Scale without Mass: Permanent Establishments in the Digital Economy
Andrés Felipe Ramírez Ocampo
Advogado e Economista formado na Universidad de los Andes, Colômbia. Mestrando em Direito Tributário Internacional e Desenvolvimento pelo IBDT. Advogado em São Paulo. E-mail: email@example.com
Resumo. Este paper contém uma visão geral das diretrizes atualmente existentes sobre os estabelecimentos permanentes, conforme estabelecido no Comentário à Convenção Modelo da OCDE sobre Renda e sobre Capital, e sua aplicabilidade à economia digital. Utilizo estas informações para comparar as instruções da OCDE com a proposta de diretiva do conselho europeu que estabelece regras relativas à tributação das empresas de uma presença digital significativa (2018/0072 (CNS)) emitida em março de 2018. Finalizo com conclusões sobre a aplicabilidade e os efeitos dessas considerações para os países em desenvolvimento.
Palavras-chave: estabelecimento permanente, comentário da OCDE, economia digital, diretiva de presença digital, países em desenvolvimento
Abstract. This paper contains an overview of the currently existing guidelines on permanent establishments as set forth in the Commentary to the OECD Model Tax Convention on Income and on Capital and their applicability to the digital economy. These insights are used to compare current OECD standards with the Proposal for a council directive laying down rules relating to the corporate taxation of a significant digital presence (2018/0072 (CNS)) issued in March 2018, for discussion in the context of the European Union. The conclusion analyses the applicability and effects of these considerations for developing countries.
Keywords: permanent establishment, OECD commentary, digital economy, digital presence directive, developing countries
Business models of the digital economy showcase one of the characteristics of global markets: firms can grow in trans-continental scale without an increase in mass – “Mass” referring to a firms’ physical presence in the location of the user or the customer’s market. This phenomenon is called “scale without mass”.
Scale without mass is problematic because the current definition of nexus under Article 5 of the 2017 OECD Model Tax Convention and the attribution of profits under Articles 7 and 9 are both based on physical characteristics used to align profits (and taxation) with value creation. If highly digitalized business models manage to outsource or eliminate their physical characteristics in any given jurisdiction where they nevertheless perform a significant economic activity, then there are no tools left for tax administrations of that jurisdiction to apply the Permanent Establishment concept over that same economic activity.
The aim of this paper is to explore the current state of the commentary and supplementary guidelines issued by the OECD to pinpoint the extent of the scale without mass problem, particularly as regards certain highly digitalized business models as described in the OECD 2018 Interim Report: “Tax Challenges Arising from Digitalization”, and future proposals to amend them.
The extent of the physical nexus requirement
The current version of the Commentary to the OECD Model Tax Convention on Income and on Capital is one of the most relevant and comprehensive texts issued by a multilateral organization in world history. It is the result of an ongoing process more than fifty years in the making, with dozens of nations rising to the daunting task of ascribing meaning in practice and in detail to vague and general notions which could otherwise be rendered inconsequential due to hermeneutical ambiguity.
This is exemplarily true as far as the Commentary to Article 5 is concerned, particularly when one reviews the current mechanics of the Permanent Establishment concept and their potential to tackle base erosion and profit shifting, even when compared to the 2014 version of the same text (this is a direct result of the BEPS initiative).
Pursuant to the 2017 version of the Commentary, as outlined on the Additional Guidance on the Attribution of Profits to Permanent Establishments issued by the OECD in March 2018, the current application of the Permanent Establishment concept follows a holistic approach to economic activity that aligns source taxation to the factors leading to the actual realization of profits, regardless of previously existing limitations induced by the wording of the 2014 OECD Model Tax Convention and its predecessors.
An example of this is the current scope of the definition of “preparatory or auxiliary” activities, which has been considerably reduced. In the words of the OECD, “activities previously considered to be merely preparatory or auxiliary in nature may nowadays correspond to core business activities” 1.
The reach of Paragraph 4 exceptions was severely limited too, to accommodate a broader Permanent Establishment concept. Additionally, the threshold to become a dependent agent was materially lowered.
Therefore, the range of the current Permanent Establishment definition is better suited to address the challenges of the digital economy than its predecessors, particularly because of the relatively lower degree of physical nexus required for economic activity to qualify as a Permanent Establishment and to have profits attributed to it.
The above is better illustrated by two cases described in the Commentary and subsequent guidelines regarding the attribution of profits to Permanent Establishments.
