The treaty access problem: why your holdco may not get the lower rate

You set up a holding company in a treaty-friendly jurisdiction, expecting to pay 5% withholding tax on dividends instead of 15% or 20%. Your adviser told you the structure works. Then a tax authority challenges the claim, denies the reduced rate, and you're hit with a full withholding assessment plus interest. This scenario plays out more often than most founders expect, and it's accelerating. The treaty access problem is one of the most misunderstood risks in cross-border structuring. Tax authorities in the UAE, UK, EU, and across Asia have sharpened their tools considerably since the OECD's BEPS reforms took full effect. If your holding company exists primarily on paper, with no employees, no real decision-making, and no function beyond collecting and redistributing income, you are exposed. The days of parking a shell entity in the Netherlands, Luxembourg, or even the UAE and claiming treaty benefits by virtue of tax residency alone are effectively over. What follows is a practical breakdown of the specific rules that trip up holding structures, what substance actually looks like to an examiner, and how to build a holdco that earns its treaty position rather than merely borrowing one.

‍

The principal purpose test in one paragraph

The principal purpose test, commonly called the PPT, is the single most powerful anti-avoidance tool available to treaty partner countries. Introduced through the OECD Multilateral Instrument (MLI) and now embedded in most bilateral treaties signed or modified since 2017, the PPT asks a deceptively simple question: was one of the principal purposes of the arrangement to obtain a treaty benefit? If the answer is yes, and granting that benefit would be contrary to the object and purpose of the treaty, the benefit is denied. There is no bright-line threshold. Tax authorities assess the totality of the facts: the timing of the structure, the flow of funds, the location of decision-making, and whether the entity would exist at all if the treaty rate were unavailable. The PPT anti-avoidance provisions have already been applied in practice across multiple jurisdictions, including India, France, and the UK. If your holdco was incorporated shortly before a dividend distribution and has no other commercial rationale, expect scrutiny. The test is subjective by design, which means it gives revenue authorities broad discretion to deny claims they consider abusive. You can check whether a specific treaty has been modified by the MLI using the OECD MLI Matching Database.

‍

Beneficial ownership: the rule that empties many holding companies

Even if you survive the PPT, there's a second gate: beneficial ownership. Most tax treaties only grant reduced withholding rates to the "beneficial owner" of the income. A holding company that receives a dividend and immediately passes it upstream to a parent in a non-treaty jurisdiction is not the beneficial owner. It's a conduit. Tax authorities look at whether the recipient entity has genuine control over the income, whether it can decide how to use the funds, and whether it bears economic risk. A treaty shopping holding company that acts as a pass-through, contractually or practically obligated to remit income onward, will fail this test.

Courts have been remarkably consistent on this point. The Danish beneficial ownership cases from 2019 remain the leading authority, where the European Court of Justice denied treaty benefits to holding entities that were interposed purely to access lower withholding rates. Since then, India's tax tribunals have applied similar reasoning to Mauritius and Singapore structures, and HMRC has challenged several UK-facing arrangements. If your holdco's bank account shows dividends arriving and departing within days, with no retention, reinvestment, or independent treasury function, the beneficial ownership argument collapses quickly. The same dynamic is central to what economic substance actually looks like in practice for any cross-border structure.

‍

What commercial substance looks like to a treaty examiner

Substance is the word everyone uses but few define precisely. A treaty examiner is looking for evidence that the holding company is a real business presence, not a mailbox. The specifics vary by jurisdiction, but the pattern is consistent across most treaty partner countries.

Here's what examiners typically evaluate:

• Physical office space: not a virtual address, but premises where people actually work
• Qualified local employees or directors who make genuine decisions about the entity's investments, financing, and distributions
• Board meetings held in the jurisdiction of residence, with contemporaneous minutes showing real deliberation rather than rubber-stamping
• Independent decision-making on when and how much to distribute, reinvest, or lend
• Local bank accounts with meaningful balances that reflect the entity's actual economic activity
• Proper books, records, and locally filed tax returns showing the entity as a functioning taxpayer

A single director who also serves on 40 other boards, signing pre-prepared resolutions by email from a different country, does not create substance. Neither does renting a desk in a serviced office that you visit twice a year. The standard has moved well beyond box-ticking. Examiners in 2026 are trained to look through form to economic reality, and they have access to automatic exchange of information data that makes cross-referencing trivially easy. The OECD Model Tax Convention provides the interpretive framework examiners use when assessing these questions, and its commentary on beneficial ownership and substance has grown considerably in recent editions.

‍

The structures regulators have already taken apart

Real examples are more instructive than theory. Consider the classic "Dutch sandwich" structure, where a company routes royalties through a Netherlands entity to benefit from the Dutch treaty network and a low effective rate on outbound payments. The Netherlands itself tightened its withholding tax rules in 2021, imposing withholding on payments to low-tax jurisdictions, and the PPT now catches most remaining variations. If you're assessing where to locate IP assets within a group, the four-country comparison for IP holding companies covers the current landscape including how each jurisdiction approaches substance.

Mauritius-India structures tell a similar story. For years, Mauritius holding companies claimed capital gains exemptions under the India-Mauritius treaty. India renegotiated the treaty, and from 2017 onward, capital gains became taxable in India regardless of the Mauritius entity. Investors who relied on the old structure without restructuring found themselves with unexpected tax bills.

