Tax Challenges Arising from Digitalisation – Report on Pillar One Blueprint

2. Scope

2.1. Overview

21. The new taxing right established through Amount A only applies to those MNE groups that fall within the defined scope of Amount A. The scope of Amount A is based on two elements: an activity test and a threshold test.

2.1.1. In-scope activities

22. The definition of the scope responds to the need to revisit taxing rules in response to a changed economy. The existing international tax rules generally attach a taxing right to profits deriving from a physical presence in a jurisdiction. However, given globalisation and the digitalisation of the economy, businesses can, with or without the benefit of local physical operations, participate in an active and sustained manner in the economic life of a market jurisdiction, through engagement extending beyond the mere conclusion of sales, in order to increase the value of their products, their sales and thus their profits. Such participation is attributable to the nature of what is being supplied, how it is being supplied and the active interaction or engagement with market jurisdictions. This means that the allocation of taxing rights and taxable profits can no longer be exclusively circumscribed by reference to physical presence.

23. As such, the Outline has proposed that the scope for the application of Amount A be based on an activity test. The definition of the in-scope activities reflects the types of activities where this policy challenge is most acute: ADS and CFB. Within each, the definitions are designed with sufficient breadth to accommodate new business models (and thus ensure a level playing field over time). Considerable technical work has been devoted to defining ADS and CFB activities. However, as noted in the executive summary, scope is one of the key pending political issues and the section needs to read within this context.

2.1.2. Automated digital services

24. The approach to defining ADS recognises that certain MNEs can generate revenue from the provision of ADS (including revenue from the monetisation of data) that are provided on an automated and standardised basis to a large and global customer or user base and can do so remotely to customers in markets with little or no local infrastructure. They often exploit powerful customer or user network effects and generate substantial value from interaction with users and customers. They often benefit from data and content contributions made by users and from the intensive monitoring of users’ activities and the exploitation of corresponding data. In some models, the customers may interact on an almost continuous basis with the supplier’s facilities and services. The ADS scope is not limited with respect to whether the services are provided to consumers, but goes beyond this, given that the possibilities for significant and sustained engagement in the market by ADS businesses is not dependent on the type of customer.

25. The definition of ADS is comprised of a positive list of ADS activities; a negative list of non-ADS activities; and a general definition. Providing a positive list and a negative list provides certainty to MNEs and tax administrations in respect of existing business models; as well as flexibility for the future, as the lists can be updated from time to time as business models evolve. The general definition ensures that a rule remains in place to address rapidly changing business models not otherwise addressed by the positive or negative list. In practice, businesses and tax administrations will apply the rules by going through the following process. First, they will identify whether an activity is on the positive list (i.e. included); if it is, it is an ADS business. Second, if the activity is not on the positive list, they will need to check whether it is on the negative list (i.e. excluded); if it is on the negative list, it is not an ADS business. Only if an activity is not on either list is it necessary to consider whether it meets the conditions in the general definition. Finally, all questions of scope are covered by the early tax certainty process discussed in Chapter 9(including cases where an MNE ceases to be in scope).

26. The general definition of ADS (which also informs the positive and negative lists) is built on two elements:

• automated, i.e. once the system is set up the provision of the service to a particular user requires minimal human involvement on the part of the service provider; and

• digital, i.e. provided over the Internet or an electronic network.

27. The first part, “automated,” reflects the fact that that the user’s ability to make use of the service is possible because of the equipment and systems in place, which allow the user to obtain the service automatically, as opposed to requiring a bespoke interaction with the supplier to provide the service. In determining whether a service requires minimal human involvement the test only looks to the supplier of the service, without regard to any human involvement on the side of the user (e.g. where the user may input certain parameters into an automated system to obtain a customised result). The definition focuses on the provision of the service and therefore does not include human interventions in creating or supporting the system, such as setting up the system environment needed for the provision of the service, maintaining and updating the system environment, dealing with system errors, or making other generic, non-specific adjustments unrelated to individual user requests. Typically, such businesses have an ability to scale up and provide the same type of service to new users on an automated basis with minimal human involvement, with nil or minimal marginal cost.

28. The second part, “digital,” distinguishes it from other service provision methods, such as the on-site physical performance of a service.

29. The positive list includes online advertising services; sale or other alienation of user data; online search engines; social media platforms; online intermediation platforms; digital content services; online gaming; standardised online teaching services; and cloud computing services.

30. The negative list of non-ADS activities includes customised professional services; customised online teaching services; online sale of goods and services other than ADS; revenue from the sale of a physical good irrespective of network connectivity (“Internet of things”); and services providing access to the Internet or another electronic network.

2.1.3. Consumer-facing businesses

31. The inclusion of a broader group of CFB in the scope of Amount A recognises that the ability to participate in an active and sustained manner in the economic life of a market jurisdiction goes beyond businesses that provide ADS. Although this set of businesses includes traditional businesses that have been disrupted to a lesser degree by digitalisation, CFB are able to engage with consumers in a meaningful way beyond having a local physical presence and can thereby substantially improve the value of their products and increase their sales. This significant and sustained engagement is able to take place and create value for consumer-facing MNEs because of the broader digitalisation of the economy, as technology facilitates more targeted marketing and branding, and the collection and exploitation of individual consumer data – all of which can be achieved with greater efficiency and remotely. Consumer relationships, interactions with users and consumers and wider consumer-facing intangibles drive value for these businesses. However, the current tax rules allow them to earn residual profits associated with such intangibles remotely and without a commensurate share assigned to the market jurisdiction.

