A group operating in eight countries does not have eight transfer pricing problems. It has one, described eight times. The distinction matters, because the way most mid-sized multinationals organize the work assumes the opposite: a local adviser in each territory, each drafting a local file to local taste, each running its own comparables search, and none of them reading what the others produced. The result is documentation that is technically complete and substantively incoherent, which is the least defensible combination available. The architecture introduced by BEPS Action 13 points in the other direction, and the function it describes can be run from a single seat.
What Action 13 actually centralizes
The three-tiered structure, set out in Chapter V of the OECD Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations, is often summarized as though each tier belongs to a different owner. It does not. The Master File is a single group document. It sets out the organizational structure, the description of the group’s businesses, its intangibles, its intercompany financial arrangements, and its financial and tax positions. There is one of them, and it is prepared once.
In the Netherlands the Master File and Local File obligations attach where the multinational group had consolidated group revenue of at least EUR 50 million in the preceding reporting year, the threshold in the fourth paragraph of article 29g of the Wet op de vennootschapsbelasting 1969. Country-by-Country Reporting works the other way round: article 29c of the same law switches the filing obligation off below EUR 750 million, the figure the OECD Guidelines recommend, which puts most mid-sized groups in the first category and outside the second. Neither file is submitted with the return; both must be in the taxpayer’s records within the period set for filing the corporate income tax return, as article 29g requires.
Local Files are prepared per jurisdiction, which is where the misunderstanding starts. A Local File is jurisdiction-specific in its presentation, not in its substance. It describes the local entity, the controlled transactions it enters into, the method selected, the tested party, the financial data, and the comparables relied upon. Every one of those elements derives from the same functional analysis and, in the ordinary case, the same benchmarking set. If the German limited-risk distributor and the Polish limited-risk distributor perform the same functions, assume the same risks and use the same assets, they should be tested against the same comparable set at the same interquartile range. Two different local advisers will produce two different sets, two different ranges and, eventually, two different characterizations of the same business model.
The problem the first cross-border audit exposes
Tax administrations exchange information. The Master File is filed or made available in multiple jurisdictions and reads the same everywhere. Local Files do not, and inspectors read them alongside each other. The recurring failure pattern is not aggressive pricing; it is inconsistent narrative. The Dutch entity is described as the entrepreneur bearing market risk. The French Local File, drafted eighteen months earlier by a different firm, describes the French entity as carrying its own inventory risk and customer credit exposure. Both cannot be true, and once one administration has identified the inconsistency, the group is defending a characterization it never consciously chose.
The same fragmentation shows up in intangibles. DEMPE analysis, covering development, enhancement, maintenance, protection and exploitation, comes out of BEPS Actions 8 to 10 and sits in Chapter VI of the OECD Guidelines, which allocates the return from an intangible according to functions performed, assets used and risks assumed. It requires a group-level view of who does what and who controls the associated risks. It is not capable of being answered correctly by four advisers each looking at one entity. Where the legal owner of a trademark sits in one country and the marketing spend, the decision-making and the enforcement budget sit in three others, the allocation of return follows the functions performed and the risks controlled rather than the registration. That determination has to be made once, at group level, and then reflected consistently in every Local File.
Documentation is not defended jurisdiction by jurisdiction. It is defended as a single account of how the group actually works, and it fails at the weakest description of it.
Why the Netherlands is a workable seat for the function
There is nothing in Action 13 that requires the transfer pricing function to sit anywhere in particular, and the choice is practical rather than legal. Four features make the Netherlands a serviceable base for a group whose parent may sit anywhere. The first is the treaty network, which matters when a corresponding adjustment has to be sought against a counterparty jurisdiction through a mutual agreement procedure. The second is alignment with the OECD Guidelines. The Dutch transfer pricing decree of 14 June 2022, published in Staatscourant 2022, no. 16685, treats those Guidelines as an appropriate interpretation of article 8b of the Wet Vpb 1969, and Dutch practice works from the methods in Chapter II: comparable uncontrolled price, resale price, cost plus, transactional net margin and transactional profit split. A file built to Dutch standards is intelligible to any administration reading from the same text.
The third is working language. Article 29g requires the Master File and the Local File to be drawn up in Dutch or in English, and the profession works in English as a matter of routine. A group with a parent in Singapore or Texas does not need a translation layer between its tax function and its documentation. The fourth is dispute mechanisms. Alongside treaty-based mutual agreement procedures, the same 2022 decree points to advance certainty through the Besluit vooroverleg rulings met een internationaal karakter of 19 June 2019. For intra-EU disagreements the route is the Wet fiscale arbitrage, which implements Council Directive (EU) 2017/1852 on tax dispute resolution mechanisms in the European Union.
