Patrick Schrievers
Akshay JahagirdarThe two cases discussed in this three-part blog series concern Luxembourg companies that had branch offices in foreign jurisdictions, namely the United States (“U.S.”) and Malaysia. The foreign branches were set up for the exclusive purpose of holding and administering equity investments. In each instance, a manager was allocated to the branch to undertake the day-to-day management of the relevant investment portfolio. The recognition of the foreign branches as Permanent Establishments (“PEs”) by the Luxembourg tax authorities (“LTA”) was critical for availing exemptions in respect of the annual net wealth tax (“NWT”).
In Luxembourg, resident companies are subject to an annual NWT, levied at a rate of 0.5% on the company’s worldwide net worth as determined on 1 January of each year. Where a foreign PE is recognised under Article 5 of the applicable tax treaty, Luxembourg must grant relief in respect of the income attributable to, and assets allocated to that PE. In the cases at hand, the Luxembourg companies operated branches in Malaysia and the U.S. that held substantial investments. Thus, recognition of the Malaysian and U.S. branches as PEs under the Malaysia-Luxembourg and U.S.-Luxembourg DTAAs was critical. Absent such treaty-based PE recognition, the assets held through those branches would not qualify for the exemption from Luxembourg NWT, resulting in the value of these investments being included in the tax base for NWT in Luxembourg.
In both cases, the Luxembourg Administrative Court (“the Court”) ruled in favour of the LTA, concluding that the U.S. and Malaysian branch offices did not constitute PEs within the meaning of Article 5 of the applicable tax treaties. Although both cases deal with the issue of (foreign) PE-recognition and the impact on the Luxembourg NWT, the recognition of the (foreign) PEs by the LTA would also be relevant for exempting income from Luxembourg CIT by virtue of the Luxembourg participation exemption. It is conceivable that, had the Luxembourg taxpayer held the investments in the Azerbaijani companies (Case #1) and the Cayman Islands company (Case #2) directly, the Luxembourg companies might not have qualified for the benefits of the Luxembourg participation exemption regime.
It is plausible (but cannot be said for certain) that the branch offices in Malaysia and the U.S. were set up to allocate the equity investments (i.e., in the Azerbaijan companies and the Cayman Island company) to the respective PEs in order to (i) avail the exemption from Luxembourg NWT; and (ii), avail the benefits of the Luxembourg participation exemptions (i.e., effectively exempting dividends received and capital gains realised under Luxembourg CIT).
This blog is the third in a three-part series discussing the two Luxembourg court cases described above. In this blog, we present our observations and key takeaways from the two decisions.
Real and effective business activity
These cases are of particular relevance because the activities of the foreign branch offices were limited to (i) holding participations in subsidiaries; and (ii) managing intra-group financing arrangements. They did not engage in any external trading, production, or customer-facing activities. Such PEs can be regarded as non-trading PEs, characterised by limited economic substance and correspondingly, narrow functional profiles with restricted assets and risks.
In both cases, the core issue before the Court was the level of business activities/substance required to constitute a PE. In the first case, one may appreciate the difficulty the Court faced in deciding in favour of the taxpayer, as the evidence was filled with inconsistencies and contradictions. In the second case, however, it is somewhat difficult to follow the Court’s reasoning.
Under Art. 5(1) of the OECD’s Model Tax Convention,[1] a “fixed place PE” means “a fixed place of business through which the business of the enterprise is wholly or partly carried on”. This definition of “permanent establishment” in Para. 1 of Art. 5 therefore provides the three essential characteristics of a permanent establishment.[2] Firstly, there must be a “place of business”, i.e., a location, such as premises or, in certain instances, machinery or equipment. Secondly, this place of business must be “fixed”, i.e. there must be a certain degree of permanence. Lastly, the business of the enterprise must be carried out from this location. Para. 2 of Art. 5 provides a non-exhaustive list of examples for a “place of business”. However, the locations specified in the list (“a branch”, “a factory”, etc.) do not constitute PEs unless they have the essential characteristics.
However, it needs to be noted that nowhere in the OECD’s Commentary[3] to Article 5(1) and 5(2) of OECD’s Model Tax Convention on Income and on Capital (“OECD MTC”)[4] does the OECD mention the level (i.e., realness and/or effectiveness) of the business activity that needs to be carried out to constitute a PE. Article 5(1) simply requires that the business of the (foreign) enterprise is carried out from the location in question. Neither Article 5 OECD MTC nor the corresponding OECD Commentary requires that a PE must demonstrate “real and effective business activity” in the sense of active trading or revenue-generating operations.