Case 1: “[…] an enterprise of State R maintains in State S a very large warehouse in which a significant number of employees work for the main purpose of storing and delivering goods owned by the enterprise that the enterprise sells online to customers in State S, paragraph 4 will not apply to that warehouse since the storage and delivery activities that are performed through that warehouse, which represents an important asset and requires a number of employees, constitute an essential part of the enterprise’s sale/distribution business and does not have, therefore, a preparatory or auxiliary character.” 2
The first case (copied from Paragraph 62 of the Commentary on Article 5) shows that despite the fact that the physical nexus in the source state is the maintenance of a warehouse for distribution purposes, such an activity cannot be deemed auxiliary or preparatory in nature (as Paragraph 4 would suggest) but essential and significant to the actual realization of profits – this, exclusively because of the characteristics of the business model. This is surprising given that the enterprise of State R does not physically perform the sale of any products to customers in Country S.
This is also true regarding the attribution of profits to such a Permanent Establishment, which would ordinarily be based on the activities of personnel stationed in the warehouse (a functional analysis) which would determine the extent of the risks assumed in Country S operations of the enterprise. Pursuant to the Additional Guidance on the Attribution of Profits to Permanent Establishments, the work performed by the employees working on the warehouse in Country S constitutes sufficient people functions to attribute the profits generated by that warehouse to the Permanent Establishment and, therefore, to tax the profits arising in that Country therein.
Note that the case above is in clear contradiction with Paragraph 4(b), which expressly excludes “the maintenance of a stock of goods or merchandise belonging to the enterprise solely for the purpose of storage, display or delivery” 3 from the definition of Permanent Establishment. It is because of the nature of the business model, which relies on delivery of goods to customers in Country S without these goods flowing through wholesale and retail as it used to be the case before e-commerce was introduced, that activities previously deemed preparatory or auxiliary in nature become essential and significant in the current version of the Commentary – the “online” nature of the business makes it more profitable but simultaneously makes the scarce physical nexus overtly relevant for Permanent Establishment purposes.
Extrapolating to other business models that require a certain degree of physical presence (the “Value Chain” digital economy 4), having this broadened Permanent Establishment concept means that the current state of the Commentary is sufficient to align economic activity with value creation (scale without mass in the retail business is not as extreme as in other industries of the digital economy, or at least the amount of mass present still sufficiently generates a physical nexus). As proven by the market share and current valuation of companies such as Alibaba and Amazon, online retail is one of the most significant sources of revenue of the digital economy which is therefore still captured by the current version of the Commentary.
The second example shows the extent towards which the guidelines have applied the 2017 Commentary, this time regarding income form advertising in online platforms.
Case 2: “SiteCo, a company resident in Country R, owns the rights in a website. SellCo, an associated company resident in Country S, performs marketing activities on behalf of SiteCo in Country S under a services agreement with SiteCo that provides for the fee payable to SellCo to be equal to a percentage of the sales revenue received by SiteCo from sales of advertising space to customers in Country S. The effect of the arrangement is that SellCo habitually plays the principal role leading to the routine conclusion of sales by SiteCo in Country R to customers in Country S without material modification of the terms and conditions on which the customers offer to purchase the advertising space. […]” 5
Case 2 was copied from the Additional Guidance on the Attribution of Profits to Permanent Establishments referenced above. In its analysis, the OECD concludes that, pursuant to the current version of the Commentary on Paragraph 5 of Article 5, even though SellCo does not have the power to habitually conclude contracts, its activities are sufficiently beneficial so that they lead to the conclusion of advertising agreements between SiteCo and its clients in State S.
In this regard, the current version of the Commentary 6 ignores legal powers and attributions and focuses on the relevant facts and circumstances that lead to the conclusion of contracts. If the presence of an agent in Country S (inconsequential as it may be compared to the overall business of the website) is sufficient to cause the website to profit in that Country, then a sufficient physical nexus is deemed to exist.
Note that SiteCo has no physical nexus with State S other than its contract with SellCo. The question arises then as to the extent of the activities that need to be currently performed by an Agent of a website that depends on advertising for revenue (the “Value Network” business model 7) for that Agent to be considered a Dependent Agent Permanent Establishment. Could it be stated that any website engaged in global commerce gains sufficient market visibility through word-of-mouth advertising, without the need to hire local advertising, to be successful? Hardly. On-line advertising is the fastest-growing sector in the industry, with on-line advertisers billing over USD 107 billion in the United States in 2018 8. If trends continue at their current pace, having agents performing advertisement services in source-state jurisdictions may soon become inevitable.
The attribution of profits in Case 2 is similarly comprehensive as in Case 1, based on the functional analysis of SellCo’s activities in Country S. In the end, Country S will be able to capture a significant portion of the revenue accrued by SiteCo in that country, exclusively because of SiteCo’s interaction with SellCo for local advertising. It should be noted that this whole interaction remains, in itself, entirely digital.
