Leverage is not a planning device in the Netherlands; it is a commercial fact that the tax system then rations. Since the earnings stripping rule implementing article 4 of Council Directive (EU) 2016/1164 entered Dutch corporate income tax law, the deductibility of net interest has been capped by reference to fiscal EBITDA, subject to a minimum threshold below which the restriction does not apply. Since 1 January 2025 article 15b caps it at the higher of 24.5 per cent of fiscal EBITDA, raised from 20 per cent, or EUR 1,000,000 of net interest, unchanged for 2026, and disallowed interest carries forward without time limit. The threshold means a vehicle with modest borrowings is unaffected in practice; the percentage means a leveraged one usually is not. For a trading group with modest borrowings the rule is a compliance footnote. For a property vehicle carrying senior debt and a shareholder loan, or for an acquisition company set up to hold a target, it is often the single largest reason why the tax charge diverges from what the headline rate of 25.8% in the upper bracket would suggest.
How the cap is built
The mechanism is deliberately blunt. Net interest is deductible up to the greater of two amounts: a fixed percentage of fiscal EBITDA, and a minimum threshold expressed in absolute terms. Interest above that ceiling is disallowed for the year. The threshold means that smaller structures are left alone entirely; the percentage means that larger ones are measured against their own earning capacity rather than against any external notion of what a reasonable gearing ratio might be. The percentage has moved twice, from 30 per cent in the first years to the 20 and then 24.5 per cent the Belastingdienst publishes, while the threshold has stayed at EUR 1,000,000, so the current statutory figures govern, not the ones a model was built on.
Two features of the base matter more than they first appear. First, the reference figure is fiscal EBITDA, not the EBITDA the lender modelled. The corrected profit of article 15b(3) adds back to the taxable result fiscal depreciation, write downs net of reversals, and the net interest balance, none of which track the book figures. Second, the rule applies at the level of the Dutch taxpayer, which for a fiscal unity means the consolidated Dutch perimeter rather than each company individually. Where the group has left companies outside the unity, each stands on its own EBITDA and its own threshold, and the aggregate outcome can differ materially from what a group level calculation would suggest.
What counts as net interest
The measure is net, so interest income earned by the same taxpayer reduces the exposure. The definition is economic rather than formal, and reaches beyond the coupon. In practice the calculation has to be tested against the full cost of funding rather than the cash coupon alone: borrowing costs that are amortized rather than paid, discounts and premiums on debt instruments, arrangement and commitment components that are in substance a charge for the use of money, financing elements in arrangements that are debt in economic terms, currency and hedging results to the extent they relate to the borrowing, interest capitalized into an asset, and dual inclusion income under Directive (EU) 2017/952 on hybrid mismatches. Article 15b(6) lists them, and which one an item falls into is a question of characterization, better settled before the return than after it.
The capitalized interest point catches development structures repeatedly. Interest capitalized during a construction phase enters the calculation in the year of capitalization, not in the later years when the asset is written down, and where it exceeds the ceiling article 15b(7) refuses both the capitalization and the deduction, so it does not return through depreciation. Groups that model the restriction only on cash coupon consistently understate it.
Why holding and property structures feel it first
The rule is unforgiving to entities whose income is exempt. Under the participation exemption in article 13, dividends and capital gains from qualifying shareholdings are outside the tax base, which means they contribute nothing to fiscal EBITDA. A pure holding company can therefore carry a substantial acquisition loan, receive substantial distributions, and still have almost no capacity under the cap, because the income that services the debt never enters the measure that determines how much of that debt is deductible. The exemption itself is conditional, resting on a holding of at least 5 per cent and on the participation not being a non-qualifying investment participation, so the analysis has to be run before it is relied on.
Real estate produces a different version of the same squeeze. Rental income does enter fiscal EBITDA, and depreciation is added back in arriving at it, but a highly geared asset can carry net interest that is large relative to net operating income. The result is that the restriction bites hardest precisely in the years when the structure is most leveraged, which is usually the years immediately after acquisition, before amortization of the senior facility has done any work. Refinancing on tighter terms does not necessarily help either, since what matters is the interest charge relative to fiscal earnings, not the credit quality of the borrower.
The arithmetic of a leveraged acquisition
The mechanics are easier to see in index units, holding the statutory percentage and threshold abstract. Take an acquisition vehicle whose fiscal EBITDA produces a deductible ceiling of 100 units, and which carries net interest of 150 units for the year. Fifty units of interest are disallowed. Taxable profit is therefore 50 units higher than the commercial result would suggest, and at the upper-bracket rate of 25.8% that difference carries 12.9 units of tax that the model did not contain. Expressed against the interest line, a third of the borrowing cost has become an after-tax cost rather than a pre-tax one.
