Legend has it that a medieval hero of Spain, the Castilian knight El Cid, won a battle after death when his wife strapped his dead body to his horse, knowing that his reputation as an invincible warlord would make his enemies run away in fear. Something similar is happening to Eqiom: while our tax authorities wanted to see it dead and buried by the Danish cases, its principles are still fully valid and should be applied when judging tax abuse.
As we mentioned in our last post, the Spanish tax authorities have rejoiced over the Danish cases, using controversial arguments and a particular reading of ECJ judgments, which raised so many questions that we ended up begging for the Spanish courts to redirect the situation and interpret the Danish cases properly. Apparently, this is what the Spanish High Court (Audiencia Nacional) has done in several recent judgments dealing with dividends paid out by Spanish entities to their EU parent companies controlled by non-EU shareholders. In these cases, the Court addresses the two essential issues: the burden of proof in the Spanish special anti-avoidance rule (SAAR) and the identification of an abusive situation in a foreign holding company.
On the first one, we explained in our post how the presumption of abuse included in the Spanish SAAR and used by the tax authorities did not comply with EU law before and after the Danish cases, which kept the burden of proof of abuse on the authorities. This was against (i) the European Commission’s position, which in 2005 opened an infringement procedure against this SAAR but closed it with no further action (implicitly validating SAAR); and (ii) the Spanish Supreme Court’s criterion, which expressly confirmed the compatibility of this SAAR with EU law and refused to ask for a preliminary ruling.
In its judgments of May 21 (available here), May 31 (available here) and June 10 (available here), 2021, the High Court precludes any shift of the burden of proof and demands higher efforts from the tax authorities claiming lack of business reasons when the dividend withholding tax exemption must be refused on abusive grounds. Remarkably, the High Court mentions that previous Spanish caselaw upholding the SAAR must be considered overruled based on the criteria of the ECJ in the Eqiom case, the Deister/Jühler case and the Danish cases, which is acte claire (i.e., a preliminary ruling is not necessary).
Let us be honest: it was difficult to understand why the ECJ judgment on the Danish dividend withholding taxes did not quote the Eqiom and Deister/Jühler cases. We always thought that such silence simply meant that the Danish cases had followed a different procedural approach and answered different questions, but most important, kept the key doctrine of both previous cases unaltered: the burden of proof of abuse could not be shifted to the taxpayer.
The Spanish tax authorities preferred a different interpretation: the ECJ was wrong when reviewing Eqiom and Deister/Jühler, and in the Danish cases had changed its doctrine. From then on, the EU dividend withholding tax exemption should not protect entities controlled by non-EU shareholders. However, the High Court interprets the silence about Eqiom completely opposite to the tax authorities’ view and states that the ECJ criteria in the Eqiom and Deister/Jühler cases has not been overruled by the Danish cases at all. In fact, the High Court has correctly identified the key aspect addressed by the ECJ in both sets of cases, i.e., the burden of proof on abuse should not be borne by the taxpayer and described it as so straightforward that the acte claire doctrine applied. Therefore, it is the previous caselaw of the Supreme Court that validated this SAAR that should be refused.
This shift should be welcomed. As any court, the ECJ has a difficult role and cannot always be perfectly coherent; however, accusing the ECJ of being so inconsistent within such a short period, circumventing the fact that a different approach was presented in each preliminary ruling, seemed a questionable position by the Spanish tax authorities from the beginning. By finding a coherent explanation of the doctrine in Eqiom, Deister/Jühler and the Danish cases, the High Court makes a more reasonable interpretation of the recent caselaw of the ECJ.
The second issue on the tax status of holding companies, given its factual nature, may hinder the possibility of raising general conclusions. However, in its judgment of May 31, 2021, the High Court’s opinion is clear: a Luxembourg holding company managing its shares prevents it from being a mere conduit company (“sociedad instrumental”). The High Court highlights the existence of relevant investments in different jurisdictions (some of them prior to the Spanish investment in question in the proceedings) and the use of the dividends obtained for reinvestment, as well as the independent role of its board of directors despite being controlled by a sole shareholder resident in Qatar. In this context, the existence of a corporate holding entity is a market standard and not sufficient basis for identifying tax abuse.