32. The inclusion of CFB in the scope of Amount A not only reflects this principle, but also recognises that in some cases, including both ADS and CFB can ensure consistent treatment of different business models that are sometimes combined. For example, certain online intermediation platforms provide both a marketplace for the sale of third parties’ goods to consumers (in scope as ADS) as well as their own inventory of similar goods (in scope as CFB); music, films and certain software may be delivered online (in scope as ADS) and the same material delivered by the MNE on a physical medium (in scope as CFB); entertainment companies may directly make their content available online (in scope as ADS) as well as license the same material to others (in scope as CFB). For activities that may be both ADS and CFB, the ADS definition applies.

33. CFBs are defined as those businesses that generate revenue from the sale of goods and services of a type commonly sold to consumers, including those selling indirectly through intermediaries and by way of franchising and licensing. This definition can be further broken down into the following elements:

• “Consumer” means an individual (whether or not the direct purchaser) who acquires a good or service for personal purposes, rather than for commercial or professional purposes.

• “A good or service is “of a type commonly” sold to consumers if the nature of the good or service is such that it is designed primarily for sale to consumers. This presupposes that the good or service is made available in ways capable of being for personal consumption (such as in portions, in sufficiently finished or usable form, or at purchase points accessible by an individual, as opposed to bulk or raw material accessible to wholesale traders or other businesses only). To be designed primarily for sale to consumers means that the MNE developed the goods or services to be regularly, repeatedly, or ordinarily supplied to consumers (whether directly or indirectly), such as by engaging in consumer market research, marketing and promoting it to consumers, using consumer / user data, or providing consumer feedback or support services (irrespective of the location in which such activities take place). Conversely, an unusual or infrequent transaction with a consumer does not affect the characterisation of the goods or service.

• “Sold to” includes sale, lease, license, rent or delivery, whether directly or indirectly (e.g. through a broker, agent, intermediary or representative). This means that the product or service may still be a consumer good or service even if the contracting party to the sale is not the final consumer (such as a sale via a third party distributor, or where a franchisor has created a product but the sale is contracted by the franchisee). It also covers cases where the nature of the product is such that to be exploited it is licensed, rather than sold (e.g. rights to music), if it is otherwise of a type commonly licensed to consumers.

• CFB are (i) the MNE that is the owner of the consumer product / service and holder of the rights to the connected intangible property (including franchisors and licensors), i.e. the MNE whose “face” is apparent to the consumer; and (ii) the MNE that is the “retailer” or other contractual counterparty of the consumer (if separate from the “owner”), as they have a direct relationship to the consumer (and including franchisees and licensees that sell the consumer product directly). It does not include other third party MNEs, such as manufacturers, wholesalers and distributors, which have no relationship with the customer – whether contractual or otherwise.

34. Guidance has also been developed to clarify the possibilities for applying the definition to the pharmaceutical sector; to show how this concept applies to capture certain in-scope intermediaries, franchising and licensing; to provide guidance on certain products that could be characterised as both for consumers and businesses (“dual category”); and to delineate the situations in which a component product can be in scope. This guidance is set out in this chapter below under the “activity tests.”

2.1.4. Exclusions from scope

35. As a consequence of articulating the in-scope activities and the policy rationale underpinning the approach, this chapter also identifies certain businesses that are not intended to be in scope of Amount A. The exclusions from scope and scope clarifications provide certainty for those sectors where the policy challenges of the digitalised economy do not present themselves. Sectors not in scope of Amount A include certain natural resources; certain financial services; construction, sale and leasing of residential property; and international air and shipping business.

Thresholds

36. The second element of defining the scope is the threshold test. An MNE can only be in scope of Amount A if it meets the activity test described above and two thresholds: the MNE’s consolidated revenue is above a certain threshold; and its in-scope revenue earned outside its domestic market is also above a certain threshold. This ensures that Amount A focusses on the largest MNEs that generate residual profit available for reallocation under Amount A, and that the compliance and administrative burden is proportionate to the expected tax benefits. This will also keep the number of MNEs affected at an administrable level for tax administrations (including the early tax certainty process).

37. Having regard to the economic impact assessment,1 setting a threshold below [EUR 750 million] would lead to substantial compliance burdens but without commensurate benefits in terms of the available reallocation for market jurisdictions. At the same time, even at that threshold, there is a significant number of MNEs that would be in scope, and consideration needs to be given to operating the new Amount A system, including the early certainty process, as efficiently as possible particularly in the initial years. As such, an option to operate the thresholds on a transition or phase in basis is being considered, whereby the two thresholds are initially set at a higher level and gradually reduced over time.

2.2. Activity tests

38. The activity tests proposed in the Outline align the scope of Amount A with its policy objective. They are designed to capture those MNEs that are able to participate in a sustained and significant manner in the economic life of a market jurisdiction, without necessarily having a commensurate level of taxable presence in that market (as based on existing nexus rules). This covers MNEs under the broad categories of ADS and CFB. The definitions of each are discussed in turn, followed by a discussion of the sectors that are excluded from the scope of Amount A.

2.2.1. Automated digital services

39. The proposed definition for ADS is comprised of a general definition; a positive list of services that are in any event in scope of ADS business and a negative list of services that are not in scope of ADS business.

40. The general definition, supported by updatable lists, provides certainty combined with flexibility. The benefit of using a general definition supported by updateable lists is that it will be capable of accommodating rapid changes in technology that give rise to new types of ADS not otherwise contemplated by the positive or negative list, and allow the new Amount A rules to stand the test of time. As it would not be possible for legislators to continuously maintain exhaustive positive and negative lists that keep pace with these developments, the general definition ensures that there is a rule in place to bring in scope (or exclude from scope) business models that are not otherwise explicitly contained on the positive list or negative list, until a new or evolved business model can be included on the positive or negative list. The application of the general definition will be supported by an early certainty process, to ensure that MNEs and tax administrations have certainty and consistency in the application of the scope.

