A holding company is, in the end, a promise about cash flow. It promises that a dividend leaving an operating subsidiary, or interest leaving a borrower, will reach the ultimate investor bearing the withholding cost that the applicable treaty or directive appears to permit. Whether that promise survives contact with an assessing authority no longer turns on the wording of the treaty alone. It turns on two questions that are asked before any reduced rate is applied: whether the recipient is the beneficial owner of the income, and whether obtaining the benefit was one of the principal purposes of the arrangement that produced it. Both questions are about facts, not about drafting.
What a treaty rate is not
Groups routinely describe a structure by its headline outcome: a nil or reduced rate on dividends, a reduced rate on interest, no exit tax on a share disposal. That description confuses the result with the entitlement. A treaty allocates taxing rights between two states in respect of a resident of one of them who receives income arising in the other. Every element of that sentence is a condition. Residence must be real and not merely certified, and the recipient must occupy the position the treaty contemplates, which is that of a person who receives the income for itself.
The Dutch domestic position illustrates the point rather than contradicting it. Corporate income tax applies at 25.8% in the upper bracket, with a reduced rate on the first band. Dividend withholding tax is 15% as a starting point, with reductions under treaty and exemptions within the European Union, and a conditional withholding tax has applied since 2021 to interest and royalties paid to low-taxed or listed jurisdictions. Each of those reliefs is expressly qualified by anti-abuse conditions. The relief is not the rule with abuse as an exception; the relief is available only where the conditions, including the anti-abuse conditions, are satisfied.
Beneficial ownership: control, not receipt
Beneficial ownership is the older of the two tests and the more precise. It asks who has the power to determine what happens to the income once it arrives. A person who merely takes delivery and passes the amount onward does not have it, whatever the register of shareholders or the loan documentation says.
The analysis is therefore economic. It looks at whether the recipient can decide to retain the income, to reinvest it, to apply it against its own liabilities, or to distribute it on its own terms and at a time of its choosing. Where those decisions are genuinely open, the recipient is exercising ownership. Where they are foreclosed, whether by contract, by a parallel obligation, by the terms of an upstream instrument or by a practice so settled that the outcome was never in doubt, the recipient is administering a payment rather than owning it.
Two consequences follow that groups often underestimate. First, beneficial ownership is tested at the level of the income, not the entity. A company may be the beneficial owner of one stream and not of another. Second, the test is applied at the moment the income is received. A structure that becomes substantive after the distribution does not repair the position for that distribution.
The obligation to pass the income on
The sharpest fact pattern is the one where an entity receives income and pays an equivalent amount onward under an obligation that already existed. The obligation need not be expressed as a duty to remit that particular receipt. It is enough that the entity has committed, before the income arrives, to make a matching payment whose funding depends on it. Back-to-back loans on mirrored terms, a distribution policy fixed by an upstream instrument, a mandatory redemption schedule keyed to underlying receipts: all describe an entity whose discretion has been contracted away in advance.
An entity that is bound before the money arrives has not acquired income; it has acquired an instruction.
Timing and quantum are the evidential markers. Where the onward payment follows the receipt closely, where the amounts correspond after a margin that is small and formulaic, and where the entity holds no meaningful assets from which the onward payment could otherwise be made, the inference of a prior obligation is difficult to displace. Conversely, an entity that retains earnings, that funds its obligations from a balance sheet capable of bearing them, and whose distributions are decided by a board applying commercial judgement, presents a materially different picture. Financing structures deserve particular attention here, because leverage introduces both the beneficial ownership question and the separate quantitative constraint on deductions under the ATAD earnings-stripping rule, which limits deductible net interest by reference to fiscal EBITDA subject to a minimum threshold and whose parameters have changed over time.
The principal purpose test and the weight of one
The principal purpose test operates differently and, for most groups, more severely. It denies the benefit where, having regard to all relevant facts and circumstances, it is reasonable to conclude that obtaining that benefit was one of the principal purposes of an arrangement or transaction, unless granting it would accord with the object and purpose of the relevant provisions.
