There and back again – trustee residence for treaty purposes

Haworth v HMRC [2022] UKFTT 00034 is an example of the so-called "round the world" scheme to avoid UK capital gains tax on a share disposal that was examined at length in the line of cases culminating in Smallwood v Revenue and Customs Comrs [2010] EWCA Civ 778.

The scheme

In Haworth, the Jersey trustees of a family trust held shares in a UK incorporated company that was soon to be listed during the peak of the "dot-com bubble" in 2000. The Jersey trustees took advice from UK accountants and leading counsel in relation to their potential liability to UK capital gains tax on a disposal of some of their shares. As a result, the following scheme was designed (leaving out many of the intricacies involved):

• the Jersey resident trustees would resign in favour of Mauritian resident trustees;
• the shares in the UK company would be sold by the Mauritian resident trustees; and
• the Mauritian resident trustees would resign in favour of UK resident trustees.

All of these steps were to occur in the same UK tax year.

The Mauritian resident trustees would not incur capital gains tax in Mauritius because none would be chargeable as a matter of Mauritian tax law. (The trustees would, however, be liable to Mauritian income tax on trust income.) The idea was that under the UK-Mauritius Double Tax Treaty there would also be no charge to UK capital gains tax for the Mauritian resident trustees. This was because Article 13(4) provided that capital gains should only be taxable in the contracting state of which the trustees (i.e. the alienator) were resident. At the time of the disposal, this would be Mauritius and so Mauritius would have sole taxing rights on the capital gain realised (even though no capital gains tax would actually be charged in Mauritius).

Critical to the argument of both the Smallwood cases and Haworth was the correct interpretation of Article 4(3), which provided a tie-breaker test for determining the treaty residence of a person who was liable to tax in both contracting states. Article 4(3) stated that a person "shall be deemed to be a resident of the contracting state in which its place of effective management is situated."

Who is a trustee for tax treaty purposes?

HMRC ran the argument that, simply put, the Mauritian trustees and the UK trustees were different entities. This meant that the UK trustees could not claim the benefit of the treaty because the UK trustees were not resident in Mauritius and were not liable to Mauritian tax at all. Put another way, the trustees deemed to exist for UK and Mauritian tax purposes were different persons for the purposes of the treaty.

Judge Morgan rejected this argument. Article 3(1)(e) defined "person" as including a "body of persons" and she considered that it was uncontroversial that trustees should be viewed as "a body of persons" rather than individually. Citing the Commerzbank decision, Judge Morgan noted that it was important "to interpret a treaty unconstrained by technical rules of domestic law and on broad principles of general acceptance" and pointed out that both the relevant UK and Mauritian law deem trustees to be a "trust body".

What is the relevant period for testing the treaty residence of a trustee?

One of the arguments that the taxpayers ran in both Smallwood and Haworth was that at the time of the share disposal the trustee was not UK resident and so Article 4(3) – the tie-breaker test for determining an entity’s residence under the treaty – should not be engaged. Instead, Article 13(4) should simply apply to the trustees as Mauritian resident and so prevent any UK capital gains tax liability. In both cases, this argument was rejected by the Courts.

In the Court of Appeal decision in Smallwood, Patten LJ noted that Article 4(1) defined residence in terms of liability to taxation and so it was necessary to consider whether an entity would be liable to tax in the contracting state based on any applicable residence period. This meant that even though at the date of disposal the trustees had not been UK resident, they became UK resident subsequently and so were treated as UK resident for that UK tax year. This meant that under Article 4(1) the trustees would be resident in both the UK and Mauritius when considering the disposal and so the tie-breaker test in Article 4(3) was engaged.

Judge Morgan agreed.

How should the "place of effective management" test be applied to a corporate trustee?

The Court of Appeal in Smallwood disagreed on how the tie-breaker test should be applied to a corporate trustee. Patten LJ, dissenting, considered that the "place of effective management" test should be applied to a corporate trustee in accordance with the principles of Wood v Holden. This meant considering where the corporate trustee was effectively managed and controlled at the time of the disposal, taking into account whether this management and control function had been usurped by a third party.

By contrast, Hughes LJ considered that it was the place of effective management of the trustees as a continuing body that should be tested, not the place of effective management of the corporate trustee who happened to be in office at the time of the disposal. This meant that the principles of Wood v Holden were less relevant. Instead, quoting the Commissioners at first instance, one had to ask where the "real top-level management" of the trust took place, which could be distinct from the day-to-day management of the trustees in office at the time. Hughes LJ concluded in Smallwood that there was "a scheme of management of this trust which went above and beyond the day-to-day management exercised by the trustees for the time being."

So, crucially, under this approach there was no need to find that the management and control of the Mauritian trustees in place at the time of the disposal had been usurped by the UK trustees or the UK advisers on a Wood v Holden basis. Instead, the "place of effective management" test looked at whether there was some overarching scheme of management of the trust that could be divorced from the day-to-day management of the trustee.

In applying this approach to the "place of effective management" test, Judge Morgan found in Haworth that the trust was effectively managed in the UK at the relevant times. Critical to her reasoning were the facts that:

• the scheme was "devised, decided upon, facilitated, orchestrated and superintended in the UK by the settlors and the UK advisers...on an on-going basis throughout the relevant period.";
• it was always intended that the Mauritian trustees would be in office for only a brief period of time;
• the Mauritian trustees were appointed on the expectation that they would implement the scheme; and
• the decision to initiate, orchestrate, superintend and refine the scheme on an on-going basis taken by the UK settlors and their UK advisers represented effective "top-level management" of the trust.

But was Smallwood right?

In paragraph 360 of her judgment, Judge Morgan questioned the reasoning of Hughes LJ (and the Commissioners) in Smallwood. In considering the meaning of "place of effective management", she seemed to prefer the approach taken in Wood v Holden, which would mean determining "place of effective management" on the basis of the central management and control of the corporate trustee in office at the time. However, she accepted that she was bound by the Court of Appeal’s decision in Smallwood and so conceded that her own views were not strictly relevant.

What does this mean?

For the purposes of the tie-breaker test in double tax treaties, the "place of effective management" of a trust is where the "top-level management" of the trust (not the trustees) takes place. This is on the basis that trustees should be treated as "a deemed trustee body" and not individual trustees. The "place of effective management" test is not the same as a "central management and control" test.

The facts of both Smallwood and Haworth may well be of largely historic significance since it seems unlikely that many well-advised taxpayers would still consider implementing "round the world" schemes. However, taken together with the Development Securities line of cases (as to which see our article), this case demonstrates how important it is to understand the potential effect that third parties can have on the residence position of trusts and companies alike.

Finally, Judge Morgan’s comments on whether or not the Court of Appeal majority decision in Smallwood was correct, taken together with Patten LJ’s dissenting judgment, suggest that the courts may well one day need to revisit the meaning of "place of effective management" in the context of corporate trustees.