41. The benefit of using the positive and negative lists is that it provides certainty and precision with respect to the business models that are currently known, and avoids the need for those businesses to apply the general test. As these lists are then updated as jurisdictions gain experience with the rules (including being informed by the early certainty process) and new business models evolve, those lists will also continue to provide certainty over time.

42. In practice, businesses and tax administrations could generally apply the rules by going through the following process. First, identify whether an activity is on the positive list (i.e. included); if it is, it is an ADS business. Second, if the activity is not on the positive list, identify whether it is on the negative list (i.e. excluded); if it is, it is not an ADS business. Only if an activity is not on either list is it necessary to consider whether, on first principles, it meets the conditions in the general definition. It is noted that if an MNE does not fall within the definition of ADS following this approach, it may however be in scope as consumer-facing business. The legal implementation of the positive and negative lists and the general definition (such as in a multilateral convention) remains under consideration, as well as more broadly the process and implementation of changes to the lists.

43. Although the general definition would apply only in residual cases of activities that are not identified in either the positive or negative list, and therefore as a backstop to those lists, the principle itself has informed the creation of those lists. For this reason, this chapter first presents the general definition, and then the content of the positive and negative lists, a discussion of dual category / bundled ADS, and finally other definitions that apply for ADS. In each case, the rule appears in the first box, followed by the accompanying commentary in the second box.

Box 2.1. ADS – General definition

An ADS1 is one where:

• The service is on the positive list; or

• The service is

  • automated (i.e. once the system is set up the provision of the service to a particular user requires minimal human involvement on the part of the service provider); and

  • digital (i.e. provided over the Internet or an electronic network); and

  • it is not on the negative list.

1. The provision of services may need to be defined to provide further clarity.

Box 2.2. ADS – Commentary

The definition of ADS is structured around a general definition, combined with both a positive list, and a negative list. An activity on the positive list is an ADS business. An activity on the negative list is not an ADS business.1

The first condition, being whether the service is automated, reflects the fact that that the user’s ability to make use of the service is possible because of the equipment and systems in place, which allow the user to obtain the service automatically, as opposed to requiring a bespoke interaction with the supplier to provide the service.

In determining whether a service requires minimal human involvement the test only looks to the supplier of the service, without regard to any human involvement on the side of the user (e.g. where the user may input certain parameters into an automated system to obtain a customised result). Furthermore, the definition focuses on the provision of the service and therefore does not include human involvement in creating or supporting the system, such as setting up the system environment needed for the provision of the service, maintaining and updating the system environment, dealing with system errors, or making other generic, non-specific adjustments unrelated to individual user requests. Finally, the threshold of minimal human involvement would not be crossed where the provision of the service to new users involves very limited human response to individual user requests / input at the service delivery point or where in individual cases involving particular, more complex problems, the programmes running the system direct the customer to a staff member.

A general feature of the concept of “automated” is whether there is an ability to scale up and provide the same type of service to new users with minimal human involvement. This feature aims to identify ADS businesses that benefit from significant economies of scale, rather than to suggest that there is no human involvement required in the business. For many ADS businesses, developing the system that delivers the offered service may require a large degree of upfront human involvement and capital inputs (such as creating algorithms to deliver the automated service including such features as tailoring the offering to the user’s preferences). It distinguishes ADS businesses by looking to whether the marginal cost in terms of additional human involvement of providing the same services to additional users is nil or almost nil. In other words, once the offered service of an ADS business is developed (such as the music catalogue or social media platform), then the business can provide that service to one user, or to one thousand users, on an automated basis with the same basic business processes and therefore benefit from scale without mass in the market jurisdiction; whereas many non-ADS businesses would have material per unit costs in connection with providing the services to new customers.

With respect to the second condition of the general definition (i.e. that the service is digital), it is inherent in the nature of ADS that it will be provided over the Internet or an electronic network. This distinguishes it from other service provision methods, such as the on-site physical performance of a service. For example, an online intermediary platform website that provides the service of bringing users and hotel offerings together and allowing users to book a hotel online, derives revenue for providing that online intermediation service. It is in scope as ADS. This is different from the case of the hotel itself, which earns revenue for providing physical accommodation, which is not provided over the Internet. Although the hotel may have an online booking service on its website, it is not deriving revenue from providing that function as a service per se but from providing the accommodation (and the online booking is of no use without the main activity of providing that accommodation). It is not in scope of ADS (but would be in the scope of CFB). This condition does not distinguish between different Internet or electronic network transmission methods. It is also not affected by whether or not the service provider owns, leases or otherwise controls the transmission equipment.

The third condition is that the service is not on the negative list, noted above. This ensures that the negative list takes precedence over the general definition.

1. The definition of ADS draws upon, but is different from, concepts contained in the definition of “electronically supplied services” used in the European Union VAT law. The intention is to avoid conflation between both concepts (and avoid any difficult questions as to the relevance of the associated EU case law for the Inclusive Framework).

Positive list

44. As noted above, the approach to defining ADS starts with the proposition that if an activity is on the positive list, it is an ADS (with no need to additionally apply the general definition). The positive list currently contains the following nine categories of services.

• Online advertising services;

• Sale or other alienation of user data;

• Online search engines;

• Social media platforms;

• Online intermediation platforms;

• Digital content services;

• Online gaming;

• Standardised online teaching services; and

• Cloud computing services.

45. The categories are not mutually exclusive and may overlap (for example, a digital content service could be funded in whole or in part by online advertising). Thus, an ADS business may be comprised of multiple ADS – some of which are directly revenue generating and others of which are not. As another example of such a two-sided business model, an online advertising service may obtain a higher price for its advertisements by targeting them based on data about viewer interests obtained by providing the viewers with “free” access to a search engine or social media platform. As discussed in the chapter on revenue sourcing, ADS will be categorised for revenue sourcing purposes according to the nature of the revenue source for the service. The definitions of each category, supported by a commentary to provide further guidance, are set out in turn below.

