Taxing digital presence: do we need a digital services permanent establishment in the OECD model?

Dragoș-Alexandru Iordache
Dragoș-Alexandru Iordache

As part of BEPS Action 1 concerning the digital economy, the OECD identified two major trouble areas: (i) deciding which state should be able to tax, and (ii) developing allocation mechanisms[i]. Although Action 1 has not led to amendments to the Model so far, ‘the digital economy is the entire economy’, and limited solutions towards addressing the two problems have been introduced via Action 7[ii]. In what follows, I outline the material elements of these changes, followed by the issues which are yet to be adequately solved. I then assess what other solutions have been implemented or touted at unilateral, EU, or OECD level.

Action 7 resulted in the 2017 amendments[iii] to Art. 5 covering the definition of a permanent establishment (PE) and to the associated Commentary. All exceptions in Art. 5(4) have been made subject to the ‘preparatory and auxiliary’ caveat. The new anti-fragmentation rule in Art. 5(4.1) requires that the activities of an entity in a country must be aggregated irrespective of whether they constitute a PE or not. The rule also covers ‘closely related enterprises’ (Art. 5(6)) as defined in Art. 5(8). In respect of agency, Art. 5(4) now incorporates the negotiation-based agency test[iv]. Paragraph [62] of the Commentary on Art. 5 provides that a very large warehouse from which online orders are fulfilled likely constitutes a PE. Paragraph [144] also introduces the option of a services PE.

Many problems related to the digital economy remain unaddressed. The first is that of e-sales. Paragraph [123] makes clear that a website will not be a PE. A server, analogised to a vending machine, can be a PE. Nevertheless, what is likely to happen is that the server will be placed in a low tax jurisdiction and sales will be made in the market country through the website only. Secondly, although the definition of a dependent agent in Art. 5(4) was widened, it still depends on physical presence. An electronic intermediary will not be caught. The position is oddly different from that of technical services, where the source state can withhold tax without physical presence being required. Thirdly, there is the allocation difficulty under Art.7(2). One may end up with a PE to which very few real or intra-enterprise dealings can be attributed. An indicative part-solution may be found in the UN Model version of Art. 5 (rejected by the OECD[v]), which contains a force of attraction rule. The PE state can tax not only profits attributable to the PE but also sales associated to the PE. Such a position would be helpful for dealing with contract-splitting, but it would not deal with cases where under current rules there is no PE at all.

The desire to expand the notion of PE in the digital context originates in the perceived need to give taxing rights and allocate a proportionate amount of profits to the market country. This idea is not universally shared. The US position has long been against taxing in the market country – they state that value created in the US is fully taxed there (through corporation tax and the global intangible tax regime). Yet the market country approach has slowly been gaining more traction and has been recognised in the 2015 OECD Report on Action 1[vi]. The argument is that the exploitation of data provided by users captures value in itself[vii] or transfers financial means[viii]. Nevertheless, the lack of a solution at that level (the 2018 OECD report[ix] is inconclusive) has opened the way to unilateral action. Take the UK example. Two taxes have been introduced: the diverted profits tax[x] (to defeat general avoidance of the PE status not otherwise covered) and a digital services tax[xi] (DST), which is a stop-gap to capture some of the income of digital sector MNEs which would have fallen outside the PE definition altogether for the reasons discussed above. These forms of unilateral action risk threatening the integrity of the tax treaty system because many are de facto taxes on income.

In addition to unilateral action, there have been attempts to act at EU level. The Commission proposed two directives concerning a short-term DST-like stop-gap[xii] and a more long-term solution going along the lines of the profit-split method of allocation[xiii]. But the DST has the same definitional problem – it is unconvincing to argue, as the Commission does, that it is a genuine indirect tax. The support for an interim tax is motivated more by a desire to avoid unilateral action than by coherence[xiv]. On the allocation front, the profit-split method (as proposed in Action 10[xv]) has the advantage of being useful when both parties make unique and valuable contributions. It would also be consistent with the proposed revision to the transfer pricing (TP) guidelines so that legal ownership of intangibles is no longer enough to create a right of return. The disadvantage is the difficulty posed by applying the profit-split method, both methodologically and in the absence of sufficient data.

One way forward is via the TP rules. If the changes brought about by Action 6 on anti-abuse truly worked, there would be no need for Action 1 at all[xvi]. Given that they are unlikely to have a significant effect, there is still a need to act through re-allocation mechanisms. There is an apparent conflict between Actions 8[xvii]-10, which are predicated on the assumption that income should be allocated primarily where value is created and the market state approach at the heart of Action 1, but the conceptual struggle between the two is less of a problem in practice. More important is the ditching of the blanket stop-gap taxes. Rather than going ahead with the digital PE or the equalisation tax hinted at in Action 1, a better view is revising the TP guidelines based on Actions 8-10.

If a profit-split is to be applied (regardless of the conceptual footing behind it), a way needs to be found to mitigate the difficulty in application. The 2019 OECD Programme of Work[xviii] shows the structure of the modified residual profit-split analysis, which is as follows: (i) determination of the total profit, (ii) removal of routine profit, (iii) determination of non-routine profit, (iv) splitting of the non-routine profit between the jurisdictions in question. A fractional apportionment method is also contemplated, but it would demand an unlikely agreement between states on a formula.

