EU holding company structure in treaty shopping cases

Treaty shopping gets trickier but it’s not impossible

In the last months of 2016 a couple of tax cases, one from Italy and the other from France, spread new interest on the tax status and characterisation of passive holding companies. The issue has gained momentum more recently because, among several provisions of the MLI, Article 7 is dedicated to preventing different forms of treaty abuse and contains some reference to holding companies (Article 7(10)(a)(i), MLI).

Although international tax initiatives are demonising the general use of holding companies, the art of using them is not dead yet. There is still room for them if there are genuine organisational reasons in the MNE structure for them to exist.

The Italian Supreme Court of Cassazione (case 27113/16) took a pragmatic position on the basic features of a quite common cross-border structure where an EU regional sub-holding of a multinational group was receiving dividends from its Italian subsidiary with the ultimate parent company established in the US.

Because the European holding was considered to be a conduit entity, it was refused the reimbursement of the withholding tax applied when the dividends were paid. So the company went to the tax court to obtain the recognition of its right to the reimbursement.

After two sets of unfavourable proceedings, the Supreme Court finally ruled in favour of the taxpayer because of the constitutive elements it recognised the holding company to have. Recalling that an enterprise lacking in economic substance and established just to cash-in dividends cannot take advantage of any tax treaty, it stated that due attention should be put on the "peculiarities undoubtedly derived from the nature and functions of the company", i.e. the holding.

In particular, the court pointed out that the elements characterising an operative company have a "completely different" scope in respect of a pure holding company that has "simply" to instruct and control its subsidiaries, to participate in their shareholders meetings and to cash-in dividend distributions.

Earlier, in 2011, the Italian Revenue Agency issued guidance on reimbursing withholding tax charged on dividends to EU companies. It stated that due attention should be put into evaluating the hypothesis that the cashing in of dividends was structured in a "wholly artificial arrangement" through conduit entities or conduit arrangements.

Echoing the European Court of Justice (ECJ, case C-222/04, Cassa di risparmio di Firenze), the Revenue Circular agreed that the analysis has to be conducted according to a substance-over-form approach, while the establishment in a low taxation country is not per se a symptom of being in presence of an artificial arrangement.

In the case of holding companies, the tax authority admits that even if the evaluation of the "inactive" character of a company and its potentially artificial nature must be made with "particular caution" because "usually these entities do not show a significant physical presence in conducting their activities", they should not be deemed to be only a form of abuse on the freedom of establishment for this feature alone.

The solution to this problem has been found in the distinction between an active (business) holding company and a (mere) passive holding company. The conclusion has been that whenever the company perceiving the dividends can:

• Be characterised as a "mere holder" of shareholdings;

• Be considered a resident in a country granting a comprehensive tax treatment "significantly better" than that granted in Italy to a company in a similar condition;

• Be considered not to have any other activity different from the mere holding of such shares; and

• Be considered to have the exclusive or main centre of interest in Italy because of the number and relevance of its shareholding,

the establishment of that holding company in the other country can be considered as a wholly artificial arrangement, having the aim of only avoiding the Italian domestic (and EU) withholding regime.

In one of its recent cases that is still awaiting a final ruling (case C-6/16), the EU court has to decide if a member state may refuse, on grounds of preventing tax evasion, an exemption from withholding tax that would normally be granted for dividend payments of a resident subsidiary to its non-resident parent company.

The issue at stake is that the Luxembourg-resident company, which is indirectly controlled by a Swiss entity, did not give the necessary proof that the principal purpose behind the structure of the chain of interests was not tax-related.

In his Opinion, Advocate General (AG) Kokott concluded that "an artificial arrangement can be assumed if the company is only a fictitious establishment in the form of a 'letterbox' company. But even where there is a physical presence, one might conclude, in light of the financial and staffing set-up, that the arrangement is artificial. In this regard, what appears to be relevant is, for instance, the actual authority of the company organs to take decisions, to what extent the company is endowed with its own financial means and whether any commercial risk exists" (para. 56, 57).