Box 2.3 Online advertising services – Definition

This means online services aimed at placing advertisement on a digital interface,1 including services for the purchase, storage and distribution of advertising messages, and for advertising monitoring and performance measurement. It includes related systems for attracting potential viewers of the advertisements and collecting content contributions from them and data regarding them, including via the provision of access to a digital interface, such as search engines, social media platforms or digital content services.

1. See general definitions below for definition of digital interface.

Box 2.4. Online advertising services – Commentary

This category is drafted to be broad and to cover automated online services throughout the online advertising value chain. This includes direct advertising services, such as where social media platforms, online search engines, online intermediation platforms and digital content providers directly sell advertising inventory for display on the digital interfaces they operate. It also extends to the automated systems and processes for the purchase and sale of advertising inventory (such as demand-side platforms, supply-side platforms, ad exchanges, ad verification services, etc.). Given the broad definition of “digital interface” it also includes online advertising displayed on an Internet-connected good (“Internet of things”) provided that there is such an identifiable advertising revenue stream.

Box 2.5. Sale or other alienation of user data – Definition

This means selling,1 licensing or otherwise alienating to an unrelated third party customer user data generated by users of a digital interface.2

1. It is recognised that there may be a question of whether the sale or alienation of personal data is a type of “service,” or indeed whether it can really be “sold.”

2. See general definitions below for definition of “users.”

Box 2.6. Sale or other alienation of user data – Commentary

This category is drafted to capture business models that monetise user data generated on a digital interface by selling, licensing or otherwise alienating it to unrelated third parties.

It applies to data generated by users on a digital interface. The definition of a digital interface is intended to be broad, and would cover Internet-connected interfaces embedded in a physical good (i.e. the Internet of things) regardless of whether the sale of that good itself is within the scope of Amount A.

“User data” includes information such as a user’s habits, spending, location, environment, usage of services, hobbies, or personal interests, including anonymised and aggregated data (including geolocation information and user traffic levels).

Further consideration is being given to the extent to which user data would also include the provision of data such as industrial, scientific, statistical, or other data not linked to natural persons (such as businesses that acquire and disseminate information about investments and financial markets, or scientific research). As the definitions of ADS are not mutually exclusive, such businesses could already be captured under digital content services, where they provide that data in an automated way, such as an online database or online library.

The source of data may be collected as raw data by the MNE itself (e.g. the manufacturer/seller of a home heating system collecting data about energy use, or a social media company collecting data about its users) or it may be acquired from another business. The source of the data would not be relevant for determining scope provided that it is generated by a user through a digital interface.

To the extent that the data is provided as part of a customised service, such as a professional marketing service, the guidance to be provided as set out below under dual category ADS and bundled services would apply.

Box 2.7. Online search engines – Definition

This means making a digital interface available to users for the purpose of allowing them to search across the Internet for webpages or information hosted on digital interfaces.

Box 2.8. Online search engines – Commentary

Many online search engines are monetised through online advertising services and/or services transmitting data about users. To the extent these services are funded via online advertising or the sale of data, such revenue would be treated under those respective categories (including for revenue sourcing purposes).

This category would extend to instances where an online search engine charges users for access, for example under a subscription model, or where online search engine technologies are provided for incorporation into a third-party host website (e.g. a “search box”).

This category would not include services such as online databases or ‘internal’ website search functions that are not monetised, where the search results are limited to data hosted on that same digital interface (or related digital interfaces). However, if an online database or an ‘internal’ website search function service involves monetisation of services and it falls under the positive list or meets the general definition of ADS, it will be covered under ADS.

Box 2.9. Social media platforms – Definition

This means making a platform available on a digital interface to facilitate the interaction between users or between users and user-generated content.

Box 2.10. Social media platforms – Commentary

This category would include a range of activities that rely on an active and engaged user base to create value, such as social and professional networking websites, micro-blogging platforms, video or image sharing platforms, online dating websites, platforms dedicated to sharing user reviews, as well as online call and messaging platforms, some of which could overlap with online intermediation platforms.

This category would extend to instances where the social media platform charges users for access, for example under a subscription model. To the extent these services are funded via online advertising, the sale of data, or subscription models (which follows the business model for digital content services), such revenue would be treated under those respective categories (including for revenue sourcing purposes).

This category does not extend to instances where user interaction is merely incidental to the core purpose of the digital interface, for example where a company sells its own inventory online and the website allows users to post comments or reviews or where a website allows a user to engage in an online chat with a sales representative.

Box 2.11. Online intermediation platforms – Definition

This means making a platform available on a digital interface to enable users to sell, lease, advertise, display or otherwise offer goods or services to other users. It does not include the online sale of goods and services of the platform’s own inventory.

Box 2.12. Online intermediation platforms – Commentary

This category would apply where the service enables the interaction between third party users, irrespective of the nature of the interaction, the characteristics of the users involved, whether an underlying transaction is itself in scope, or the extent of the service provider’s activities in facilitating the interaction.

This category would extend to instances where the online intermediation platform charges users for its online intermediation services, for example through commission, listing or subscription fees. To the extent the intermediation service is funded via online advertising or the sale of data, such revenue would be treated under those respective categories (including for revenue sourcing purposes).