The two examples above show that the Commentary and supplementary guidelines are currently equipped to cover most situations involving the digital economy (including hybrids that involve a sufficient degree of physical nexus), as long as the physical nexus activities cannot be deemed auxiliary or preparatory 9.
The “scale without mass” problem is therefore only really relevant regarding business models that have completely been digitalized, such as Value Chains that deal exclusively in intangibles and Value Shops 10 that are limited to the provision of digital services. Even then, most scaling requires some degree of physical presence in any given jurisdiction to comply with consumer protection legislation and other forms of local substance requirements – including advertising – and therefore most of the revenue involved can still be captured through the latest amendments.
However, one must accept that there currently are some business models with no physical nexus whatsoever to a given jurisdiction, which may be replicated and extended in the near future. These business models would completely escape taxation based on the Permanent Establishment model in its current state, and therefore their existence requires adjustments that may involve a radical shift from the physical nexus criterion: The Significant Digital Presence Permanent Establishment.
Facing completely digital business models
The future may very well witness the complete digitalization or automation of business models. If this is true, even extremely flexible physical nexus rules (as the ones currently in place) will not be sufficient to adapt the Permanent Establishment to effectively deal with the “scale without mass” phenomenon. Consequently, some conceptual amendments will have to be implemented at the core of the current definition for the Permanent Establishment to survive.
As the OECD points out in its 2018 Interim Report, certain countries are already experimenting with adjusted versions of the Permanent Establishment definition available in local legislation to rely on economic activity rather than physical activity in aligning taxation and value creation. In this regard, the Slovak Republic, Israel, and Italy have currently implemented Significant Economic Presence tests to alter their Permanent Establishment definitions, and the European Commission is presently analyzing COM(2018) 0147, the “proposal for a Council Directive laying down rules relating to the corporate taxation of a significant digital presence”, in line with foreign regulations.
The common characteristics of all of the initiatives described above are that they are not intended to conflict with treaty provisions, and in the cases of the Slovak Republic, Israel and India, are only meant to be applied in the absence of a Double Tax Treaty.
In the case of the European Union, however, the proposal explicitly states that member countries should coordinate their efforts to achieve similar modifications to the OECD Model Tax Convention: “The rules in this proposal should be integrated into Member States corporate income tax systems and the Commission’s CCCTB proposal and should ultimately be mirrored by corresponding changes in the OECD Model Tax Convention (OECD MTC) at [sic] international level.” 11
Given the influence of European countries in the policy creation process and the corresponding updates to the Model and the Commentary, this is extremely relevant signaling of what is to be expected of the 2020 version and the new Minimum Standard.
The core of the European proposal is based on a three-pronged approach to business models in the digital economy, referred to as the “user-based” approach: revenue, number of users and number of contracts for digital services. Due to the heterogeneity of the ecosystem of the internet, one would expect various combinations of these three factors as the defining characteristics outlining “scale without mass” global firms.
Pursuant to the model directive,
“The concept of significant digital presence is intended to establish a taxable nexus in a jurisdiction. Therefore, it should be regarded as an addition to the existing permanent establishment concept. The proposed rules for establishing a taxable nexus of a digital business in a Member State are based on revenues from supplying digital services, the number of users of digital services or the number of contracts for a digital service. These criteria are proxies for determining the ‘digital footprint’ of a business in a jurisdiction based on certain indicators of economic activity. […]
For the three user-based criteria mentioned above […] different applicable thresholds are set. There is a significant digital presence in a Member State if one or more of the following criteria are met: if the revenues from providing digital services to users in a jurisdiction exceed EUR 7 000 000 in a tax period, if the number of users of a digital service in a Member State exceeds 100 000 in a tax period or if the number of business contracts for digital services exceeds 3 000.” 12
The above is an eminently practical approach that effectively creates a chasm between the previous mechanics of the Permanent Establishment concept and the new ones. The new definition dangerously approaches the concept of consumption-based taxes, without considering the elements of the business model that effectively characterize a given activity as a Permanent Establishment based on tangible economic activity – the concept of auxiliary or preparatory activities becomes completely useless 13 under this approach.
Also, the “fixed place of business” will have to acquire an entirely new meaning, no longer attached to a geographical reference in the traditional sense of a “place”, but to an electronic marketplace positioned in a specific location in the physical world because of the address of end users.
