Transfer pricing questions involving intangible assets have attracted considerable attention in recent years. The allocation of economic returns from IP among group entities gives rise to some of the most complex and contested issues in international taxation.
Since the OECD’s BEPS actions 8–10 were finalised in 2015, the concept of control over risk has assumed a central role in determining which entity is entitled to the economic returns from intangible assets. Several recent Swedish court cases have raised interesting questions about how the concept of control over risk should be applied in the context of group restructurings involving IP.
The Swedish Tax Agency (STA) has in recent years recharacterised several intra-group transactions undertaken in connection with group restructurings following the acquisition of Swedish companies by foreign groups. The post-acquisition arrangements have commonly been structured as licences of the Swedish entities’ IP to the acquiring group against running royalties.
The STA has in turn recharacterised these licences as deemed disposals of the intangible assets, on the basis that control over the associated risks has been transferred to the acquiring entity. Three lower-court decisions have now upheld that approach, and a pattern is forming that carries significant implications for multinational groups with Swedish targets in their portfolios.
The current development in Swedish case law creates uncertainty for affected taxpayers, as the recharacterisation triggers immediate exit taxation at arm’s-length value of the IP. This article examines the implications of the STA’s approach and considers whether it is consistent with the OECD Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations 2022 (the OECD Guidelines).
A recurring fact pattern
The three cases – Datawatch AB (Administrative Court of Appeal, Stockholm, case No. 4775-19, decided 2020), Twilio Sweden AB (Administrative Court of Appeal, Gothenburg, case No. 17-22, decided 2024), and Perforce Sweden AB (Administrative Court, Uppsala, case No. 7426-23, decided 2025) – share a similar structure.
In each case, a Swedish software company with established products and customer relationships was acquired by a US-based group for the purpose of integrating the Swedish software into the group’s existing operations. Following the acquisitions, the US parent was granted non-exclusive rights to distribute and exploit the Swedish companies’ IP against royalties, while the Swedish entities remained the legal owners of all intangible assets. Virtually all employees who had worked for the Swedish companies before the acquisitions, including management, remained in place afterwards, continuing to perform product development and, to varying degrees, sales and marketing functions under the strategic direction of the new parent.
In all three cases, the STA argued that the US parent had assumed control over the intangible assets through the change of reporting lines to the new HQ and that the licence structure therefore did not reflect the economic substance of the transaction. The Swedish entity should be taxed as if it had disposed of all its intangibles at market value. The STA prevailed in all three cases.
In Datawatch, the court relied primarily on the US parent’s ability to veto planned product development, without analysing whether it had the product-specific expertise to make substantive risk-related decisions.
In Twilio, the court held that all control functions related to the Swedish IP had been transferred to the US parent, and that the licence agreement in all probability never would be terminated and the assets returned, making the arrangement a transfer in substance. The court also noted that the licence payments had been structured as ‘instalment payments’, which further supported the characterisation as a disposal rather than a licence.
In Perforce, all pre-existing customer contracts remained unchanged, and the Swedish entity continued managing all prior customer relationships and retained authority to enter new ones. Notwithstanding these facts, the court found that the acquisition and subsequent restructuring entailed a “total transfer” of the Swedish IP rights, including all pre-existing customer relationships and the control over the product development function. The Perforce decision has been appealed to the Administrative Court of Appeal in Stockholm.
Control over risk: formal authority is not the same as substantive control
A notable aspect of all three decisions is that the courts appear to equate a parent company’s formal authority over its subsidiary with control over the economically significant risks associated with that subsidiary’s intangible assets. However, according to the OECD Guidelines, these are two distinct concepts that should be distinguished from one another.
The six-step risk framework in Chapter I requires that a party bearing risk under a contract must also exercise control over that risk and have the financial capacity to assume it. Control over risk, as defined in paragraph 1.65, encompasses:
The capability to make decisions to take on a risk;
The capability to decide how to respond to it; and
The actual performance of those decision-making functions.
This requires more than a formal decision-making role. The OECD Guidelines clearly state that the decision-maker must possess the competence and experience to understand the consequences of risk-related decisions and to some extent be involved in the decision-making process – not merely approving outcomes or setting broad strategic direction at board level.
In all three Swedish cases, it was undisputed that the US parent entity assumed some strategic influence over the Swedish entity’s operations upon acquisition. However, the courts provided very limited reasoning as to how the parent companies’ formal authority translated into control over the risks associated with the Swedish IP. The fact that basically all pre-acquisition personnel remained in Sweden strongly suggests that the competence needed to control product development risk to some extent stayed with the Swedish entities.
Paragraph 1.93 of the OECD Guidelines recognises that more than one entity may control the same risk simultaneously. This raises the question of whether the control over risks in the three Swedish cases more accurately could be characterised as ‘shared’.
On the facts, it appears that the US parents exercised strategic oversight while the Swedish entities retained operational and technical control over the decisions that actually shaped product development risk. If so, the contractual structure – a licence with the Swedish entity retaining legal ownership – may well have been an appropriate model for allocating the economic returns from the Swedish IP between the parties.
By collapsing a complex decision-making process into a simple question of who holds formal authority, the courts may have applied the kind of form-driven analysis that the post-BEPS OECD Guidelines were intended to move beyond.