The definition would include intermediation services, and would not include the online sale of goods and services which form part of the platform’s own inventory (which may however be captured under CFB). This result would also be achieved through the requirement in the definition that there must be more than one user (i.e. the reference to “users” in the plural), read together with the definition of a “user” which specifically excludes the provider of the service or any entity in the same group. The online sale of goods and services other than ADS (e.g. the direct sale to consumers by an MNE of its own inventory on its own website) is also confirmed as out-of-scope in the negative list.

However, there may be exceptional cases where an MNE does sell its own goods or services, while essentially performing an intermediation service and remaining insulated from inventory risk. As such, further consideration is being given to address instances where an online intermediation platform such as an online marketplace operates on a resell model, without incurring the ordinary commercial risks associated with the provision of the underlying product (such as where the business only takes flash-title and is insulated from inventory risk, is fully indemnified against product liability risk and credit risk). For example, after a user secures a reservation for a car rental service through an online platform, the platform service provider may purchase the car rental service itself, after receiving the customer’s order, before immediately reselling it to the customer. However, the platform service provider is not incurring any of the ordinary commercial risks associated with the provision of the underlying car rental service (e.g. inventory risk) and its service arguably remains, in essence, one of intermediation.

Box 2.13. Digital content services – Definition

This means the automated provision of content through digital means, whether by way of online streaming, accessing or downloading digital content (e.g. music, books, videos, texts, games, applications, computer programmes, software, online newspapers, online libraries and online databases), whether for access one time, for a limited period or in perpetuity.

Box 2.14. Digital content services – Commentary

This category is drafted to capture the different forms which digital content can take when acquired by a user. This category includes, for example, music, books, videos, texts, games, applications, computer programmes, software, online newspapers, online libraries and online databases (such as subscription-based statistical or academic online databases).

It does not include simply making a digital interface available to users. The purpose from a user’s perspective in a digital service transaction is the acquisition of the digital content, whether for access one time, for a limited period or in perpetuity.

It includes streaming, accessing or downloading digital content. By way of background, the streaming and downloading consists of the same process of transmitting data, either as a continuous flow (in effect like a temporary download) or as a file saved to the user’s device available for later use. A number of streaming services allow both temporary streaming and downloading, but from the perspective of the ADS provider, the process is essentially the same. Therefore, for the purposes of the scope of Amount A, digital content streaming should include accessing and downloading, under the inclusive heading of “digital content services.” Otherwise, distorted and administratively burdensome results would follow if streaming series were only included in scope to the extent that users were temporarily “streaming” rather than downloading the material.

The sale of software was enumerated as a CFB in the Outline. However, the sale of software also meets the general definition of ADS where the provision of that software is automated requiring minimal human involvement to make it available to users and the software is delivered over the Internet. As such, software that is accessed or downloaded over the Internet would therefore qualify as ADS under the “digital content services” category (or it may also qualify as ADS under the “cloud computing” category (“Software-as-a-service”).

There are two exceptions to this. First, to the extent that software is acquired as a tangible product (e.g. a packaged product on a CD that does not require any connection to the Internet to access the software), then it would not be an ADS (although it could be included as a CFB alongside other physical, offline items such as music, films, books or computer games purchased on a physical medium.) In order not to create distortions, where the tangible medium has no other purpose than to provide the user with information required to access the digital content via Internet or electronic network (such as a printed access code for downloading the material), then it would remain as ADS.

Second, reflecting the discussion above and the “automated” aspect of the definition, highly customised software that has been designed for a particular businesses’ needs (and which may be negotiated with tailored pricing depending on the software package chosen and various elements incorporated) would not be included as ADS, even if the final product is made available online. This is because it would require significant human involvement, more akin to professional services, and would not reflect the same idea of there being limited additional unit costs to make the software available to additional users. This refers to customisation undertaken at the specific request of one customer, in response to its unique requirements. Where software has been designed with particular options or features, which a customer can choose or personalise without any significant human involvement on behalf of the service provider during the online purchase, such software would be in scope of ADS. Thus, the degree of customisation and the level of human involvement should be the primary factors to characterise the provision of software as ADS or otherwise. If the provision of the software was bundled, comprising standardised modules and customised modules, the guidance to be provided as set out below under dual category ADS and bundled services would apply.

Box 2.15. Online gaming – Definition

This means making a digital interface available for the purposes of allowing users to interact with one another in the same game environment.

Box 2.16 Online gaming – Commentary

This category would apply irrespective of whether the access to the game is by way of fee or available for free. It applies to all multiplayer gaming enabled by the Internet, such as massively multiplayer online (MMO) games, or other games that enable multiplayer functionalities, and regardless of the device or platform the game is accessed through.

The provision of in-game purchases, or any other online purchases within the game would also be in scope under this category.

This category would not generally include single-player games (which if streamed, accessed or downloaded over the Internet would be in scope of digital content services) or the purchase of a game sold on tangible media (as above for software, and which would be in scope of CFB).

Box 2.17. Standardised online teaching services – Definition

This means the provision of an online education programme provided to users, which does not require:

(i) the live presence of an instructor; or

(ii) significant customisation on behalf of an instructor to a particular user or limited group of users, whether with respect to the curriculum, teaching materials, or feedback provided.

Box 2.18. Standardised online teaching services – Commentary

This category would include pre-packaged, non-customised education products such as a pre-recorded series of lectures, and the content of which is not customised to each individual user (e.g. massive open online courses). Although these services may allow users to discuss the course content with each other on discussion forums within the platform, there is no or only limited interaction with instructors. Another key feature of standardised teaching services is that coursework completed by the user is generally not marked by the instructors, but either marked automatically, or by other users. Non-ADS online teaching services may be captured under consumer-facing business.

This category would not cover online education products that are customised to a student, or to a limited group of students, which may incorporate certain ancillary elements that are automated (e.g. a pre-recorded lecture offered as part of a customised education package; automatically graded assignments accompanying a live-streamed lecture). Such services would also not meet the general definitions of ADS, and as such are included in the negative list under the category of customised online teaching services.