The definition of the adjective “fixed” would also need to change, no longer attached to a degree of permanence in a single place 14, but to the volume of user interaction with the electronic address of the enterprise (here, one cannot help but question the difference between user 99 999 and user 100 001, which is the difference between having or not having a Permanent Establishment in a given jurisdiction).
Similarly, the phrase “business of the enterprise is wholly or partly carried out” loses meaning, being circumscribed to thresholds, and morphing into “business […] consists wholly or partly of the supply of digital services through a digital interface” 15.
In this context, we are forced to go back to the original idea behind Permanent Establishments to be able to identify one in the molds of the new European directive. Since Permanent Establishments are designed as ways to apply corporate income taxation in the source State for profits that would not be taxed in that State otherwise, then the new regulations uphold this ideal – if the physical nexus defined the previous concept, this does not mean that the objectives changed with that paradigm. The question now is whether the two definitions could coexist – is it possible to have a Significant Digital Presence Permanent Establishment as well as a Permanent Establishment based on a physical nexus?
The attribution of profits defined for Significant Digital Presence Permanent Establishments is not helpful in this regard. Bluntly put, the Authorized OECD Approach (functional analysis) would still apply, but the significant people functions that are essential to that analysis would need to be altered to fit the digital world. The Directive states:
“[…] The situation occurs where a significant digital presence operates through a digital interface without any physical presence in a certain jurisdiction or where no significant people functions are performed in the jurisdiction of the significant digital presence.
[…] The attribution of profits should take into account the development, enhancement, maintenance, protection and exploitation of intangible assets in the performance of the economically significant activities by the digital presence, even if these are not linked to people functions in the same Member State. […]
Each of the economically significant activities contributes to the value creation in the digital business models in a unique manner and is an integral part of these models. The profit split method would, therefore, often be considered as the most appropriate method to attribute profits to the significant digital presence.” 16
Shortly, in the digital world, one must replace relevant people functions with relevant economic functions performed through a digital interface. The attribution of profits, in this regard, becomes a nightmarish task: one can no longer rely on people doing things to see who creates value 17, but one must determine possible splitting factors between jurisdictions which somehow resemble a fair return based on expenses incurred weighted by market share.
Rather surprisingly, these considerations are not far from the once remote possibility of performing a formulary apportionment review to allocate taxes between OECD member countries 18. In effect, depending on the allocation of profits to Significant Digital Presence Permanent Establishments, States will end-up taxing profits based on profit splitting methods based on formulas!
One wonders how this method of attribution would function when a traditional physical nexus approach is also applicable. The margins of intersection between both considerations are enormous, particularly due to the different nature of the basis to determine profits: people functions versus digital presence.
A lot of red tape is still necessary to have a better idea of how this will look in 2020. Further versions of this paper will explore the consequences of amendments to the Directive in the future, as the outline for the future OECD version falls into its final form.
Conclusion: what about the developing world?
Completely digital business models wreak havoc in jurisdictions that are unprepared to host that type of economic activity. This is particularly true for countries that consume large amounts of digital services and have no legal tools to bring those profits within their own borders without contravening treaty limitations, resorting ultimately to excessively high digital service VAT regimes 19 to account for lost Corporate Income Tax revenue.
VAT, however, is unperfect, and unsustainable in the long run 20. This tax cannot serve as a substitute for Corporate Income Taxation, particularly because it hurts low-income consumers the most 21 and restricts their access to technology 22 – a key engine of economic development.
Assuming that a new version of the Permanent Establishment will surface in the near future as an alternative in the OECD Model Tax Convention, developing countries will be faced with the daunting task of implementing and monitoring digital presences that generate significant economic activity without owning a single brick (or mortar) in those countries.
This task, however, will probably be met with enthusiasm, given the potential revenue involved in taxing Permanent Establishments based on the number of users in a given country. Populous nations will likely end up with a large part of the tax revenue that would otherwise be beached in tax havens in eternal deferral loops or repatriated with low, exceptional, tax rates in developed countries.
If the transformation of business models does imply a full transition to purely digital economic activity, developing countries will benefit the most from a well-designed Significant Digital Presence regime that coexists with pre-2020 tax treaties and is based on a formulary apportionment-style mechanism to allocate profits among the myriad Permanent Establishments that would come into existence worldwide. Countries must, therefore, prepare themselves to update their local legislations and recognize the existence of Permanent Establishments internally if they wish to apply this form of legislation to companies in non-treaty jurisdictions.
