Treaty Shopping for Normal People

Double Tax Agreements (DTAs) are a bit like VIP passes for cross-border income. They can reduce withholding taxes, determine which country gets to tax particular income and, most importantly, stop the same income being taxed twice.

But what happens when you arrange your affairs specifically to get hold of a particularly attractive VIP pass?

That is where treaty shopping comes in.

What is treaty shopping?

Imagine that Country A charges 15% withholding tax when a dividend is paid directly to a shareholder in Country B. But Country A has a tax treaty with Country C under which the withholding tax is only 5%.

It might be tempting for the shareholder in B to insert a company in C:

Country A → Country C company → Country B shareholder

The dividend now appears to qualify for the 5% treaty rate.

Historically, structures like this could work surprisingly well. Today, however, simply putting a company in the middle is nowhere near enough.

Depending on the treaty and the type of income, tax authorities may ask several separate questions:

• Is the company actually tax resident in the treaty country?
• Where the relevant treaty article requires it, is the company the beneficial owner of the income?
• Does it satisfy any Limitation on Benefits (LOB) requirements?
• Does the arrangement survive the Principal Purpose Test (PPT) or other anti-abuse rules?

These tests overlap, but they are not the same thing.

Beneficial ownership: who really controls the income?

Beneficial ownership is particularly relevant to treaty articles dealing with dividends, interest and royalties.

Very broadly, the question is whether the treaty-country company genuinely has the right to use and enjoy the income, or whether it is really just receiving it on somebody else’s behalf.

A classic warning sign is a conduit company which receives EUR 10 million of interest on Monday and is legally or commercially obliged to pay substantially the same EUR 10 million onwards on Tuesday.

But simply paying dividends to your own shareholders does not automatically mean that you are not the beneficial owner.

That distinction was illustrated in the well-known Canadian case Prévost Car. A Dutch holding company received dividends from a Canadian subsidiary and ultimately distributed money to its UK and Swedish shareholders. The Canadian courts nevertheless accepted the Dutch company as the beneficial owner: it was the legal owner of the dividends and was not simply an agent or nominee required to pass each payment straight through.

At the other end of the spectrum is Indofood, a UK case involving Indonesian debt and an attempted restructuring through Mauritius. The proposed Mauritian intermediary did not have sufficient freedom over the interest it would receive and was effectively a conduit.

The practical point is important: beneficial ownership is about the recipient’s substantive right to use and enjoy the relevant income, not merely who ultimately owns the company receiving it.

The principal purpose test: why did you do it?

The modern treaty-shopping analysis goes considerably further.

Following the OECD’s BEPS project, many treaties now contain a Principal Purpose Test, or PPT.

The test is deliberately broad. In general terms, a treaty benefit may be denied where, having regard to all the 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 the benefit would nevertheless be consistent with the object and purpose of the relevant treaty provisions.

There are two important points here.

First, tax does not have to be the only purpose.

Second, having tax considerations does not automatically cause a structure to fail. Businesses are entitled to consider tax when deciding where and how to invest. The second limb of the PPT matters: the question is also whether allowing the benefit is consistent with what the treaty provision was intended to achieve.

Mauritius has adopted the PPT as its principal anti-abuse mechanism under the OECD Multilateral Instrument for treaties covered by the MLI. The India-Mauritius treaty is a special case: it was not modified through the MLI because Mauritius did not list India as a covered treaty. India and Mauritius instead signed a separate Protocol in 2024 to introduce a PPT, and Mauritius approved ratification of that Protocol in 2026.

And then there is the LOB test

Some treaties also contain a Limitation on Benefits, or LOB, provision.

The distinction is useful:

• PPT asks why the arrangement was structured that way.
• LOB asks whether the taxpayer satisfies specified objective conditions for the treaty benefit.

An LOB may look at matters such as ownership, stock-exchange listing, active business activity or onward payments to persons who would not themselves qualify for the benefit.

So a company can be a genuine resident of a treaty country and still fail a particular LOB provision.

Tiger Global: the case everyone using Mauritius should know about

That brings us to the major recent case.

Tiger Global involved Mauritian investment companies in the Flipkart structure. In 2018, when Walmart acquired Flipkart, the Mauritian entities sold shares in the Singapore company through which the Indian business was held.

The companies held Mauritian Global Business licences and Mauritian Tax Residency Certificates. They sought to rely on the India-Mauritius treaty in relation to the gains.

In 2024, the Delhi High Court found in Tiger Global’s favour. Among other things, it placed considerable weight on the companies’ Mauritian Tax Residency Certificates and rejected the proposition that investing through a tax-favourable jurisdiction was, by itself, enough to establish treaty abuse.

