The question is almost always asked in the wrong form. A principal arrives having compared corporate tax rates across four jurisdictions and wants to know which is best. The rate is a fact about a jurisdiction; it is not a fact about the structure, and on its own it settles nothing. Two groups with identical rates in front of them will properly reach different conclusions, because the rate is one input among several and rarely the binding one.
What settles the question is a short set of enquiries about the group as it actually operates, asked in an order that matters. Answered honestly, they usually reduce the field to two jurisdictions. The final choice between those two is then a judgement about which set of constraints the group is better placed to carry over the life of the structure.
The six questions, in the order they bind
Where do the assets sit, and what are they? Shares in operating subsidiaries, real estate, intellectual property and portfolio investments each behave differently on disposal and on death, and each is treated differently by the jurisdiction in which it is located rather than the one in which it is held. Where are decisions actually taken, and by whom? A holding company managed from a country other than the one it is registered in may be treated as resident where its directors meet, whatever its certificate of incorporation says.
Which treaties does the group need, and can it satisfy them? A treaty network is only as useful as the substance behind the claim. What will the banks accept? A jurisdiction that is technically efficient and practically unbankable for the group’s sector is not a solution. What must be disclosed, and to whom? Registers differ in what they publish and who may search them. And who inherits? A structure that works for twenty years and fails at succession has been solved for the wrong horizon.
A jurisdiction is chosen for the constraints the group can live with, not for the rate printed at the top of the comparison.
Why the rate comes last
The headline rate is the most visible figure and the least decisive one. It applies to profits the holding company itself earns, which for a pure holding vehicle are often limited to dividends and gains, both of which are frequently exempted or relieved under participation regimes that vary far more between jurisdictions than the headline rates do. The effective position therefore turns on the exemptions, the conditions attached to them and the group’s ability to meet those conditions year after year.
There is a second reason for its position in the order. Rates change, and a structure formed for a rate is a structure formed for something the group does not control. Ownership, control, disclosure and succession change more slowly and are within the group’s power to arrange. Building on the stable elements and accepting the rate as a variable produces a structure that survives a change of policy; building on the rate produces one that has to be reconsidered whenever the policy moves.
What the answers usually leave
Worked through properly, the six questions rarely leave a single answer, and a process that produces one should be treated with suspicion. They leave two credible candidates, each with an identifiable weakness: one with a stronger treaty position and heavier substance requirements, the other easier to administer and more exposed on disclosure or on recognition of the vehicle by a particular counterparty.
The decision between them is properly made by the principal, not by the adviser, once the trade-off has been set out in writing. What the firm contributes is the discipline of asking the six questions before the incorporation rather than after the first bank review, and of recording the reasoning at the time. A structure whose rationale exists only in the memory of whoever formed it becomes very difficult to defend when the person asking is an auditor, a registrar or a successor.
This note is general in application and does not constitute legal, tax or regulatory advice. It describes practice observed across institutions and should not be relied upon in relation to any particular structure.