None of this is a planning proposition. It is a statement about where the drafting, the benchmarking and the defence of a group’s position can most efficiently be housed. Groups that are already considering where to locate a European holding entity, or that run an Asia to Europe structure through a Dutch B.V., tend to find that the transfer pricing function sits naturally alongside the entity that already carries the group’s European reporting.
The Dutch obligation underneath
Article 8b of the Wet Vpb 1969 does two things. Its first paragraph states the arm’s length principle; its third imposes a documentation duty to keep in the records showing how the transfer prices came about and whether they are conditions independent parties would have agreed. Critically, article 8b carries no turnover threshold. It applies to a Dutch entity with EUR 4 million of intercompany transactions exactly as it applies to one with EUR 400 million. The EUR 50 million and EUR 750 million figures in articles 29g and 29c govern the formal Master File, Local File and Country-by-Country tiers; they do not switch the underlying duty on or off. This is set out at greater length in our note on the article 8b documentation obligation.
Two further points shape how a centralized function has to operate. First, the Dutch anti-mismatch rules restrict downward adjustments. Article 8bb of the Wet Vpb 1969, in force since 1 January 2022, disregards a downward profit adjustment unless the taxpayer makes it plausible that a corresponding upward adjustment is taxed at the associated body. The era of informal capital contributions producing unilateral Dutch deductions has ended, and any position that depends on an unmatched downward adjustment should be assumed to fail. Second, and consequently, the correspondence between the Dutch position and the foreign position is no longer a matter of tidiness. It is a condition of the Dutch outcome.
A worked example
Take a group with a parent outside Europe, consolidated revenue of EUR 210 million, and operating subsidiaries in the Netherlands, Germany, France and Poland. It is above the EUR 50 million threshold in article 29g of the Wet Vpb 1969 and below the EUR 750 million figure in article 29c, so a Master File and Local Files are required and no Country-by-Country Report is due. The figures that follow are illustrative rather than quoted, and adviser pricing varies widely.
Under the decentralized approach, assume four local advisers each preparing a Local File at roughly EUR 22,000 and each commissioning a separate benchmarking search at roughly EUR 9,000, with the parent’s adviser assembling a Master File at EUR 30,000. Total annual spend of approximately EUR 154,000, producing four documents with four comparable sets and no single owner of the analysis article 8b of the Wet Vpb 1969 requires them to evidence.
Under the centralized approach, one functional analysis and Master File at approximately EUR 45,000, one consolidated benchmarking exercise at approximately EUR 20,000, and four Local Files prepared as wrappers around the common core, at approximately EUR 8,000 each. Total of roughly EUR 97,000, a difference of about EUR 57,000 a year.
The fee difference is the smaller half of the argument. Assume the German Local File describes the German subsidiary as a full-risk distributor while the group’s Dutch documentation treats it as limited-risk. On audit, the German administration reprices on the basis it finds in the German file and raises the German taxable base by EUR 1.2 million across the period reviewed. The corresponding downward adjustment in the Netherlands is not automatic, and under article 8bb it has to be substantiated by the foreign upward inclusion rather than asserted. Until that is achieved, the same EUR 1.2 million of profit remains in the Dutch base. On the 2026 rates, which the Belastingdienst publishes as unchanged from 2025 and which the rate table in article 22 of the Wet Vpb 1969 sets at 19 per cent up to EUR 200,000 and 25.8 per cent above, that profit falls in the upper bracket, or EUR 309,600, on top of the German liability. Add interest, the cost of running a mutual agreement procedure, and the internal time absorbed over the two to three years such a case typically occupies. A EUR 57,000 saving is not a saving.
What centralization does not mean
It does not mean dispensing with local input. Filing formats, deadlines, penalty regimes and language requirements differ, and a local signature is often still required. What changes is the direction of travel: the functional analysis, the method selection, the benchmarking and the characterization of each entity are decided once, at the centre, and local advisers are engaged to conform the output to local requirements rather than to originate the analysis. That is a materially different, and materially cheaper, scope of work.
It also does not mean the file is written once and left alone. Intercompany arrangements change, entities take on new functions, and comparable sets require refreshing. A centralized function makes that maintenance tractable, because there is one document set to update rather than four to reconcile. Groups that treat documentation as an annual filing exercise generally discover the gap at the point where it is most expensive to close.
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. Each engagement is subject to scope and applicable regulation.