[Guest Post] Tax Court of South Africa wrestles with “arm’s length” royalties rate for IP licenses

One of the sure things in the world of royalties and royalty rate calculations is taxes. Katfriend Thato Moloto (Trademarkia) writes on the methods used in calculating royalty rates for IP licences to enable the determination of appropriate taxes. Here’s what Thato writes:

Tax Court of South Africa wrestles with “arm’s length” royalties rate for IP licenses
by Thato Moloto

On 14 February 2024, the Tax Court of South Africa handed down a judgement in the seminal transfer pricing case of ABD Limited v The Commissioner of the South African Revenue Service (IT 14302). This case revolved around determining the market price at which a parent company would have charged its foreign subsidiaries for the use of intellectual property in market-based arm’s length negotiation. The court, in its deliberation, scrutinized various methods and concluded that a valuation using the Comparable Uncontrolled Price Method (CUP) was apt for this specific scenario.

Let's delve deeper into the proceedings.

Background

ABD Limited (ABD), a South African telecommunications company with global subsidiaries, referred to as "Opcos”, granted licenses for use its intellectual property to these Opcos. These Opcos have local shareholders in each country, with ABD being a significant shareholder. In return for these licenses, the parent company received royalties, calculated at 1% of the Opcos' profits, a figure it settled on based on recommendations from a consultancy hired by ABD.

If AI did the work, would we pay taxes...

Initially, the Commissioner of the South African Revenue Service (the Commissioner), accepted ABD's calculations following the advice of its own expert. However, after a transfer pricing audit, the Commissioner issued an additional assessment based on the evaluation of a second expert, determining that the 1% royalty rate was not a fair market price. Instead, the Commissioner set the royalty rate at 3%, resulting in a much greater liability for ABD.

Choice of method
The Court had to determine which royalty rate most closely indicated that the fair market price was charged for the royalties. A necessary corollary of this enquiry being an evaluation of the methods used to determine the value of the intellectual property as applied by the different experts.

ABD and the Commissioner both subscribed to and accepted the OECD Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations (OECD Guidelines) as guidance for calculating the fair market price of the intellectual property.

The OECD Guidelines provide five common methods that are acceptable to nearly all tax authorities. What was at issue between the parties in the present case was whether the correct method in this matter was either the Transactional Profit Split Method (TPS method) or the Comparable Uncontrolled Price Method (CUP).

Transactional Profit Split Method (TPS method)
The Commissioner relied on the TPS method for his additional assessment. This method is understood to involve examining the net operating profits realised from the transactions between ABD and its Opcos and then splitting those profits between these entities on a basis that approximates the division of profits that would have been agreed to during market-based negotiations.

The TPS method-related calculation involved firstly investigating how much more a buyer of ABD branded products was willing to pay versus unbranded products. This was termed the willingness to pay techniques (WTP). It began with surveying some nearly 4000 direct consumers and business customers in six of the fourteen markets ABD was active to determine the ‘value of the brand’.

Based on this, what follows is an accounting exercise which aims to isolate the branded value from the non-branded value, ascertain the revenue from the sale of ABD branded products, then approximating a reasonable share of the profit to be apportioned to each party.

There were multiple criticisms levelled against this expansive evaluation exercise. The most fatal criticism was that the expert premised his valuation on the legal error that the trade marks and other intellectual property licenced to the Opcos were transferred with the goodwill in ABD’s business. The transfer of goodwill was in fact explicitly excluded from all but one of the branding licensing agreements.

Given the value of goodwill to businesses, this mistake would naturally lead to an inflated projected royalty rate. The assumption related to this mistake also contaminated the rest of the process from the onset, including in the framing of the survey questions issued to ABD’s customers.

Comparable uncontrolled price method (CUP method)
ABD relied of the CUP method. This method involved comparing the royalty rate charged by ABD to a comparable royalty rate charged previously in similar circumstances. The comparison seeks to replicate how market forces have acted, with contingencies for the differences between the past and present transaction built into the calculation.

The comparable transaction identified by ABD involved the sale of one of its subsidiaries based in Cyprus to a third party (the Cyprus transaction). The now independently owned entity in Cyprus then concluded a brand licence agreement with ABD on substantively the same terms (and royalty rate) as ABD’s Opcos. ABD argued that this brand licence agreement was a suitable CUP candidate as the parties were at arm’s length when the agreement was concluded.

Overall, although there were some valid criticisms levelled against some of assumptions and techniques applied when this method was used, the Court found the Cyprus transaction’s CUP the most persuasive. The Court held that the result most resembles what would been achieved in a marked-based arms-length negotiation. This contrasts with the TPS method which, whether completed using steps used by the Commissioner’s expert or with another technique, yielded widely different outcomes,

Comment
The Tax Court of South Africa is specialist statutory court that is restricted to the issue under review and does not create legal precedent. Nevertheless, although this judgment is not binding, its reasoning may be persuasive and even instructional to IP valuation practitioners given the depth of the Court’s expertise and its sound analysis.

Similarly, inasmuch as the CUP method was suitable where an appropriate past transaction was available for the comparison, the circumstances of other matters may not be as favourable.

Furthermore, as is evident from the vague language used by the experts in describing a bundle of intellectual property rights ( “brand”, “brand licence”, “brand value” and the like), it is easy in trade mark valuations to conflate the value of the potential market with the clearly delineated and strictly licensable intellectual property rights.

Considering this and in circumstances such as these where intellectual property is the primary subject matter of the valuation or the dispute, the parties would benefit greatly from consulting an intellectual property practitioner to avoid basic errors in law that other type of experts may inadvertently fall into.