In its judgment of January 21, 2020 (Santander case, available here), the European Court of Justice (ECJ) not only prevented the Spanish Central Tax Tribunal (Tribunal Económico-Administrativo Central – TEAC) from requesting a preliminary ruling due to its lack of juridical independence (para. 77), but it also recalled its obligation to ensure that EU law is applied and to override national provisions that are contrary to EU law provisions that have a direct effect (para. 78). We address how the TEAC faced this task in a couple of resolutions dated October 8, 2019, dealing with the payment of Spanish-sourced dividends and interest to EU taxpayers and the application of the ECJ’s doctrine in the well-known Danish cases (ECJ’s judgments of February 26, 2019).
Given the particularities of the Spanish withholding exemption on interest, we will focus on the case involving a Spanish listed company engaged in the energy sector that paid dividends and attending fees to its Luxembourgish parent company. Apart from other investments, the Luxembourgish shareholder held around 9.5% in the stake of the Spanish company, which in Spain is a qualified percentage to be entitled to apply the dividend withholding exemption.
However, provided that the Luxembourgish company was an EU holding company of a third State’s sovereign fund, the withholding agent found applicable the Spanish special anti-avoidance rule (SAAR). In line with the French SAAR analyzed in the ECJ’s judgment of September 7, 2017 (‘the Eqiom case’ – available here), this rule denies the withholding exemption applicable to dividends paid to EU parent companies when they are controlled directly or indirectly by non-EU shareholders. In the tax periods concerned, it provided three safe harbors (including a valid business reasons test) that allowed the application of the dividend withholding exemption when the taxpayer gave evidence that any of them was met.
In its resolution, the TEAC refused to allow the refund of the withholding tax based on several arguments that deserve strict attention, because it is likely that they will guide the Spanish tax auditors’ activity.
Before describing the judgment, we want to stress how it is drafted. The TEAC goes far beyond the task of an administrative review body: it ignores that its job is just to review the claimant’s appeal and the soundness of its grounds. Instead, the judgment is drafted as a 32-page tax paper, where the key point is not analyzing the facts underlying the case, but spreading this joyful news to the tax inspection: the Court of Justice was wrong when issuing its decision in the Eqiom case, but in the Danish cases it found its way to Damascus: the EU dividend withholding exemption should not be applied to companies controlled by non-EU shareholders.
The TEAC’s reasoning started by confronting the ECJ’s criteria in the Danish cases (we will stick to the decision concerning the Parent-Subsidiary Directive – ‘the PSD judgment’, available here) with its criteria in the Eqiom case, and concluded by identifying a remarkable change in the ECJ’s approach. In its view, in the PSD judgment, the ECJ took a stricter approach before the protection of the ‘freedom of establishment’ (sic) where non-EU shareholders use EU intermediate holding companies. The TEAC highlighted the reinforcement of the general principle of prohibition of abuse of EU law, the lack of any reference to the Eqiom case and the fact that the ECJ did not follow AG Kokott’s Opinion. As expected, the TEAC also referred to the denial of the withholding exemption when the beneficial owner is resident in a third State (para. 111 of the PSD judgment), adding that if that situation is merely argued, in such a case the burden of proof is on the taxpayer.
Of course, the interplay between the Danish cases and the ECJ’s previous case law is not an easy question. It is a possibility that the ECJ has broadened the concept of abuse. However, the TEAC gave no relevance to the fact that, in the Eqiom case, the ECJ was confronted with a French SAAR, similar to the Spanish SAAR, that was found to be incompatible with the EU law as it presumed abuse when the parent company receiving the dividends was controlled by non-EU investors and, because of this, it shifted the burden of proof to the taxpayer.
Remarkably, at this point, the TEAC stopped the confrontation test and did not refer to the ECJ’s doctrine on the burden of proof, which is addressed in both the Eqiom case and the Danish cases. By contrast, the TEAC referred to the Spanish Supreme Court’s case law, which in the past (and prior to the Eqiom case) validated the shift of the burden of proof included in the Spanish SAAR. However, we are of the option that this presumption of abuse, included in the Spanish SAAR, and allowed by the TEAC, is against the EU law before and after the Danish cases, which keeps the burden of proof of abuse on the tax authorities (para. 117 of the PSD judgment).
One may argue, as the TEAC did, that the taxpayer does not have the option of relying on the fundamental freedoms to call into question the Spanish legislation because an abuse had been identified (para. 123 of the PSD judgment). In the case analyzed by the TEAC, only the fundamental freedoms could have prevented the application of this presumption, given that the Luxembourg company had a stake below 10% (so the Parent-Subsidiary Directive was not strictly applicable). However, in our opinion, this is not as easy as what the TEAC derived from the Danish cases, and a couple of (group of) questions must be raised.
First, can the fraud or abuse that precludes the fundamental freedoms be identified by using a general presumption that shifts the burden of proof to the taxpayer? Is this a proper interpretation of the Danish cases? Can the French SAAR scrutinized in the Eqiom case now be valid depending on the facts and circumstances of the Luxembourg holding company that invested in the French subsidiary?