This position represents a sort of step backwards in respect of previous (quite consolidated) ECJ precedents (case C-208/00, Überseering; case C-212/97, Centros) according to which the fact that a company was formed in a particular member state for the sole purpose of enjoying the benefit of more favourable legislation does not constitute abuse even if that company conducts its activities entirely or mainly in a second state (ECJ, case 167/01 Inspire Art, paragraph 97).

Kokott's perspective, if confirmed by the court, seems a bit at odds with the European Commission's and EU Council's one. In fact, the Commission, in its recommendation of January 28 2016 against tax treaty abuse, introduced requirements that make reference to the authenticity of the economic activity performed by the taxpayer and such a requirement is larger than the "fraudulent" one the ECJ has made reference to in its jurisprudence.

Consequently, if setting up a mere conduit company can hardly be considered as a "fraudulent arrangement", it can nonetheless be characterised as an arrangement devoid of a genuine economic purpose and so be deprived of any tax treaty benefits.

The recent EU Anti-Tax Avoidance Directive I (ATAD I) considered that "within the Union, GAARs should be applied to arrangements that are not genuine; otherwise, the taxpayer should have the right to choose the most tax efficient structure for its commercial affairs… When evaluating whether an arrangement should be regarded as non-genuine, it could be possible for member states to consider all valid economic reasons, including financial activities".

The prescriptive result of this approach is Article 6 of ATAD I, according to which a "main purpose test" has been introduced to evaluate the "genuine" nature of an arrangement. The point here is that if an arrangement is to be "regarded as non-genuine to the extent that they are not put into place for valid commercial reasons which reflect economic reality", the consequence might seem an enlargement of the previous requisite of the "wholly artificial arrangement" introduced in 2006 by the Cadbury Schweppes decision. In case of passive holding entities, such requirements as activity performed, staff and personnel and physical premises remain hardly detectable. Interestingly, this evolution of the Commission's approach reflects the analogous perspective of the OECD.

After its 1986 Conduit Report, the OECD BEPS Action 5 Report took a revised approach to holding companies, stating that "the substantial activity factor requires, at a minimum, that the companies receiving benefits from [holding] regimes respect all applicable corporate law filing requirements and have the substance necessary to engage in holding and managing equity participations (for example, by showing that they have both the people and the premises necessary for these activities)".

Specific attention for preventing treaty abuse is discussed in the BEPS Action 6 Report, in which paragraph 22 suggests that, at a minimum, states should implement a principal purpose of transactions or arrangements test (PPT) or a PPT and a limitation on benefits clause (LoB). One of the most relevant requirements of the LoB clause is the "active conduct of a business" (the so-called active business test) according to which granting treaty benefits to an entity resident of a contracting state is possible only if the entity actively carries on business activities in that state where an entity will be considered to be engaged in the active conduct of a business only if persons through whom it is acting (employees of a company) "conduct substantial managerial and operational activities".

Some problems admittedly arise in respect of a passive holding company because "making or managing investments" will be considered a "business" only when the relevant activities are part of banking, insurance or securities activities conducted by a bank or financial institution. "Such activities conducted by a person other than a bank (or financial institution agreed to by the contracting states), insurance enterprise or registered securities dealer will not be considered to be the active conduct of a business," the Action 6 report states.

The other cornerstone in the Action 6 anti-treaty shopping strategy is the introduction of a PPT. According to the proposed rule, "the benefits of a tax convention should not be available where one of the principal purposes of certain transactions or arrangements is to secure a benefit under a tax treaty and obtaining that benefit in these circumstances would be contrary to the object and purpose of the relevant provisions of the tax convention".

All the above considerations have been assumed to identify the so-called "minimum standard" of the MLI. Moreover, Article 7(10)(a)(i) of the MLI states that an entity defined as a holding company will not be seen as performing an "active conduct of a business". This exclusion seems quite peremptory unless the discretionary rule of Article 7(11) of the MLI applies. In this case, a contracting state may grant the benefits of a tax treaty "only if such a resident demonstrates to the satisfaction of such competent authority that neither its establishment, acquisition or maintenance, nor the conduct of its operations, had as one of its principal purposes the obtaining of benefits under the covered tax agreement". The commentary on this provision clarifies that in order to be granted benefits, the person must establish, to the satisfaction of the competent authority of the state from which benefits are being sought, that there were clear non-tax business reasons for its formation, acquisition, or maintenance. It is not necessary to find conclusive proof of intent, but the competent authority must be able to conclude, after an objective analysis of the relevant facts and circumstances, that none of the principal purposes for the establishment, acquisition or maintenance of the person and the conduct of its operations was to obtain benefits under the convention.

