By Aitor Navarro, Assistant Professor, Carlos III University, Madrid
Spanish tax authorities on December 16, 2021, published a panel report on the applicability of the domestic GAAR to a conduit arrangement aimed at benefiting from an exemption at source on the payment of interest income to EU residents. The text of the report is available here (in Spanish).
Background
The arrangement refers to interest payments undertaken in fiscal year 2015 to 2019 by a Spanish subsidiary to a Dutch sister company, both being controlled by a US company. The interest remunerated two participating loans and revolving credit facilities and was payable every six months. The taxpayer considered that the payment was exempted because of the applicability of a provision allowing so when the payment recipient is an entity resident in an EU member state (Article 14.1.c of the Non-Resident Income Tax Act). Although such an exemption provision reflects the content of the EU Interest and Royalties Directive, it was introduced before the directive was issued. Plus, the scope of the exemption is wider than that the directive defined, as it covers interest payments any EU member state obtained, and not only the companies listed in the directive. More importantly, the provision does not require the recipient of the income to be the beneficial owner for the exemption to apply.
The tax inspector referred to the domestic GAAR—article 15 of the general tax code— because it was considered that the Dutch entity was not the beneficial owner of the income but a pass-through entity without substance. Specifically, to ascertain the existence of tax abuse, the GAAR requires that the transactions undertaken are “notoriously artificial or improper for achieving the result obtained” and that “no relevant legal or economic effects result from their use, other than tax savings and the effects that would have been obtained with the usual or proper acts or business.” To enforce the GAAR, the acting inspector must request approval from a tax official’s panel that determines whether the conditions are met in the specific case. The analysed document constitutes the anonymized version of a real case such a panel assessed.
The implications of the report’s publication are twofold. First, the tax authorities publicly positioned themselves in favour of applying the domestic GAAR in conduit scenarios, even if the domestic interest exemption provision does not include a beneficial ownership requirement. Second, the publication allows the tax authorities to impose penalties in future assessments of substantially identical structures.
Features of the Dutch conduit that led Spanish authorities to apply the GAAR
The Dutch entity that received the payment was qualified as a financial service under applicable Dutch regulations. The Dutch tax authorities spontaneously exchanged information with their Spanish counterpart because of the lack of substance requirements posed in the Decree to the International Assistance for the Levying Taxes Act (Uitvoeringsbesluit internationale bijstandsverlening bij de heffing van belastingen). Specifically, the requirement that was not met referred to the fact that the Dutch entity’s accounting records were prepared and kept in the US.
The Spanish authorities collected data on the transactions the taxpayer had undertaken and concluded that the indicia that were found—taken as a whole—showed the Dutch entity’s lack of substance and economic activity, as it operates merely as an instrumental entity. Specifically, the aspects that were considered of relevance referred to the Dutch entity’s substance, its accounting records, and the content of the finance agreements signed with its US parent and the Spanish sister.
Substance: The company’s domicile was shared with the other 4,255 holding companies. Most of the Dutch entity directors were also directors in some of the other companies. The entity had no workers. The only function the entity performed was to channel income from the US parent to the Spanish subsidiary.
Accounting records: The income and expenses were only of a financial nature. There were minor general administration expenses. The entity held no assets aside from credits and cash. The financing the Dutch entity granted to the Spanish entity derived exclusively from funds obtained from the US ultimate parent.
Finance agreements: The only difference between both credit streams was that the US-Dutch deal was LIBOR+1.85% and the Dutch-Spanish deal was LIBOR+2% in the case of the revolving credit facility and LIBOR+4% LIBOR+6% in the case of the participating loans. The spread was the profit the Dutch entity obtained in accordance with its functional profile. Yet the report did not discuss whether such remuneration was at arm’s length. The proceedings were transferred to the US parent either as a dividend (a total of US $18.5 million) or through a loan remunerated at a 10% annual rate (US $5.2 million). The governing law of the contracts was that of the state of New York.
The tax authorities’ assessment and the taxpayer’s arguments
The panel report leaned on the mentioned markers to sustain that “the main purpose is to create an artificial scenario to obtain a tax advantage by creating a certain company structure that does not correspond to the economic reality […] It is an abusive structure as a result of a wholly artificial arrangement […] These loan transactions are purely formal and artificial, lacking any economic and commercial justification.”
Those fragments are representative of the panel’s tone of analysis. Interestingly, in addition to the enforcement of the GAAR, the tax authorities considered that the beneficial ownership requirement was implicit in the domestic provision granting the exemption to interest payments, even if such a condition was not present in its wording. This line of reasoning follows that of the European Court of Justice in the so-called Danish cases, which were explicitly mentioned in the report.
To counteract the assessment, the taxpayer highlighted that the Dutch entity obtained profits that were taxable in the Netherlands. Moreover, the fact that the profits obtained were distributed three years after they were obtained showed that the income was not automatically transferred and that there was no obligation to do so–neither at a contractual level nor at a factual one. It showed that the income may be enjoyed, reinverted, and can generate additional returns until the moment in which the shareholder agrees to distribute a dividend. The panel report disregarded the argument by simply stating that the Dutch entity “is not the ‘real’ beneficiary of the income in the years under investigation.”