But in January 2026 the Indian Supreme Court overturned that judgment.

The important point is not that the Supreme Court announced a rule that Mauritian structures do not work, or that a Mauritian TRC is worthless. It did neither.

The case arose from applications for advance rulings. The Authority for Advance Rulings had refused to entertain them because it considered the transactions prima facie designed for the avoidance of Indian income tax. The Supreme Court held that the AAR was entitled to do so and restored that conclusion.

In reaching its decision, the Court was prepared to look well beyond the formal existence of the Mauritian companies.

The wider structure was ultimately connected to Tiger Global in the United States. The authorities had found that significant control sat outside Mauritius, including authority over major banking transactions, and that the Mauritian entities’ investment activity was concentrated in Flipkart. The Court treated these and the surrounding circumstances as relevant to whether the structure had genuine commercial substance or represented an impermissible tax-avoidance arrangement.

The lesson from Tiger Global therefore needs to be stated carefully.

A Tax Residency Certificate remains important evidence of residence. But it is not necessarily an impenetrable shield against an enquiry into treaty abuse, commercial substance or where meaningful control actually sits.

Residence, beneficial ownership and PPT are different questions

This is where people often get confused.

A company can be incorporated in Mauritius, tax resident in Mauritius and hold a valid Tax Residency Certificate, but another treaty question may still remain.

Residence asks: Where is this company resident for treaty purposes?

Beneficial ownership, where the relevant treaty article uses that concept, asks: Does this company genuinely have the right to use and enjoy the income?

The PPT asks: Was obtaining this treaty benefit one of the principal purposes of the arrangement and, if so, would granting it nevertheless be consistent with the object and purpose of the treaty provision?

An LOB asks: Does the taxpayer satisfy the treaty’s specified eligibility conditions?

Passing one test does not necessarily answer the others.

Substance helps, but ticking boxes isn’t enough

This is also why the modern discussion about “substance” needs some care.

Substance is not one universal treaty test. Rather, the underlying facts may provide evidence relevant to residence, beneficial ownership, the PPT, control and commercial purpose.

Having Mauritian directors, a bank account, accounting records, audited financial statements and an office can all be relevant. In many cases some of these are also regulatory or residence requirements.

But the paperwork needs to reflect what actually happens.

If board minutes say that Mauritian directors made a USD 100 million investment decision, while the emails show that somebody in London, New York or Johannesburg made the decision and simply told the directors what to sign, beautifully drafted Mauritian board minutes will not solve the problem.

The stronger structures therefore tend to have a commercial explanation which makes sense before anyone calculates the withholding-tax saving.

• Why is the company in Mauritius?
• What does it actually do there?
• Who makes its important decisions?
• Does its board have real authority?
• Can it decide what happens to its money?
• Does it bear meaningful risks?
• Would there still be a sensible commercial reason for the structure if the treaty benefit were reduced?

Those questions increasingly matter more than the number of pages in the company’s “substance file”.

So is treaty planning dead?

No.

There is an important difference between using a treaty and abusing a treaty.

DTAs exist precisely to facilitate cross-border investment, allocate taxing rights and prevent excessive or duplicate taxation. Businesses are entitled to structure investments with knowledge of the applicable tax consequences.

What has become much harder is inserting an entity whose principal function is simply to borrow another country’s treaty network.

The progression can perhaps be summarised like this:

• Old world: “We have a Mauritian company and a TRC, so we get the treaty.”
• Better analysis: “Is the Mauritian company actually entitled to the relevant treaty benefit, including any beneficial-ownership requirement?”
• Modern analysis: “Is it genuinely resident, is it the beneficial owner where that test applies, does it satisfy any LOB, what are the principal purposes of the structure, who actually controls it, and is granting the treaty benefit consistent with the purpose of the treaty?”

That is a much more demanding analysis.

The takeaway

Treaty planning has not disappeared. But access to treaty benefits has become much more sophisticated.

Prévost tells us that paying income onwards does not automatically turn a holding company into a conduit.

Indofood shows why a recipient with insufficient freedom over income may not qualify as its beneficial owner.

The PPT asks us to examine why the structure exists and whether the claimed benefit is consistent with the treaty’s purpose.

And Tiger Global is a timely reminder that licences, local directors and a Tax Residency Certificate do not necessarily prevent a tax authority or court from examining the commercial reality behind a structure.

For businesses using Mauritius as a holding, financing or investment jurisdiction, the question should therefore no longer simply be:

“Can we access this treaty?”

It should be:

“Why are we using Mauritius, what genuinely happens here, and can we defend the structure when somebody looks beyond the paperwork?”

That is the question modern treaty planning needs to answer.

Get in touch if you would like to chat further about this topic.