Box 2.19. Cloud computing services – Definition

This means the provision of network access to on-demand standardised information technology (IT) resources, including infrastructure as a service, platforms as a service, or software as a service (such as computing services, storage services, database services, migration services, networking and content delivery services, webhosting, and end-user applications and software).

Box 2.20. Cloud computing services – Commentary

The network access to on-demand standardised IT resources includes all types of standardised cloud computing services, including computing services, storage services, database services, migration services, networking and content delivery services, webhosting, and end-user applications and software.

• Computing services include virtual servers in the cloud, the ability to run and manage web apps using remote computing, the ability to run code on remote computers in response to events and the ability to run batch code jobs at scale;

• Storage services include storage in the cloud and data transport;

• Database services include data warehousing, database management and caching systems;

• Migration services include database migration and data transport;

• Networking and content delivery services include access to a virtual private cloud (an isolated cloud that the customer can control) and use of a global content delivery network (whereby content such as videos are transferred to viewers at high transfer speeds);

• Web hosting services include website and webpage hosting; and

• End-user applications and software services include systems permitting users to develop access or use software and applications.

Cloud computing services1 are typically provided in a standardised and highly automated way. Standardised cloud computing services may be ‘assembled’ or configured together for a particular customer (whether by the service provider or by the customer on a self-serve basis). For example, a business may choose to obtain a certain combination of data storage, data security and web-hosting services from a cloud computing provider. Each such cloud computing service is standardised, irrespective of the particular combination of services chosen by the client. In other words, although a customer’s activities may be different (e.g. making available different music on their platform or hosting a different website), from a cloud computing company’s perspective the act of providing the data storage or hosting the website in the cloud is essentially the same.

Some cloud computing services, however, involve a high degree of human involvement to customise the service to the needs of a particular client. For example, a hospital may hire a cloud computing service provider to develop a bespoke cloud-based IT system to manage its complex operations (e.g. software to track patient care and medical records, IT security solutions to respect the applicable legal requirements for patient confidentiality, etc.). Such services are closer to engineering and consulting services, i.e. professional services, which are on the negative list.

In order to ensure that the cloud computing services category on the positive list only captures those that reflect the principles of ADS, the proposed definition refers to ‘standardised’ cloud computing services. This way, a bespoke cloud solution involving a high degree of human involvement on behalf of the provider’s staff (e.g. engineers or consultants) to create a new computing solution (as opposed to configuring existing solutions) would not be included. To the extent that the human involvement relates only to the configuration of standardised cloud computing products, the integration of standardised cloud computing products into a customer’s existing IT architecture, or ancillary customer support, the human involvement will be considered ancillary to the cloud computing service, which would be covered by this category.

Where a cloud computing service provider is in the business of providing a bundled service comprising both standardised and customised services (including where customised services are built upon standardised modules), the guidance to be developed under “dual category ADS and bundled services” below would apply.

1. For the avoidance of doubt, the reference to cloud computing services refers to the MNE that provides the cloud services. Even though from a user perspective, accessing another type of ADS (such as digital content) could be (in an indirect sense) accessing a cloud service, for present purposes the question is about whether the revenue is generated from the business activity of providing cloud computing services, as opposed to generating revenue from a business activity which is in turn supported by a cloud computing service.

Negative list

46. If an activity is not on the positive list, the next step would be to determine if it is on the negative list. If so, it is not an ADS. The proposed negative list currently contains the following five categories of services:

• Customised professional services;

• Customised online teaching services;

• Online sale of goods and services other than ADS;

• Revenue from the sale of a physical good, irrespective of network connectivity (“Internet of things”); and

• Services providing access to the Internet or another electronic network.

47. This section proposes definitions for the services that would be contained on the negative list, together with a Commentary.

Box 2.21. Customised professional services – Definition

This means customised professional services, whether provided individually or as a firm, such as legal, accounting, architectural, engineering and medical services.

Box 2.22. Customised professional services – Commentary

This category confirms that customised professional services are not within the general definition of ADS. Although such services may be delivered online (e.g. legal advice sent by email, an architect sending drawings in PDF format; or an accountant sending calculations in a spreadsheet), they require customisation to each client, through the tailored exercise of professional judgment and bespoke interactions. These services are not automated and require more than minimal human involvement on behalf of the professional individual or firm. They would also not be scalable without additional human involvement.

Where a professional service relies heavily on ADS, for example if a law firm relies on artificial intelligence software (AI) to conduct due diligence, or an architect’s firm to draw plans, revenue from the provision of the professional service itself will remain out-of-scope of ADS. This is because human involvement is required on behalf of the professional to use the AI and exercise professional judgment in order to provide the final service product to the client. Payments made by such a firm to a third-party AI provider, however, will generally be captured under ADS (as cloud computing or digital content services).

By contrast, where a user directly accesses an automated service online that may be equivalent to a professional service (e.g. if a user self-serves legal advice on a dedicated platform) then such service would qualify as ADS to the extent that it meets a category on the positive list or the elements of the general definition.

Box 2.23. Customised online teaching services – Definition

This means live or recorded teaching services delivered online, where the teacher customises the service (such as by providing individualised, non-automated feedback and support) to the needs of the student or limited group of students and the Internet or electronic network is used as a tool simply for communication between the teacher and student.

Box 2.24. Customised online teaching services – Commentary

This category confirms that customised teaching services delivered online would not be within the general definition of ADS where the Internet or electronic network is used as a tool simply for communication between the teacher and student.