The 2020 OECD Statement[xix] may clarify this to some extent by providing a classification of what can be allocated to the market jurisdiction. It features: (i) a share of residual profits at MNE level for active and sustained participation in the economy (for automated digital services, consumer-facing business), (ii) a fixed remuneration based on arm’s length pricing for a baseline set of activities consisting of distribution and marketing, (iii) any additional profit where in-country functions exceed the baseline compensated under (ii). This is not to deny that computation difficulties persist, and the question is often asked whether there are any truly ‘residual’ profits in the first place.

Therefore, while it cannot be said that the Action 7 amendments adequately resolved the challenges arising out of the digital economy, it was perhaps wise to confine them to matters on which agreement could be reached. A Commentary solution akin to the services PE might have done little more than sow additional uncertainty.

[i] OECD (2015), ‘Addressing the Tax Challenges of the Digital Economy’, available at https://www.oecd-ilibrary.org/docserver/9789264241046-en.pdf?expires=1596714830&id=id&accname=guest&checksum=4456BFCF92A15479E8AA42BA57E438FD
[ii] OECD (2015), ‘Preventing the Artificial Avoidance of Permanent Establishment Status’, available at https://www.oecd-ilibrary.org/docserver/9789264241220-en.pdf?expires=1596714984&id=id&accname=guest&checksum=DDEF8CF426662995E83B5BABE68C9A9A
[iii] OECD (2017), ‘2017 Update to the OECD Model Tax Convention’, available at http://www.oecd.org/ctp/treaties/2017-update-model-tax-convention.pdf
[iv] This test was first outlined by the Tax Court of Canada in Knights of Columbus v The Queen [2008] 10 ITLR 827 (TC).
[v] See Peter Harris, International Commercial Tax (2nd ed, Cambridge University Press 2020), p. 179
[vi] OECD (2015), “Addressing the Tax Challenges of the Digital Economy, Action 1 – 2015 Final Report”, available at http://www.oecd-ilibrary.org/taxation/addressing-the-tax-challenges-of-the-digital-economy-action-1-2015-final-report_9789264241046-en
[vii] Werner Haslehner, ‘Taxing where value is created in a post-BEPS (digitalized) world?’ (Kluwer International Tax Blog, 30 May 2018), available at http://kluwertaxblog.com/2018/05/30/taxing-value-created-post-beps-digitalized-world/
[viii] Wolfgang Schoen, ‘One Answer to Why and How to Tax the Digitalized Economy’ (2019) Working Paper of the Max Planck Institute for Tax Law and Public Finance 2019-10, available at https://ssrn.com/abstract=3409783
[ix] OECD (2018), ‘Tax Challenges Arising from Digitalisation –Interim Report 2018’ (OECD Publishing, Paris), available at http://dx.doi.org/10.1787/9789264293083-en
[x] Finance Act 2015, s. 71. See the guidance issued by HMRC (2018), available at https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/768204/Diverted_Profits_Tax_-_Guidance__December_2018_.pdf
[xi] Finance Act 2020, s. 39. HMRC (2020) have published an interim manual, available at https://www.gov.uk/hmrc-internal-manuals/digital-services-tax
[xii] European Commission, Proposal for a Council Directive on the common system of a digital services tax on revenues resulting from the provision of certain digital services (2018/0073), available at https://ec.europa.eu/taxation_customs/sites/taxation/files/proposal_common_system_digital_services_tax_21032018_en.pdf
[xiii] European Commission, Proposal for a Council Directive laying down rules relating to the corporate taxation of a significant digital presence (2018/072), available at https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52018PC0147
[xiv] Murray Clayson, ‘Tax reform in the digital economy: recent OECD and EC activity’, Tax Journal, No 1393, pp. 9-13
[xv] OECD (2018), ‘Revised Guidance on the Application of the Transactional Profit-split Method – BEPS Action 10’, available at https://www.oecd.org/tax/beps/revised-guidance-on-the-application-of-the-transactional-profit-split-method-beps-action-10.pdf
[xvi] Martin Jimenez, ‘BEPS, the Digital(ized) Economy and the Taxation of Services and Royalties’, UCA Tax Working Papers 2018/1, available at https://ssrn.com/abstract=3215323
[xvii] For the current state of play, see OECD (2020), ‘Inclusive Framework on BEPS: Progress Report July 2019-July 2020’, pp. 20-21, available at https://www.oecd.org/tax/beps/oecd-g20-inclusive-framework-on-beps-progress-report-july-2019-july-2020.pdf
[xviii] OECD (2019), ‘Programme of Work to Develop a Consensus Solution to the Tax Challenges Arising from the Digitalisation of the Economy’, available at https://www.oecd.org/tax/beps/programme-of-work-to-develop-a-consensus-solution-to-the-tax-challenges-arising-from-the-digitalisation-of-the-economy.pdf
[xix] OECD (2020) ‘Statement by the OECD/G20 Inclusive Framework on BEPS on the Two-Pillar Approach to Address the Tax Challenges Arising from the Digitalisation of the Economy’, available at https://www.oecd.org/tax/beps/statement-by-the-oecd-g20-inclusive-framework-on-beps-january-2020.pdf

Dragoș-Alexandru Iordache

* An earlier draft of this article was submitted as part of the Cambridge LLM examination in International Commercial Tax (2020).

Related posts