Instead, the Commentary to Article 5 OECD MTC states that for a place of business to constitute a PE, it need not be of a productive character.[5] The Commentary further states that where tangible property has been let or leased out to third parties through a fixed place of business, then this activity shall render the place of business a PE.[6] The OECD further highlights that the same logic is also applicable in the case where capital is made available through a fixed place of business.[7] Hence, even the passive activity of extending capital through a fixed place of business may constitute a PE. The OECD does not require this to be substantiated by way of carrying out any real and effective activities. The only qualitative limitation relevant here is the preparatory or auxiliary carve-out of Article 5(4) OECD MTC. The examples listed in the OECD Commentary do not encompass holding and finance functions.
In the second case, the activity of the U.S. Branch consisted of holding and managing Company C (a Cayman Islands company). The taxpayer in the first case argues that the activity of the branch could be comparable to that of a “Luxembourg Soparfi” holding shareholdings, so there was no need for a large staff. The same argument could also be extended to the second case, suggesting that the activities of the U.S. Branch are akin to those of a passive holding company. As such, it does not require significant activities to be carried out for the purpose of holding and managing its assets.
Implicit substance threshold for non-trading PEs
Generally, tax residency disputes rarely arise in cases involving holding companies that merely own subsidiaries or intra-group loans. Revenue authorities typically do not challenge the tax residency of such entities on the basis that their activities are limited to administrative or passive holding functions. Instead, when such entities are scrutinised, the focus is usually on treaty entitlement, for instance, whether the entity qualifies as the beneficial owner of dividends or interest, or on the application of anti-abuse provisions, such as those denying treaty benefits in cases of insufficient substance or treaty shopping.
Against this background, if the corporate groups in Case 1 and Case 2 had instead operated through holding companies established respectively in Malaysia and the U.S., it is unlikely that the LTA would have challenged the residency of those entities under Art. 4 of the Malaysia-Luxembourg DTAA or Art. 4 of the U.S.-Luxembourg DTAA. In such scenarios, the discussion would more plausibly have centred on insufficient substance and treaty shopping, thereby denying treaty benefits based on provisions related to treaty entitlement, beneficial ownership, or anti-abuse measures.
Applying the same rationale to the cases at hand, it may be argued that, by examining whether the PEs possessed sufficient substance, the LTA sought to assess whether the establishment of the PEs amounted to an abuse of law or treaty shopping. In support of this position, the LTA successfully challenged the nature and scope of the PEs’ activities. While the LTA’s approach is, in that sense, understandable, it is unclear why no explicit anti-abuse provision was invoked. At least at first sight, the dispute appears to concern the characterisation of a factual constellation and whether it gives rise to a PE.
In that regard, it is also questionable whether a fact pattern involving a PE that merely holds shares in lower-tier companies and or is party to a substantial loan agreement would, in principle, be materially different from the fact patterns in both the Azerbaijan PE (Case #1) and the U.S. PE (Case #2). So, the reasoning of the Court and the LTA prompts a broader conceptual question: does the PE concept contain an implicit qualitative requirement that effectively excludes non-trading PEs, namely branches whose activities are confined to holding shares or intra-group loans without engaging in trading or service activities. If so, the distinction between a non-trading PE and a resident holding company with comparable functions and substance becomes difficult to sustain on principled grounds.
Accordingly, the Court’s decision may be read as suggesting an implicit substance threshold embedded within the PE concept. Does the reasoning of the Luxembourg Court and the LTA imply that “holding” or “financing” PEs fall outside the scope of Article 5(1) of the OECD Model Tax Convention? Put differently, the issue is whether the mere legal ownership of participations and the existence of loan agreements can, in themselves, displace the functional assessment that typically underpins the PE analysis.
Alternative focus: Allocation of ownership of loans and shares
Beyond the threshold question of whether a PE exists as a matter of Article 5(1), the LTA and the Court could also have examined whether, and to what extent, the lower-tier shareholdings and the loan agreements were attributable to the PE’s “material” enterprise. In other words, the dispute could have been framed not only as a question of PE qualification, but also as a question of allocation or attribution of the underlying assets and legal relationships to the PE, including which entity, in substance, should be regarded as the relevant owner or holder of the participations and receivables: the PE’s or the Luxembourg taxpayer.