On the other hand, the taxpayer considered that because Spain unilaterally grants an exemption on interest paid to any resident of an EU member state, such treatment should be extended to third countries under the EU freedom of movement of capital. The tax authorities answered that the existence of abuse excludes the applicability of those fundamental freedoms, in accordance with the CJEU jurisprudence on the so-called Danish cases. Anyhow, the free movement of capital did not seem to apply in the described setting, but the freedom of establishment—not extensible to third countries—as the US holds 100% participation in its EU-located subsidiaries.
Implications and ongoing concerns
The panel report validated the enforcement of the GAAR to the described structure and denied the applicability of the exemption on the payment of interest to the Dutch entity. It considered that the beneficial owner was the US parent. Thus, the 10% withholding tax limitation on interest envisaged in the Spain-US tax treaty should have been applied.
The report’s outcome has relevant implications for similar conduit structures. As stated, the report’s publication signals the willingness of the tax authorities to resort to the domestic GAAR as a form of bringing the beneficial ownership concept to play in the context of clauses that lack such a requirement in their wording. Such an assessment tool adds to the use of the beneficial ownership concept—present in almost all of the tax treaties of the Spanish network—as an anti-abuse tool that the Spanish judiciary has admitted since the Real Madrid beneficial ownership cases. Plus, since the CJEU decision on the Danish cases was published, the tax authorities have leaned on their outcome in at least three cases concerning beneficial ownership issues. The tax authorities also used transfer pricing regulations to disregard the deductibility of payments and achieve a similar outcome. The tax authorities make use of all available paths to fight against these tax optimization structures.
Additionally, the report leaves two relevant concerns unresolved that should be assessed.
First, it is somehow contradictory to argue that the beneficial ownership requirement is implicit in the domestic provision granting the exemption to interest payments while at the same time applying the GAAR to the described conduit arrangement. If the Dutch entity is not the beneficial owner, the consequence should be to deny the applicability of the exemption. Why resort to the GAAR in the first place? The only logical reason that may come to mind is that the tax authorities want the GAAR to be yet an additional instrument to fight against conduit entities. If courts consider that the beneficial ownership requirement cannot be inferred from the domestic provision, the GAAR still may deny the granting of the withholding tax exemption.
In fact, the tax authorities tried to defend the implicit existence of the beneficial ownership requirement in the Colgate-Palmolive case, in the context of the Spain-Switzerland tax treaty, on what regards source taxation of dividends. The dividends provision of such a treaty does not incorporate the beneficial owner concept. Thus, it restricts source taxation when the recipient is resident in the other contracting state, even if it cannot be labeled as the beneficial owner of the income. The tax authorities resorted to the OECD MTC commentaries and tried to argue that the beneficial ownership requirement is implicit in the wording of the provision. Contrary to such a line of reasoning, the Supreme Court determined that the applicability of a tax treaty cannot be denied based on a text that does not constitute a source of law. In such a tax treaty context, it remains unclear whether the court would have admitted the applicability of the GAAR–especially so if one considers that the Supreme Court has admitted the enforcement of the GAAR in tax treaty scenarios.
Second, the tax authorities decided to disregard a thorough analysis of the requisites the GAAR posed—artificiality and effects other than tax savings— and simply agree on the provision’s applicability by resorting to the mentioned indicia taken as a whole. Thus, it is impossible to determine which elements are key to delineating a beneficial ownership concept that satisfies the requirements posed in the report. For instance, would the outcome be different if the company had substance in the form of employees, local directors not ascribed to other companies, or own premises? If the EU directive proposal against shell entities is adopted, would its substance markers impact the analysis? On the other hand, the fact that the company retained its profits for three years seems to be irrelevant in the report’s analysis. Would the outcome be different if, instead of distributing it as a dividend, the income was reinvested? Would the Dutch entity be considered the beneficial owner if it owned assets other than credits and cash? The report’s outcome generates additional uncertainty in this respect.
Moreover, regulations on tax penalties allow the imposition of fines to structures that resemble those of the panel reports that were published, in which the domestic GAAR was considered to apply. Notwithstanding, regulations require that the assessed arrangements are “substantially identical” to the case to which the published report refers. As stated, the specific impact of each of the features of the commented case remains unclear. Thus, it is certainly impossible to ascertain when a case could be considered substantially identical. Therefore, the issue lies not only with such a formulation’s lack of determination but also with the indiscernible relevance of the structure’s features assessed in the published report.
In sum, the published report on the domestic GAAR’s applicability to conduit structures—even in the absence of a beneficial ownership requirement in the applicable regulations— constitutes yet another relevant development in this rapidly evolving field. The significant degree of uncertainty in assessing these arrangements anticipates further litigation. It is yet to be seen whether the proposed directive against shell entities helps to bring order or increases the complexity of assessing the subject matter.
-
Aitor Navarro is an assistant professor at Carlos III University in Madrid.
Be the first to comment