Welcome to the concept of beneficial ownership ‐ one of the most nebulous concepts in intellectual property tax law ‐ which surely must echo the voice of Nietzsche's in Why I Write Such Excellent Books: "When Doctor Heinrich von Stein once honestly complained that he understood not one word of my Zarathustra, I told him that was quite in order: to have understood, that is to say experienced, six sentences of that book would raise one a higher level of mortals than "modern" man could attain to."
Although the concept of beneficial ownership casts a shadow over international royalty, dividend and rental payments, this article is concerned only in its application to royalties.
Payment of royalties by a local licensee to a foreign licensor triggers withholdings tax. In the case of South Africa, withholdings tax is set by our Income Tax Act at 12%. However, Double Tax Agreements ("DTA") often reduce this tax … in many instances to 0%. For example, Article 12.1 of the South African / Mauritian DTA provides:
"12.1. Royalties arising in a Contracting State and paid to a resident of the other Contracting State shall be taxable only in that other State if such resident is the beneficial owner of the royalties."
As such, where a South African licensee pays royalties to a Mauritian licensor, provided that the Mauritian licensor is the "beneficial owner" of the royalties, the royalties are taxable only in Mauritius ‐ South African withholdings tax is reduced to 0%.
Presuming that I wish to conclude a licence between an Australian licensor and a South African licensee: the South African / Australian DTA reduces withholdings tax payable to SARS from 12% to 10%. Assuming that the Australian / Mauritian DTA mirrors the South African / Mauritian DTA and zeros withholdings taxes, would it not be preferable instead of concluding a licence directly between the Australian licensor and the South African licensee rather to conclude a parent licence between the Australian licensor and the Mauritian licensee and a sub‐licence between the Mauritian licensee and the South African sub‐licensee? By doing so, royalty payments could flow from the South African sub‐licensee via the Mauritian licensee to the Australian licensor without triggering any withholdings taxes whatsoever.
In an attempt to counter such tax planning, the OECD introduced the concept of "beneficial ownership" into article 12 of the model DTA. Applying this concept, the parties to my hypothetical example would only circumvent withholdings taxes if the Mauritian licensee was found to be the beneficial owner of the royalties paid by the South African sub‐licensee. If the tax authorities succeed in arguing that the beneficial owner of the royalties paid by the South African sub‐licensee is in fact the Australian licensor, the South African / Australian DTA would apply and the South African sub‐licensee would once again find itself liable to SARS for the 10% withholdings tax.
Having introduced this concept in 1995, one could reasonably expect some degree of clarity today. But, the term has yet to be defined in the OECD Model Double Tax Treaty, the OECD guideline or our law, so what does it mean?
One large auditing firm's interpretation of beneficial ownership is that as long as the intermediary makes a market‐related margin for its services, it is regarded as the beneficial ownership of all payments made to it. If this is correct, a margin of between 5% and 10% retained by the Mauritian licensee in our example should successfully circumvent withholdings tax. But, I have serious misgivings regarding this interpretation ‐ it is doubtful that the test for beneficial ownership is the same as that for transfer pricing.
Turning to other countries for assistance, where analysis has focussed largely on the "beneficial ownership" of dividends:
Apart from concluding that:
cannot be regarded as the beneficial owner of the asset, it is difficult to draw a common thread between the above interpretations ‐ each country appears to be tugging in a different direction. For this reason, Charl Du Toit proposes a hybrid test. According to Du Toit's catholic interpretation:
"The beneficial owner is the person whose ownership attributes outweigh that of any other person."
For ownership attributes, Du Toit gives no preference to any one country's interpretation. But have our Courts not favoured the restricted English concept of legal ownership over the practical US concept of economic ownership?
Now that I have managed to make a readable phrase less clear, let's try and break into the light by considering a few examples discussed by Du Toit:
Example 1: A(US) licenses intellectual property ("IP") to B(ZA) for a period of 5 years with fixed annual royalty payments. One year later, B wishes to discontinue its business, but is bound to continue paying fixes annual royalties to A for a further 4 years (without the right to delegate its obligations in terms of the licence). B sells the assets of the business to C(UK) and grants C a sub‐licence in consideration for the same royalties that B is obliged to pay A.
Ans.: According to Du Toit, B remains the beneficial owner of royalties paid by C as B is not legally obliged to pay the royalties received from C on to A. B just happens to have a matching liability in favour of A.
But what if C gives B promissory notes in respect of the future royalty payments, which notes B in turn endorses to A? Have the tables not turned?
Example 2: A licenses IP to B, who sub‐licences the IP to C. B is contractually bound to pass all royalties received from C on to A, who in turn pays B a commission of 20% of royalties received. (A highly improbably scenario)
Ans.: A is clearly the beneficial owner of royalties paid by C.
But what if B is required to pay royalties to A irrespective of whether C performs? Is A still the beneficial owner of royalties paid by C?
And what if, irrespective of the contractual obligations, B on‐pays 90% of royalties received from C to A? It is anyone’s guess … depending on the terms of the licences, of course.
In the absence of judicial interpretation, it is doubtful whether we will achieve a greater degree of clarity regarding beneficial ownership of royalties. Until SARS starts challenging a few licensing arrangements, we may never be able to delineate the boundaries of this doctrine.
You may recall that the original draft of our R&D tax incentive provision (s11D) included a reference to beneficial ownership. It was specifically due to this high degree of uncertainty surrounding this phrase that it was quickly removed from the section … broadening the scope of that section considerably.