Second, did fraud or abuse exist, as defined in para. 97 of the PSD judgment, if one takes into account that the third-State investor could already rely on the free movement of capital to avoid the Spanish withholding tax, so that the use of an EU intermediate holding company (‘the artificial arrangement’) does not allow it to obtain any advantage?
In fact, the TEAC assumed that the 9.5% stake in the Spanish company did not allow any influence over it (doubtful, by the way), so that such an investment would have been a protected portfolio investment made by a resident in a third State (see ECJ’s judgment of November 13, 2012, FII (2) case, para. 99, available here) entitled to the same withholding exemption as in the domestic case. However, conveniently, the TEAC did not quote para. 110 of the PSD judgment, where the ECJ stated that in a case where the dividends would have been exempt had they been paid directly to an investor resident in a third State, a group cannot be reproached for having chosen a particular EU structure rather than a direct payment.
All this conceptual discussion is not a theoretical exercise in this case, because both the taxpayer and the tax authorities submit several arguments on the sound business reasons for the structure and its artificiality, respectively. On the one hand, the taxpayer argued that the Luxembourg company was a permanent, multijurisdictional vehicle with various investments, and the shares in the Spanish listed company amounted to around 49% of its balance sheet and part of its income. It was also argued that this investment was made a year and half after the holding company was incorporated, and regulatory requirements for investments exceeding 10% in companies engaged in the energy sector also explains why an EU vehicle was needed for this investment. Following these requirements, the TEAC discarded the latter argument.
The tax auditors argued that the holding company was incorporated with the minimum share capital, and that its registered office was a trust company’s office that contributed two ‘class-B’ directors, with the remaining two ‘class-A’ directors being from the third-State fund group, as well as the fact that the company had no employees. The tax auditors highlighted that the majority of the investments corresponded to companies not included in the group, and they pointed out certain transactions made by the holding company with other assets as potential evidence of artificiality. Particularly, the financing of the Luxembourg company by way of convertible certificates (CPECs) was highlighted and identified as a ‘non-formal but actual dividend payment’ to its parent company.
On these grounds, the TEAC concluded that the Luxembourgish holding company failed to give evidence of the economic reasons behind its incorporation. Apparently, it assumed that a holding company equals abuse, a conclusion that opposes the ECJ’s case law (judgment of December 20, 2017, Deister/Jühler case, para. 73 – available here), and that it cannot be found as such in the Danish cases. That is to say, in a case that was not ‘crystal clear’, it was the taxpayer which suffered the consequences. However, if the burden of proof is on the tax authorities, this should mean that not only do they have to establish the existence of elements constituting an abusive practice, but they also have to bear the consequences of a case where the existence of fraud or abuse is not so self-evident.
The TEAC’s conclusion may be based on the fact that since the Luxembourgish holding company was not the beneficial owner of the dividends, it was not entitled to the PSD exemption. We have to set aside the controversy of this ECJ argument: referring now to AG Kokott’s Opinion seems to be too melancholic, and even Eqiom finally lost its case before the French Supreme Court because of this (judgment of June 5, 2020, available here). However, in the absence of an analysis of what actually happened with the dividends paid by the Spanish company and the contractual conditions of the CPECs, its mere existence alone should not ground that argument. Additionally, the requirement to have human and material resources in a structure that would have been protected by the free movement of capital raises many questions.
On a separate note, the taxpayer also argued the application of article 10.2 of the Double Tax Treaty signed between Spain and Luxembourg (DTT), which establishes a 15% withholding rate on dividends paid to beneficial owners resident in Luxembourg. After arguing that the attending fees are also included in the concept of dividend for treaty purposes, the TEAC discarded the application of the DTT given the lack of business reasons for the incorporation of the Luxembourg holding company and the funding via CPECs.
Sometime ago we wrote in this blog (available, here) that the ECJ made a mistake in the Danish cases when using the same signs of abuse relating to dividend and interest withholding taxes. The “copy&paste” technique that the ECJ used to replicate the signs of abuse in both the interest and dividends cases is not only embarrassing, it is wrong. We can agree that an immediate on-payment of interest can be considered an indication of a lack of beneficial ownership and, therefore, of abuse. But, we firmly disagree that a holding company paying dividends from its profits (which by nature can only be dividends and gains) is an abuse of EU fundamental principles.
Right or wrong, the ECJ’s position of giving the same treatment to interest and dividends is favorable to the tax administration: if a company that receives interest or dividends pays them to its financiers or shareholders, it is no longer the beneficial owner and therefore it is carrying out an abuse. In the case underlying the TEAC’s judgment, the situation was not that clear-cut because the holding company was financed by debt in the form of CPECs. The TEAC uses only one line to solve this potential weakness in all of its reasoning: though the Luxembourgish company did not formally distribute dividends out of the dividends received from Spain, by serving the CPECs, it was doing so economically.
To sum up, in this resolution, the TEAC showed how the tax authorities rejoiced over the Danish cases. It is obvious that these decisions provide tax auditors with arguments that they did not have before in the ECJ’s case law, so it makes sense that they harden their position. However, we think that this resolution went too far, accusing the ECJ of having conflicting case law and cherry-picking its criteria in the Danish cases against the taxpayer. We hope that Spanish courts will redirect the situation and interpret the Danish cases properly.