This includes online education packages that are significantly customised to a student, or to a limited group of students, even where certain ancillary elements of the product are automated (e.g. a pre-recorded lecture offered as part of a customised education package; automatically graded assignments accompanying a live-streamed, customised lecture).

However, standardised online teaching service with ancillary interaction with an instructor is contained in the positive list.

Box 2.25. Online sale of goods and services other than ADS – Definition

This means the sale of a good or service completed through a digital interface where:

• the digital interface is operated by the provider of the good or service;

• the main substance of the transaction is the provision of the good or service; and

• the good or service does not otherwise qualify as an ADS.

Box 2.26. Online sale of goods and services other than ADS – Commentary

This category would apply to sellers that use a digital platform to sell their own non-digital goods and services to customers. While the sale can be transacted over the Internet, these businesses are sellers of non-digital goods and non-digital services, rather than offering a digital service per se. Applying the general definition of ADS, the provision of the intended good / service (e.g. the use of the hotel) is not of a type that is automated but requires additional human interventions to make that service available to additional users. As such, for the sake of clarity, the sale of an MNE’s own goods and services where the order / booking and processing is done electronically are included on the negative list.

See also above on the positive list for online intermediation services.

Box 227 Revenue from the sale of a physical goods, irrespective of network connectivity (‘Internet of things”) – Definition

This would apply irrespective of the network connectivity of that physical good, provided that there is no separately identifiable ADS revenue stream attached to that physical good (either at the time of purchase or a later date).

Box 2.28 Revenue from the sale of physical goods, irrespective of network connectivity (‘Internet of things”) – Commentary

Increasingly, physical goods may be connected to the Internet, or bundled with an online service. There are broadly three categories for analysing that service, as follows:

First, beyond the sale of the physical good, such goods can be additionally monetised with a customer beyond the purchase of the physical good through different revenue streams (whether at the outset at the time of purchase or at a later date), and those revenue streams are captured by existing ADS categories on the positive list. Common examples include:

• Sale or other alienation of user data. The monetisation of data collected from the connected object is ADS, where that meets the definition above. For example, if an automobile maker designs an Internet-connected car that collects location data and the company sells the data about the user’s habits, location and so forth to third parties for marketing purposes, this would meet the definition above, including because the definition of digital interface above is sufficiently broad to capture the monetisation of information collected from an Internet of things device.

• Online advertising services. Online advertising revenue relating to advertisements displayed on the connected object is ADS, captured under online advertising services (e.g. supermarket adverts displayed on an Internet-connected fridge’s interface). This is because the definition of online advertising refers to displaying an advert on a digital interface, which is defined in a sufficiently broad way to include Internet connected devices.

• Other ADS. The user of the connected object may pay for different types of ADS relating to, and/or to be accessed through, the connected object (e.g. subscription payments for a tracking application to be used on a personal bracelet; streaming music through a virtual assistant device; online purchase of upgraded software to activate or enhance a product). These would be captured under the relevant ADS category (e.g. digital content services), which apply regardless of the type of physical good that supports the network connection through which such ADS is delivered.

To the extent that these revenue streams are separately identifiable from the sale price of the physical good, that ADS revenue stream would be captured as ADS. Some typical cases where this is likely to arise are noted in the commentary on the relevant ADS definitions on the positive list above. Further consideration is required to address cases where a separate revenue stream can be inferred even if not explicitly identified as such. For example, if the good is sold in two versions in which the main distinguishing feature is that one includes a subscription to a digital service and the other does not, the price difference may be treated as the revenue associated with the digital service.

Second, there are certain types of machinery and industrial products that may contain a digital component. For example, monitoring the performance of an engine and providing remote technical support. This will typically require significant human involvement to provide the core function, which is using that information to conduct maintenance and repairs on the machinery. This is related to the operation of the machinery, rather than the service provider separately monetising that data in an automated way with a third party. This means that even if the Internet-enabled functionality of the machinery were separately tested under the dual category ADS and bundled services analysis, it would not meet the general definition of ADS or any item on the positive list, meaning the entire item is out-of-scope. For the purpose of certainty, this will be added to the negative list, and revisited at a future point in time if any changes are required.

Third, there are certain consumer products, known as the Internet of things, that provide a network of everyday devices, appliances, and other objects equipped with computer chips and sensors that can collect and transmit data through the Internet, which enables additional features of the product to be used. For example, fitness trackers may give access to an online platform that analyses data, allows interaction with other users, and provides information on workout programmes. It may be that many consumer goods now contain some software and may in the future be Internet-enabled, and bringing all such items into the scope of ADS would be over-broad having regard to the general definition of ADS above, given that the sale of a physical good is not ADS because it is not a service, nor is it provided over the Internet or through an electronic network. One option to address this would be to say that where the physical good embodies an ADS, part of the sale price of that physical good could nevertheless be attributable to that ADS. However, treating a portion of Internet of things goods as within the scope of ADS would be difficult in practice and highly fact intensive for every given product that exists or will exist – and would be likely to generate disputes – as it would require trying to separate the value of the digital component of the service as the ADS, as opposed to the value created by the other parts of the good / service. In any case, the sale of the object, to the extent it is otherwise a consumer-facing product, would already be included in scope under CFB.

Based on the three categories above, the scope of ADS would include the revenue from the Internet of things to the extent separately identifiable as another ADS on the positive list (the first category above), and to exclude the revenue from the physical sale of goods such as industrial or consumer goods. This could be revisited in the future as the Internet of things evolves.

Box 2.29. Services providing access to the Internet or electronic network – Definition

This would apply to the provision of access (i.e. connection, subscription, installation) to the Internet or electronic network, irrespective of the delivery method.