This is particularly true in countries that rely exclusively on source taxation mechanisms to capture outbound payments for the importation of foreign goods and services over the internet. So far, this strategy has worked fine for countries like Brazil, where the concept of Permanent Establishment is completely underdeveloped, existing only in the wording of Double Tax Treaties signed by Brazil and never enforced or audited outside the very narrow scope of the local version of dependent agent taxation or far fletched administrative rulings. Brazil has very much to gain in terms of tax collection from the adoption of a Significant Digital Presence Permanent Establishment, even more so than it currently collects in WHT and other taxes on imports, since its tax revenues would be directly proportional to the size and relevance of its internal market – by far the largest in the region 23.
Regarding the physical nexus Permanent Establishment, world industries are still heavily dependent on some form of human interaction to maximize profits. Although the current version of the Commentary and the Model will need to be carefully and extensively amended to accommodate the Significant Digital Presence test, this does not mean that the traditional Permanent Establishment will cease to exist. Quite to the contrary, if the Authorized OECD Approach remains essentially unchanged, much of what is applicable today (subject to certain adjustments) will still be applicable for decades to come.
The Permanent Establishment in the age of “scale without mass”, it seems, is alive and well.
ACEMOGLU, D.; JOHNSON, S.; e ROBINSON, J. A. Institutions as a fundamental cause of long-run growth. In: AGHION, P.; e DURLAUF, S.N. Handbook of economic growth. Amsterdam: Elsevier, 2005.
AVI-YONAH, R. S.; e CLAUSING, K. A. A proposal to adopt formulary apportionment for corporate income taxation: The Hamilton Project. University of Michigan Public Law, Working Paper No. 85, 2007.
CEPAL. Panorama fiscal de América Latina y el Caribe. Santiago: CEPAL, 2019.
EUROPEAN COMISSION. Proposal for a Council Directive laying down rules relating to the corporate taxation of a significant digital presence. Brussels: European Comission, 2018.
GRAVELLE, J. G. The distributional case against a VAT. The VAT reader, 2011.
IAB. IAB internet advertising revenue report 2017 full year results. New York: PwC, 2018. Disponível em: <https://www.iab.com/wp-content/uploads/2019/05/Full-Year-2018-IAB-Internet-Advertising-Revenue-Report.pdf>.
OECD. 2010 report on the attribution of profits to permanent establishments. Paris: OECD Publishing – CTPA, 2010.
______. 2017 update to the Model Tax Convention. Paris: OECD Publishing, 2017.
______. Model Tax Convention on income and on capital: condensed version 2017. OECD Publishing, 2017.
______. Additional guidance on the attribution of profits to permanent establishments, BEPS Action 7. Paris: OECD Publishing, 2018.
______. Tax challenges arising from digitalisation – interim report 2018: inclusive framework on. Paris: OECD Publishing, 2018.
TODER, E.; e ROSENBERG, J. Effects of imposing a value-added tax to replace payroll taxes or corporate taxes. Washington D.C.: Tax Policy Center, 2010.
- OECD, 2018, p. 9. ↩
- OECD, 2017 update to the Model Tax Convention, 2017, p. 95. ↩
- OECD, 2017 update to the Model Tax Convention, 2017, p. 16. ↩
- OECD, Tax challenges arising from digitalisation – interim report 2018: inclusive framework on, 2018, p. 36. ↩
- OECD, Additional guidance on the attribution of profits to permanent establishments, BEPS Action 7, 2018, p. 19. ↩
- OECD, Model Tax Convention on income and on capital: condensed version 2017, 2017, p. 154. ↩
- OECD, Tax challenges arising from digitalisation – interim report 2018: inclusive framework on, 2018, p. 38. ↩
- IAB, 2018, p. 27. ↩
- OECD, Model Tax Convention on income and on capital: condensed version 2017, 2017, p. 154. ↩
- OECD, Tax challenges arising from digitalisation – interim report 2018: inclusive framework on, 2018, p. 34. ↩
- European Comission, 2018, p. 6. ↩
- European Comission, 2018, p. 7-8. ↩
- OECD, 2010 report on the attribution of profits to permanent establishments, 2010, p. 21-27; 67-71. ↩
- OECD, 2017 update to the Model Tax Convention, 2017, p. 152. ↩
- European Comission, 2018, p. 16. ↩
- European Comission, 2018, p. 9. ↩
- OECD, 2010 report on the attribution of profits to permanent establishments, 2010, p. 16. 24OECD, 2010 report on the attribution of profits to permanent establishments, 2010, p. 16. ↩
- Avi-Yonah e Clausing, 2007. ↩
- CEPAL, 2019. ↩
- Gravelle, 2011; Toder e Rosenberg, 2010. ↩
- Gravelle, 2011, p. 103. ↩
- Acemoglu, Johnson e Robinson, 2005, p. 395. ↩
- CEPAL, 2019. ↩