immediate recipient of the dividends, but of the entire architecture of the transaction. In this sense, the Opinion provides an important systemic indication, which is likely to affect the future scope of application of the anti-abuse clause in the field of intragroup dividends.

[1] According to the reconstruction of the Lithuanian tax authorities, PPH had liabilities to PG corresponding to the purchase price for the acquisition of Retal and, at the same time, claims against PG arising from loans granted and claims assigned to PG for amounts which, overall, corresponded to the dividend flows received. Those positions were subsequently offset against each other, without any distribution of dividends by PPH.