Three features of that formulation do the work. It is one of the principal purposes, not the sole or dominant purpose; a structure can have genuine commercial objectives and still fail. The standard is that it is reasonable to conclude, which is a lower threshold than proof and which permits inference from objective facts. And the saving limb requires the taxpayer to show that the benefit is consistent with the object and purpose of the provisions, which is a positive case to be made rather than a presumption to be enjoyed.
The practical effect is that intention is reconstructed from evidence. Board minutes recording that a jurisdiction was selected for the treaty network, advisory material comparing withholding outcomes across candidate locations, an interposition that coincides with a transaction rather than with any operational change: each supports the inference. Purpose is not what the group believed; it is what the record shows the group was doing.
Substance as a condition of access
Substance used to be discussed as a defensive add-on, ticked to strengthen a position already thought sound. It is now closer to a condition of access. The two tests converge on the same evidence: a company that has people with the authority and competence to take decisions, that takes them where it is resident, that bears the risks associated with the assets it holds, and that has the financial capacity to meet its obligations, is a company that can plausibly claim to control its income and to exist for a reason other than the treaty.
Formal indicators still matter, but their weight has shifted. Local directors, a registered office, accounts prepared and filed in the jurisdiction and banking operated from it are necessary but not sufficient. What is examined is decision-making: whether the board has anything real to decide, whether it exercises judgement rather than ratifying instructions from elsewhere, and whether its members could explain the group’s financing and holding arrangements without assistance. Our discussion of the current Dutch substance requirements sets out the components in detail.
The Dutch ruling practice reflects the same standard from the administration’s side. Since July 2019, advance certainty is conditioned on a real economic nexus with the Netherlands; it is not available where the decisive motive is tax saving, nor in relation to listed jurisdictions. A group that cannot obtain a ruling because it lacks nexus should treat that as information about its position generally, not merely about its access to a ruling.
Business reason, documented when it is formed
The commercial rationale for a structure is most persuasive when it is recorded before the structure is used. A holding platform established to consolidate ownership across several jurisdictions, to provide a single counterparty for lenders, to house a management team, to standardise governance across acquisitions, or to prepare for an eventual exit, is doing recognisable work. That work should appear in the corporate record at the time: the resolution that approves the incorporation, the mandate given to the board, the service agreements that give it something to administer.
Contemporaneous documentation also aligns with obligations that operate independently of treaty access. Transfer pricing documentation under article 8b applies without threshold and must reflect what the entity actually does, a point developed in our note on the Dutch transfer pricing obligation. Reportable cross-border arrangements fall within DAC6 by reference to hallmarks, not by reference to whether the group considers the arrangement aggressive. A file constructed after an enquiry begins invites the reading that the rationale was constructed then too.
What follows for structure design
The design question has changed from whether a structure qualifies on its face to whether it would be recognised as genuine by someone reading the file without goodwill. That reframing has consequences. Interposing an entity solely to reach a better treaty rate is not a plan. Layering entities to achieve incrementally better outcomes multiplies the points at which beneficial ownership can be challenged. Financing arrangements that mirror one another across a chain create the very correspondence that the analysis looks for.
What tends to survive scrutiny is the structure that would exist in some form even if the withholding position were neutral: a holding company that owns qualifying participations and may rely on the participation exemption because it holds them for its own account and meets the exemption’s own conditions, that is capitalised to carry the risks it takes, and whose board makes decisions that would be recognisable as decisions to any commercial observer. The starting point for that analysis is set out in our note on the Dutch participation exemption.
Access is granted on the basis of what a company is, not what it is called. Substance and commercial rationale are not the decoration on a structure; they are the conditions on which its central promise depends.
Montclare structures and operates Dutch and cross-border platforms for international groups. Our services are set out on our services page.
This article is informational and does not constitute tax, legal or investment advice. Each engagement is subject to scope and applicable regulation.