In that respect, the indicia developed by the CJEU in the “Danish cases” (C-116/16 & C-117/16; C-115/16, C-118/16, C-119/16 & C-299/16) on dividend and interest flows, in the context of beneficial ownership and abuse, offer a potentially useful analytical template. On that approach, the assessment would not be confined to whether the PE displays sufficient substance, but would extend to whether the PE is, in functional and economic terms, the entity to which the relevant participations and debt claims should be attributed, and whether it exercises the hallmarks of control and decision-making consistent with such attribution. Among others, the LTA could have investigated:
- Conduit features: legal/contractual obligations or de facto practice to pass income upstream; short holding periods around distributions.
- Back-to-back financing: mirroring terms (amounts, maturities, rates) between inbound and outbound loans; thin spreads not reflecting real risk.
- Limited substance: absence of meaningful premises, staff, or decision-making capacity; key functions performed elsewhere.
- Cash-flow patterns: dividends/interest timed with upstream payments; funds rapidly transferred to non-EU/third-country owners.
- Control and risk: strategic decisions taken outside the U.S./Malaysia; no genuine assumption of credit/equity risk.
Substance requirements based on BEPS Action 5
The requirement of economic substance is often a de minimis threshold to prevent the misuse of shell companies for tax evasion and avoidance. Substance also plays a crucial role in assessing beneficial ownership and principal purpose tests. However, the test of economic substance for the determination of a PE is uncommon, to say the least. A threshold predominantly used in abusive cases appears to have been applied in this case. As stated above, this is a test that has not been prescribed in the case of determining a PE in the OECD Commentary.
If parallels were to be drawn between a passive holding company and the U.S. Branch in this case, one could fall back upon the guidance provided by the OECD in its BEPS Action 5 report. The report outlines the substance requirements for pure equity holding companies. It states that pure equity holding companies, which hold only equity participation and earn only dividends and capital gains, must comply with all applicable corporate law filing requirements to meet the substantial activities requirement.[8] It suggests that they should have the people and the premises for holding and managing equity participations.[9] If one were to apply this particular test of economic substance to the U.S. Branch in question, one could argue that the U.S. Branch complies with the substance requirements. Thus, the court’s judgment could be questioned.
2010 Report on Profit Attribution to Permanent Establishments
The 2010 Report on the Attribution of Profits to Permanent Establishments (“the Report”) defines the Authorised OECD Approach and provides specific and practical guidance on the application of the Authorised OECD Approach to PEs of banks and PEs of enterprises carrying on global trading of financial instruments. Under Part I of the Report, the functional and factual analysis determines the attribution of profits to the PE in accordance with its functions performed, assets used and risks assumed by the PE.[10] It aims to identify the key entrepreneurial risk-taking functions of the enterprise and the extent to which the PE undertakes these functions.[11] This is because it is the performance of those functions that leads to the assumption of the most significant risks, and the authorised OECD approach attributes economic ownership of the income-generating assets, with those functions and risks, to the part of the enterprise which performs those functions.[12]
With respect to banks, the Report emphasises that the creation and the subsequent maintenance of financial assets constitute the core value drivers of a banking business. In relation to the creation of financial assets, the report identifies a set of key entrepreneurial risk-taking functions, including, i.a., sales/marketing, sales/trading, trading/treasury, and sales/support. For the maintenance phase, the Report similarly highlights risk-taking functions: loan support, monitoring risks assumed, managing risks initially assumed and subsequently borne, treasury (managing the bank‘s overall funding position) and sales/trading (deciding whether to renew or extend the asset and, if so, on what terms), etc.[13] The attribution of profits to a banking/finance PE is determined primarily by where these functions are performed, which risks are assumed, and which assets are owned.
Importantly, however, in banking and other finance-driven structures, value creation tends to be concentrated at the point of origination or acquisition and in the corresponding assumption of risk.[14] In that sense, the initial decision to acquire, hold, or dispose of a financial asset may be significantly more value-relevant than it would be in the context of a conventional, tangible business PE, where value is often generated through continuous operational activity. This observation is particularly relevant for a “holding PE”, such as the Malaysian PE holding the Azerbaijan companies: one of the decisive moments is likely the acquisition of the Azerbaijan participations. Events occurring thereafter may carry comparatively limited weight (not least because a typical holding structure often exhibits minimal ongoing activity). In practical terms, therefore, a functional inquiry that focuses on the acquisition of the Azerbaijan companies may be more informative than an inquiry that primarily tests the day-to-day substance of the PE.
