Of all the intercompany flows a Dutch inspector can open a file on, the management fee is the one that gets looked at first. It is visible, it is recurring, it is deductible in the paying jurisdiction, and it is very often reconstructed after the fact. A financing arrangement at least leaves loan documentation and interest schedules. A service recharge frequently leaves nothing behind but an invoice line reading “management fee” and a figure that has not moved in four years. That combination draws the question, and the taxpayer is then the one who has to explain how the price was arrived at.
What article 8b actually asks of a service charge
Article 8b of the Wet op de vennootschapsbelasting 1969 does two things at once, and groups routinely remember only the first. It requires that conditions between associated enterprises be arm’s length, and its third paragraph imposes a standing obligation to carry in the company’s own records the information showing how those conditions were determined. There is no turnover threshold. A Dutch B.V. with two employees and a single intercompany invoice is subject to the same substantive obligation as a group with revenue in the billions. What scales with size is only the form the documentation takes: Master File and Local File become mandatory under article 29g of the same Act for group entities of a multinational group with at least EUR 50 million of consolidated group revenue in the reporting year preceding the return, and Country-by-Country Reporting is required under article 29c once that consolidated figure reaches EUR 750 million. Below those thresholds the duty to substantiate does not disappear; it simply has no prescribed template, which is precisely why so many files below the line are empty. This is developed in our note on the article 8b obligation, the foundation of this series.
The five things that make a charge defensible
A charge that survives examination is one that can be taken apart and reassembled by someone who was not there when it was set. Five elements carry the weight.
The first is benefit to the recipient. The test is whether an independent enterprise in the recipient’s position would have been willing to pay for the activity, or would have performed it in-house. Duplicative activities and incidental benefits arising simply from group membership fail this test. On-call arrangements are chargeable only where the standby capacity meets an identified need and an independent party would have paid a retainer for it, which is the position taken in the OECD Transfer Pricing Guidelines on intra-group services, where standby charges are treated as unlikely between independent parties when the potential need is remote or the same capacity can be obtained promptly elsewhere.
The second is an identifiable cost base. The charge has to trace back to a defined pool of costs in the accounts of the provider, with a clear statement of what has been excluded and why. A base that cannot be reconciled to the general ledger is not a base.
The third is a rational allocation key. Headcount, seats, external spend, revenue, transaction volume; the key must bear a demonstrable relationship to the benefit received, be applied consistently, and be capable of producing a different answer next year when the underlying driver changes.
The fourth is a mark-up supported by something. For routine support services the natural methods are cost plus or a TNMM applied on a net cost plus basis. A CUP is available where genuinely comparable third-party services exist. Resale price rarely fits services, and profit split is appropriate only where both parties make unique and valuable contributions, which is almost never true of shared-service centres. Where the services genuinely fall within the narrow category of low value-adding intra-group services, the Verrekenprijsbesluit 2022 endorses the simplified approach under which the relevant costs are recharged through an appropriate allocation key with a fixed mark-up of five per cent and no benchmarking study behind it. That concession is confined to services of that kind, and it is not a general licence.
The fifth is a contract that matches what happened. Where the written arrangement and the conduct of the parties diverge, the conduct governs the analysis. If the agreement describes strategic oversight and the evidence shows bookkeeping, the file will be analysed as bookkeeping.
If what is being invoiced as a “management fee” in substance covers the development, enhancement, maintenance, protection or exploitation of intangibles, it leaves the services box entirely. The DEMPE analysis set out in the OECD chapter on intangibles then determines where the return belongs, and a cost plus five per cent recharge is unlikely to be the answer.
Shareholder costs, and why they never leave home
The single most common defect in a management fee file is the presence of shareholder costs in the cost base. These are activities the parent performs in its capacity as owner rather than as a provider of services: preparation of the group’s consolidated accounts, the parent’s own audit and statutory compliance, general meetings and board of the parent, investor relations and listing obligations, and the costs of managing and restructuring the group’s own shareholdings, including acquisitions and disposals of participations. None of these confer a service on the subsidiary that the subsidiary would have bought from a third party. They are borne by the shareholder, and they stay with the shareholder. The list of shareholder activities in the Dutch transfer pricing decree reaches the same conclusion and states plainly that no charge should be raised to other group companies for them, while treating consolidation work and merger and acquisition work as mixed activities that have to be split, so the part of an acquisition that demonstrably serves the acquiring subsidiary can still be charged on.
A cost is a shareholder cost when the only party whose interest it serves is the shareholder. It does not become a group service merely because it was inconvenient to leave it in the parent’s own profit and loss account.
The distinction matters more than its size in the accounts suggests: a base contaminated with shareholder costs undermines confidence in the whole exercise, and an inspector who finds group consolidation costs in the pool will not stop there.
A worked example
Take a group whose central service company sits outside the Netherlands and whose Dutch operating company is one of several recipients. The arithmetic applies whichever side of the invoice the Dutch entity is on.