The High Court introduces ECJ’s caselaw as conceptual background but does not quote it in its factual analysis, although it could support some aspects of its reasoning. In any event, the conclusion is that there is no tax abuse in those structures where real holding companies are involved. This is consistent with the previous argument: if the burden of proof is on the tax authorities, not only do they have to prove the existence of elements constituting abusive practice, but they also must bear the consequences where fraud or abuse is not so obvious. There could be a misleading aspect, though: while the High Court considers shareholding a real economic activity in this judgment of May 31, it rejects such conclusion in its judgment of May 21 based, among other arguments, on the nature of this activity for VAT purposes. While this apparent contradiction could be solved by considering that the Court was analyzing different safe harbors of the SAAR, it is still a striking outcome when considering its recent caselaw jointly.
The Spanish authorities are likely to appeal these High Court decisions before the Supreme Court, although such an appeal may not actually be reviewed by the Supreme Court because, since 2015, only general interest proceedings get access to the Supreme Court in Spain. If it considered that it should issue doctrine on these cases, it would take a few years before we have definitive caselaw on the proper interpretation of all the ECJ cases (not only the Danish cases, as the tax authorities would prefer).
In the meanwhile, one may argue that the tax authorities should re-assess their understanding of ECJ caselaw, especially considering that what made them rejoice over the Danish cases was viewed the other way around by the High Court in its judgment of May 31 (simply, the proceedings correspond to different tax periods). In fact, we could find a different approach in a public report on the interest withholding tax exemption for EU lenders and tax abuse published on December 16 by the Spanish tax authorities (available here in Spanish). Not only does the tax inspection offer a complete factual analysis (including cooperation with other EU tax authorities), but also highlights the tax treatment that would have applied in case of a direct payment to the non-UE lender under the double tax treaty signed between Spain and the non-UE jurisdiction (remember that the ECJ did not have strong feelings on it; e.g., para. 110 of the PSD judgment).
Finally, on a separate note, we refer to a couple of High Court decisions dated June 18 (available here) and July 22 (available here) of 2021. These cases also deal with the SAAR applicable to the dividend withholding exemption on EU dividends. The High Court’s formal approach to determine the scope of the SAAR is significant: it applies if the Spanish company is controlled by non-UE shareholders, even if the economic ownership over those shareholdings (through a trust agreement) corresponds to an individual resident in the EU. However, it is even more interesting to stress that the High Court follows the criteria of the Spanish Supreme Court (judgment of September 23, 2020, available here) on the condition that, as a rule, the beneficial ownership test that must be included in the wording of the law applies. Although this caselaw deals with the application of double tax treaties, we are of the opinion that such a principle (if the applicable rule lacks a legal beneficial ownership test, only general antiabuse rules could apply) should be extended to withholding tax exemptions on interest and dividends paid to EU lenders and shareholders. Not only does the High Court’s position support this approach (see our post), but also such a rationale explains why the Spanish tax authorities have released the public report on tax abuse referred above.
To sum it up: the Danish cases were clearly landmark judgments in the ECJ’s caselaw. However, they do not totally erase the previous doctrine of the ECJ itself, but rather complements it. Therefore, the burden of proof of abuse remains on the tax authorities, as the Court explained in the Eqiom case, and, of course, not all holding entities controlled by non-EU shareholders are abusive. Finally, at least from our side of the Pyrenees, the beneficial owner clause is not a natural law norm that prevails over the wording of the tax law. Tax authorities cannot use the beneficial ownership principle automatically, its application requires complying with the procedural guarantees required for fraus legis or substance-over-form general anti-abuse rules.