It is worth noting that this rule does not seem to consider the question of performing "genuine economic activities" when assessing whether or not a transaction passes the test, as it has been done in line with the EU's ATAD I. In this case, the MLI's PPT rule might present a different personal scope if compared to the ATAD I's PPT because, according to the former, "it is sufficient that at least one of the principal purposes was to obtain the benefit".

However, from international case law, these justifications could be easily overruled by tax authorities if the tax advantage is considered to be "excessive" in comparison exactly with those non-fiscal "relevant facts and circumstances", as the Prevost case in Canada showed in 2010 (Prévost Car Inc. v. The Queen).

In this situation, tracing a clear line of demarcation among the different forms of holding companies and their uses in MNE strategies is extremely hard, and an attempt may (and probably should) be done to identify the hypotheses of non-abusive holding structures through the examination of the precedents of domestic court cases and international materials.

A synthesis of this perspective can be seen in Table 1 (A catalogue of good practice).

Table 1
Italian cases law		ECJ jurisprudence		OECD references
Freedom of adopting laws and regulations	Law 212/2000, Article 10 bis	Actual authority of the company organs to take decision	C-6/16, Holchim	Reliable legal system, business friendly environment,
	Cass. no. 1372/2011	Endowed with financial means		Business friendly environment
Functionality and adequacy of the adopted instrument in respect of the economic objective	Cass. no. 21390/2012	Assumptions of commercial risks		Political stability	OECD Final Report on BEPS Action 6, paragraph 26; Commentary, paragraph 14, example G )
Normal market practice	Cass. no. 439/2015	Registered office in the EU and freedom of establishment	C-208/00, Uberseering	Membership of a regional grouping
Effective ownership of the shareholding and use of dividends received	Cass. no. 1465/2009	Direct and indirect involvment in the management of the companies controlled		Sophisticated banking industry
Effective legal and economic availability of shareholding income	Cass. no. 27113/2006	Supply of financial services	C-108/14, Larenia	Comprehensive double taxation treaty network” of the country of establishment
	Cass. no. 25281/2015

The tax treatment of a passive holding company remains quite controversial or, at least, uncertain. Even though there is a sort of unanimity in affirming that these structures are inevitably "slim" and their business does not require extensive presence of personnel and equipment, the same characteristics are seen by tax administrations (and some judges) as a "proof" of wholly artificial arrangements or, at least, not genuine.

Nevertheless, it can be assumed that a company merely holding property (securities, debt instruments, real estate, intellectual properties) may not carry on a business or an undertaking and, consequently, its income will not be classified as business income and therefore, it may be seen as a sort of "vehicle for imputation of enjoyment rights".

Recent Italian tax law cases seem to demonstrate that for passive holding companies to be recognised as not "abusive", they should only possess a sufficient degree of autonomy in assuming relevant decisions in the management of the controlled companies and in the use and investment of the (passive) income perceived, i.e. to perform functions of management and coordination of the group. Put it simply, to avoid any criticism relative to their economic functions, holding entities should be able to demonstrate that they effectively manage the participating entities.

Any other lack of structural and organisational set-up should not matter for characterising a passive holding as a non-genuine or abusive arrangement.

The only remaining problematic case may be represented by the sole shareholder holding company when it seems to have no other function except to localise income in a particular jurisdiction, without having any effective link with that "market". But even in this event, an analysis grounded on the "facts and circumstances" of the case could prove the soundness of the organisational structure adopted by the undertaking.

In this hypothesis, however, the holding company might find its way if its role in the multinational group actually responds to sound organisational reasons after having considered the "reliable legal system, business friendly environment, political stability, membership of a regional grouping, sophisticated banking industry and the comprehensive double taxation treaty network" of the country of establishment, as the OECD suggests (Final Report on BEPS Action 6, paragraph 26; commentary, paragraph 14, example G).