The central functions cost EUR 4,200,000 in the year: group IT, HR administration, procurement support, treasury operations and financial reporting support. The group identifies EUR 400,000 of shareholder costs and removes them. It also identifies EUR 300,000 of third-party software licences procured centrally and recharged as pass-through, at cost and without mark-up; the Dutch guidance on cost bases and costs of a disbursement character allows those amounts to sit outside the base on which the mark-up is computed, but the intermediary characterization that permits this has to be documented rather than assumed. The marked-up cost base is therefore EUR 3,500,000.
A blended key built from headcount, seats and external spend allocates 27 per cent of that base to the Dutch B.V.: EUR 945,000. A benchmarked net cost plus mark-up of 5 per cent adds EUR 47,250, giving EUR 992,250. The pass-through element adds EUR 81,000 at cost. A separate direct charge of EUR 90,000 plus 5 per cent, EUR 94,500, is raised for legal work on a corporate project. Total invoiced to the B.V.: EUR 1,167,750.
On examination, two elements do not hold. First, a further EUR 240,000 of parent-level costs is found to belong in the shareholder category; the B.V.’s 27 per cent share is EUR 64,800, and with mark-up EUR 68,040 falls away. Second, the legal work relates to the disposal of a sister company; the B.V. derives no benefit from it, and the full EUR 94,500 is disallowed. The correction to the Dutch taxable base is EUR 162,540. At the 25.8 per cent rate that the Belastingdienst publishes for 2026 above the EUR 200,000 first bracket, the additional Dutch corporate income tax is roughly EUR 41,935, before interest and any penalty.
The cash figure is not the real cost. The counterparty has already recognized the corresponding income and will not reverse it spontaneously, so the group is left with economic double taxation and a mutual agreement procedure to run, under the applicable treaty or the dispute resolution mechanism established by Council Directive (EU) 2017/1852. Where the counterparty is the shareholder, an amount paid in excess of arm’s length can also be recharacterized as a hidden profit distribution, since the Wet op de dividendbelasting 1965 brings direct and indirect distributions of profit into charge under whatever name or in whatever form they are made, and Dutch dividend withholding tax then enters the frame. A EUR 162,540 documentation failure becomes a multi-year, multi-jurisdiction problem.
What sinks a charge in practice
Inspectors do not begin with theory. They begin with three observations visible from the outside. A round figure that repeats year after year, EUR 500,000 in each of four years while headcount, revenue and activity all moved, tells them the number was decided rather than calculated. An absence of any record of who did what, no time capture, no service catalogue, no evidence of requests made and work delivered, means the benefit test cannot be evidenced at all. And shareholder activities invoiced as group services, particularly anything touching acquisitions, disposals, group financing or reporting to the parent’s own stakeholders, indicates that the base was never properly built.
Two further defects appear regularly: an intercompany agreement signed years ago describing services the provider no longer performs, and a mark-up quoted as market practice with no benchmarking study behind it. Neither is fatal on its own. Together with the first three, they leave nothing to defend.
Adjustment mechanics and the 2022 mismatch rules
The Wet tegengaan mismatches bij toepassing zakelijkheidsbeginsel entered into force on 1 January 2022 and applies to financial years beginning on or after that date. It limits downward adjustments where the taxpayer cannot show that a corresponding upward adjustment is picked up in a profits tax at the associated party. In the service-charge context this removes a symmetry groups used to rely on. A Dutch entity that has been undercharged, or overcharged, cannot assume that restating the position to arm’s length in the Netherlands will be accepted if the other side does not take the matching amount into its own base. The explanatory memorandum to that Act is explicit that the informal capital structures which previously produced a Dutch profit reduction with no pick-up abroad are no longer available.
The practical consequence is that pricing errors in service flows now tend to be one-directional in effect. Upward corrections come readily; downward corrections require the counterparty position to move first. That asymmetry should shape how carefully the base and the key are built at the outset, particularly in structures spanning several jurisdictions, a theme we return to when discussing cross-border holding arrangements.
Building the file before it is asked for
The work required is unglamorous and finite. Define the service catalogue. Reconcile the cost pool to the ledger and document the exclusions, shareholder costs first. Choose keys that reflect drivers and record why each was chosen. Benchmark the mark-up and refresh the study on a stated cycle. Rewrite the intercompany agreements so they describe the services actually delivered, and keep contemporaneous evidence of delivery. This is best fixed when the structure is set up rather than four years later, and the same discipline applies when establishing a Dutch holding structure.
None of this makes a charge immune from challenge. It changes what the challenge is about: a discussion of allocation keys and comparables, conducted from a file, rather than whether anything was provided at all.
Montclare has published a short self-assessment, the Transfer Pricing Readiness Check, which sets out ten questions that identify where a group’s exposure sits. It can be downloaded from our transfer pricing page.
This article is informational and does not constitute tax advice. Rates and thresholds cited are those in force in the Netherlands in 2026 and were verified against the official sources linked above. Each engagement is subject to scope and applicable regulation.