The disallowance also changes the shape of the return, not merely its size. Cash is used to service debt that no longer shelters income, so the vehicle needs more cash to reach the same distributable position. Where disallowed interest can be carried forward, its value depends entirely on future headroom, which in turn depends on future fiscal EBITDA; a carry forward in a structure that will remain geared for years is an asset in name and a deferred cost in substance. Article 15ba also blocks balances that arose before a substantial change in the ultimate interest in the taxpayer, subject to the usual escapes. Any deferred tax recognized on it should be tested against a realistic projection of capacity rather than against the assumption that the debt will eventually amortize.
Where transfer pricing meets the cap
The second exposure sits on the intragroup portion of the debt. Article 8b requires related party dealings to be priced at arm’s length and imposes a documentation duty with no turnover threshold, so a shareholder loan or a group facility must be supported by an analysis of the borrower’s standalone creditworthiness, the terms actually available to it, and the conduct of the parties. A rate asserted rather than evidenced is exposed on its own terms, and the exposure compounds because the same interest is being tested more than once, under different rules, with different consequences. A related party loan tied to a distribution, a capital contribution or an acquisition that makes a company related must also clear article 10a, which refuses the interest unless the debt and that transaction are predominantly business driven, or the recipient bears a levy that is reasonable by Dutch standards.
Continue the illustration. Suppose 40 of the 150 units of net interest are shareholder interest, and a review concludes that only 25 units are supportable. Fifteen units fall away as non-arm’s length before the cap is even applied. Net interest becomes 135, the ceiling is still 100, and 35 units are disallowed under the earnings stripping rule. The total non-deductible amount is unchanged at 50. The group has paid for an inflated rate and received nothing for it.
Where the cap already binds, an inflated intragroup rate buys no additional deduction at all; it buys only the exposure that comes with it.
That exposure is real. A downward pricing adjustment can carry a secondary characterization as an informal distribution, with the dividend withholding tax analysis that follows at the general rate of 15%, subject to treaty reductions and to the exemptions available within the EU. Separately, the conditional withholding tax in force since 2021, which since 2024 reaches dividends as well as interest and royalties, applies where the recipient sits in a low-taxed or listed jurisdiction, and in the abuse and hybrid cases the statute lists, regardless of how the deduction is treated at the Dutch level. Structures that place the lender somewhere other than where the funding is actually managed tend to accumulate all of these questions at once. The pricing and documentation of intragroup loans, guarantees and cash pooling arrangements is therefore not a separate workstream from the interest limitation; it is the same file viewed from a different angle.
Documentation, certainty and the limits of both
Documentation obligations scale with the group. The article 8b duty applies to everyone, while Master File and Local File requirements attach from EUR 50 million of consolidated revenue and country by country reporting from EUR 750 million. Groups reaching that figure in at least two of the four preceding years also fall within the scope of the Dutch minimum tax act, where the 15% computation starts from each group entity’s accounting result under the standard used for the consolidated accounts and behaves differently from the fiscal measure the cap uses. The two need to be reconciled rather than assumed to move together. The scope and evidential standard of the Dutch transfer pricing documentation obligation is the practical starting point for any leveraged structure.
Advance certainty is available but narrower than it once was. Under the ruling policy in force since July 2019, an advance pricing agreement requires genuine economic nexus with the Netherlands, is refused where tax saving is the only or decisive motive, and is not available for direct transactions with entities in listed jurisdictions, although a 2023 amendment carved exceptions out of two of the exclusions. A rate that only works because of where the lender is registered will not survive that filter. Where the funding function is real and the question is the margin rather than the substance, an advance pricing agreement can remove a recurring argument about a single figure.
What defensible leverage looks like
None of this argues against gearing, and none of it is a route to a lower charge. It argues for modelling the restriction at the point the debt is sized rather than at the point the return is filed. That means forecasting fiscal EBITDA rather than book EBITDA across the hold period, testing the fiscal unity perimeter against the threshold rather than assuming consolidation helps, treating capitalized and amortized finance costs as interest from the outset, and pricing intragroup debt on evidence that would survive being read by someone who did not write it. Groups that do this find the cap expensive but predictable. Groups that do not find it expensive and late.
Montclare structures and operates Dutch and cross-border holding platforms for international groups. Our services are set out on our services page.
This article is informational and does not constitute tax advice. Each engagement is subject to scope and applicable regulation.