For intangible assets specifically, the analysis warrants particular attention. Control over DEMPE-related risks demands specific knowledge of the relevant asset, such as its technical architecture, market dynamics, competitive environment, and development trajectory. It may be questioned whether full control over, for instance, a product development function transfers to an acquiring entity at the time of the acquisition, given that extensive product knowledge and development experience remains within the acquired entity. The courts’ limited engagement with this distinction raises questions about the consistency of the Swedish approach with the OECD Guidelines.
Recharacterisation requires more than a shift in control
Even if one were to accept the courts’ characterisation of where control over risk resided after the acquisitions, a further question arises as to whether the facts met the threshold for recharacterisation of intra-group transactions.
The OECD Guidelines permit recharacterisation of controlled transactions only under exceptional circumstances. Paragraph 1.142 sets out two cumulative conditions for when a transaction can be recharacterised:
The arrangements, viewed in their totality, must differ from what commercially rational independent parties would adopt; and
The structure must prevent the determination of a price acceptable to both parties given their respective perspectives and realistically available alternatives.
The OECD Guidelines further recognise that recharacterisation carries a significant risk of double taxation and should therefore be a measure of last resort.
The decisions from the Swedish courts contain limited analysis of whether the licence structures were commercially irrational. None of the courts examined whether independent parties in comparable circumstances might have adopted the same structure, or whether arm’s-length pricing could have been achieved within the existing framework – for example, through an adjusted royalty rate. The OECD Guidelines are explicit that the mere difficulty of finding comparable uncontrolled transactions does not render a controlled structure commercially irrational.
There are, in fact, sound commercial reasons why an independent licensor might grant broad exploitation rights to a licensee that assumes operational control. The licensee may command a larger distribution network, a complementary product portfolio, or superior market access – all of which were present in the Swedish cases.
A royalty structure enables the licensor to participate in the future upside generated from its historical investment in building the intangible, and to adjust its returns as the value of the asset evolves. This is particularly relevant for software assets, whose value at the point of a restructuring may be highly uncertain but whose upside potential may be substantial. By focusing on who performs and controls value-creating functions at the time of the analysis, the STA’s and the court’s interpretation of the DEMPE framework risks discounting the economic contribution of historical development activities. A licence that compensates the original developer through ongoing royalties may be entirely arm’s length – even where the licensee exercises substantial operational control going forward.
Practical implications for multinational groups
This line of case law has practical implications for multinational groups. Any post-acquisition restructuring that shifts strategic decision-making to a foreign parent risks being treated by the STA as a deemed disposal of all intangible assets at market value. For companies with valuable IP, the resulting tax liabilities can be substantial, and they materialise at the point of acquisition before the group has had any opportunity to generate returns from the integrated business.
The STA’s and the courts’ approach also does not fully account for the fact that the economic returns generated by an asset may materialise long after the underlying economic activity has been performed. Where an acquired entity has invested years of effort into developing a product, a significant portion of the post-restructuring returns may be attributable to that historical investment. Given that unrealised economic value remains embedded in the asset at the time of transfer, independent parties in comparable circumstances could well structure such a transaction as a licence, allowing the original developer to share in future upside through ongoing royalties.
The STA’s approach to recharacterisation therefore carries a risk of conflicting with the OECD Guidelines, creating uncertainty for taxpayers.
There is also a risk of divergent tax treatment across jurisdictions. Where the STA treats a transaction as a sale, the counterpart jurisdiction may well continue to treat it as a licence, potentially giving rise to double taxation.
Groups contemplating acquisitions of Swedish companies with valuable IP should consider whether a direct acquisition of the intangible assets at arm’s length – rather than a post-acquisition licence arrangement – may provide greater certainty and reduce the risk of protracted disputes with the STA.
Where a licence structure is nonetheless adopted, the post-acquisition division decision-making authority should be documented with particular care, ensuring that the Swedish entity retains substantive DEMPE functions and the personnel to perform them. Any royalty rate should also be tested against the functions the Swedish entity continues to perform, and transfer pricing documentation should anticipate the STA’s likely line of inquiry.
Other risk-mitigating actions could also be considered. For example, an open disclosure in the relevant Swedish entity’s tax return could reduce the risk of tax surcharges if the STA imposes an adjustment. Where potential STA adjustments are expected to be material, or where a high degree of certainty is otherwise desirable, applying for an advance pricing agreement should also be considered.
Looking ahead
No Swedish Supreme Administrative Court ruling has yet addressed the recharacterisation of IP licences following a group restructuring. The Perforce decision is on appeal to the Administrative Court of Appeal in Stockholm, and the outcome will be closely watched. Until the Supreme Administrative Court provides clearer guidance, the lower-court decisions establish a pattern that the Tax Agency is likely to continue following and extending.
The current state of Swedish case law is a relevant consideration for multinational groups with Swedish operations and supports the case for clearer guidance from the OECD and the Swedish courts on how control over risk should be applied in the context of intangible assets and group restructurings.
A well-reasoned decision from the Supreme Administrative Court could further clarify:
Where the line falls between a pricing adjustment and a recharacterisation;
What level of substance is required before formal authority translates into control over risk; and
How shared control scenarios should be analysed.
The status quo creates uncertainty as to how similar transactions will be treated from a tax perspective, which underscores the need for more explicit guidance.