A more recent example involves UAE-based holding companies claiming treaty benefits on dividends from European subsidiaries. Several EU member states have questioned whether UAE entities have sufficient substance, particularly where the entity has no employees and the ultimate beneficial owners reside in high-tax countries. The fact that the UAE now imposes corporate tax at 9% has helped establish economic substance arguments, but it hasn't eliminated scrutiny. If the holdco's only activity is holding shares and receiving dividends, the question remains: why is this entity here, and would it exist without the treaty rate? The full list of UAE double taxation agreements is maintained by the UAE Ministry of Finance and shows which treaties have been modified by the MLI.

‍

Limitation on benefits, principal purpose, and borrowing a treaty

The PPT is the dominant anti-abuse mechanism globally, but US treaties use a different approach: the Limitation on Benefits clause, or LOB. The LOB is mechanical rather than subjective. It requires the treaty claimant to satisfy specific tests, such as being a publicly traded company, having a threshold percentage of ownership by residents of the treaty state, or demonstrating an active trade or business that is substantial relative to the income in question. The IRS overview of treaty LOB provisions sets out the standard tests in detail.

The distinction matters because the LOB is harder to satisfy but more predictable. You either meet the criteria or you don't. The PPT, by contrast, involves judgment calls by tax authorities and courts, making outcomes less certain. Many modern treaties include both: an LOB provision as the primary gate, with the PPT as a backstop. The 2017 Multilateral Instrument gave countries the option to adopt the PPT alone, the LOB alone, or both. Most chose the PPT, which is why it dominates practical discussions.

Borrowing a treaty, which means routing income through a jurisdiction solely to access its treaty network, is precisely what these provisions target. If your structure's primary logic is "Country A has a treaty with Country B, so we'll put an entity in Country A," you are describing treaty shopping. The entity needs an independent commercial reason to exist in that jurisdiction. Tax savings alone, no matter how well documented in your adviser's memo, are not enough. This is a reason why the choice of holding company jurisdiction has to be led by commercial logic, not treaty rate comparison.

‍

Designing a holdco that earns its treaty access

The difference between a holdco that survives scrutiny and one that doesn't usually comes down to intentional design from the start rather than retrofitting substance after a challenge arrives. The full picture of how to build a holding structure is set out in how to structure a UAE holding company, and the treaty access point is one of the central considerations there.

A well-designed holding company has a genuine regional management function. It employs people who manage the group's investments, evaluate acquisition targets, negotiate financing, and make capital allocation decisions. It holds cash reserves and reinvests income rather than distributing everything immediately. Its directors live in the jurisdiction and bring independent judgment to board decisions.

The intercompany agreements governing management fees, licensing, or service charges between the holdco and its subsidiaries should be bespoke, professionally drafted documents that reflect actual services rendered at arm's length prices. Generic templates downloaded from the internet are a red flag in any transfer pricing review. The holdco should maintain its own financial records, engage local auditors, and file tax returns that demonstrate real taxable activity.

If you're establishing a UAE holdco, the 9% corporate tax regime actually strengthens your position, provided the entity has genuine operations. A Free Zone entity claiming the 0% rate while also claiming treaty benefits faces a harder argument, because the tax authority in the source country may question whether a zero-tax entity is truly "liable to tax" as required by most treaties. The UAE FTA publishes guidance on corporate tax registration and qualifying income rules that directly affects how Free Zone entities should position treaty claims.

The treaty access problem intersects with BEPS Pillar Two for larger groups as well. Once a group exceeds the €750 million threshold, the 15% global minimum tax changes the calculus on which structures are worth maintaining, because top-up taxes in high-tax jurisdictions can erode the benefit of a reduced treaty rate entirely.

‍

A pre-clearance checklist before you rely on a treaty rate

Before you file a withholding tax reclaim or instruct a subsidiary to apply a reduced rate at source, run through these questions honestly:

• Does the holdco have at least two qualified employees or directors resident in the jurisdiction who make real decisions?
• Are board meetings held locally with documented agendas, discussions, and independent votes?
• Does the entity retain a meaningful portion of income for reinvestment or working capital?
• Can you demonstrate that the holdco would exist even if the treaty rate were unavailable?
• Are intercompany agreements bespoke, reflecting actual services at arm's length prices?
• Has the entity filed local tax returns and paid tax where applicable?
• Is there a commercial rationale for the structure beyond withholding tax reduction?

If you answered "no" to more than one of these, your treaty claim carries material risk. The cost of a denied claim is not just the tax differential: it includes penalties, interest, and the reputational damage of a structure that looks like it was built to avoid tax rather than to operate a business. The UK's list of active double tax treaties is a useful reference if any of your group entities are UK-resident or receive income from UK sources, as HMRC's approach to beneficial ownership and PPT challenges has become more systematic since 2022.

The treaty access problem is real, and it catches well-intentioned founders as often as aggressive planners. The fix is not to avoid holding companies: it's to build them properly. Invest in local substance, hire real people, make real decisions in the jurisdiction, and document everything. If your holdco earns its place in the structure, the treaty rate follows naturally. If it doesn't, no amount of legal drafting will save it when an examiner starts asking questions. Get professional advice before you rely on a treaty rate, not after the assessment lands on your desk. Cosmos builds and registers holding structures and manages ongoing tax compliance for founder-led groups across the UAE, UK, and beyond.