Box 2.30. Services providing access to the Internet or electronic network – Commentary

This category is drafted to clarify that services providing access to the Internet or to an electronic network would be out-of-scope of ADS. This is in line with the policy of Amount A as the provision of such services typically requires a degree of local infrastructure and is subject to local telecommunication regulations. This category covers the provision of access (i.e. connection, subscription, installation) to the Internet or electronic network irrespective of the delivery method, namely over wire, lines, cable, fibre optics, satellite transmission or other means, although this could be revisited if needed as technology advances.

Internet Service Packages (ISPs) in which the Internet access component is an ancillary and subordinate part (i.e. a package that goes beyond mere Internet access comprising various elements (e.g. content pages containing news, weather, travel information; games fora; web-hosting; access to chat-lines etc.)) would not be covered by this category. In such cases, the rules applicable to dual category ADS and bundled packages would apply.

Dual category ADS / bundled packages

48. Where an MNE is engaged in multiple activities, and those are clearly identifiable as separate, stand-alone services by reference to revenue streams, the definitions apply to each activity separately. This includes cases where an MNE provides more than one type of ADS (e.g. an online intermediation platform could be simultaneously operating the intermediation service and running online advertising).

49. However, some MNEs may be engaged in activities that are not clearly severable, and represent a “dual category” or “bundled package.” There are two ways that dual category ADS or bundled services can occur: (i) an aspect of the service meets the ADS definition, but it comes with a non-ADS service; (ii) a physical good that comes with a service, where that service may or may not meet the ADS definition. For example, in the first case, a cloud computing MNE may provide a package that is built on standardised modules, but where a customer can also select highly customised elements which would require significant human involvement. An example of the second case is an Internet-connected physical device, which is discussed further above under the negative list.

50. In some cases, and perhaps increasingly in the future, the ADS and non-ADS elements may be highly integrated and thus be considered to be a single service. Where ADS represents a substantial part of the overall service, and the non-ADS elements derive significant benefits from their connection to the ADS elements, then the overall service might be considered as ADS. By contrast, where the ADS elements are merely ancillary or a technical support feature for the rest of the service (e.g. an automated chat function to screen a user’s request as an entry point to the service), and the rest of the service requires human involvement to provide the service as described above, the overall service may not be considered to be ADS, given the relative substantive contribution of the ADS within the bundled package.

51. Work is ongoing to address this issue. The starting point being contemplated is whether there are multiple supplies that are identifiable and substantive in their own right (which could be evaluated, for example, by whether such supplies generate a separate revenue stream or are invoiced or priced separately for the customer, having regard to the need to avoid planning opportunities, such as arbitrary splitting of invoices), in which case each individual supply would be tested against the definitions. If the supplies are not separately identifiable and substantive in their own right, there could be an evaluation of the supply as a whole. Further guidance will be developed to determine the appropriate materiality threshold above which the ADS element may be considered to be a “substantial part” or highly integrated as part of an overall service, or where elements are ancillary. Consideration is also being given to whether there are ways to provide certainty where a substantive part of the supply is on either the positive or negative list. Guidance will take into account the need to have simple, administrable and implementable rules that create certainty and consistency, as well as the documentation requirements to substantiate the application of the guidance in a given case.

Other definitions for ADS

Box 2.31. Other definitions

For the purposes of ADS, the following definitions would apply:

“Digital interface” means any programme or other system allowing access by users to software, content or other information that is accessible by users online, such as websites and mobile applications, regardless of the type of physical support enabling such access. The definition of a digital interface is intended to be broad, and would cover Internet-connected interfaces embedded in a physical good (i.e. the Internet of things) regardless of whether the sale of that good itself is within the scope of the new taxing right of Amount A.

“Online” means over the Internet or an electronic network;

i. “User” means any individual or business accessing a service, but does not include the provider, or a member of the same MNE group as the provider, of that service;

ii. an employee of a person referred to in paragraph (i) acting in the course of that person’s business.

2.2.2. Consumer-facing businesses

Defining consumer-facing businesses

52. CFBs would include businesses that generate revenue from the sale of goods and services of a type commonly sold to consumers, i.e. individuals that are purchasing items for personal use and not for commercial or professional purposes.

53. The proposed definition of CFB breaks down the elements of “commonly sold to a consumer” as well as provides clarity on the types of MNE businesses that are “consumer-facing.”

Box 2.32. Consumer-facing business – General definition

A consumer-facing business is a business that supplies goods or services, directly or indirectly, that are of a type commonly sold to consumers, and / or licenses or otherwise exploits intangible property that is connected to the supply of such goods or services.

“Consumer” means an individual (whether or not the direct purchaser) who acquires a good or service for personal purposes, rather than for commercial or professional purposes.

A good or service is “of a type commonly” sold to consumers if the nature of the good or service is such that it is designed primarily for sale to consumers. This presupposes that the good or service is made available in ways capable of being for personal consumption (such as in portions, in sufficiently finished or usable form, or at purchase points accessible by an individual, as opposed to bulk or raw material accessible to wholesale traders or other businesses only). To be designed primarily for sale to consumers means that the MNE developed the goods or services to be regularly, repeatedly, or ordinarily supplied to consumers (whether directly or indirectly), such as by engaging in consumer market research, marketing and promoting it to consumers, using consumer / user data, or providing consumer feedback or support services (irrespective of the location in which such activities take place). Conversely, an unusual or infrequent transaction with a consumer does not affect the characterisation of the goods or service.

“Sold to” includes sale, lease, license, rent or delivery, whether directly or indirectly (e.g. through a broker, agent, intermediary or representative). This means that the product or service may still be a consumer good or service even if the contracting party to the sale is not the final consumer (such as a sale via a third party distributor, or where a franchisor has created a product but the sale is contracted by the franchisee). In other words, an MNE is in scope based on the nature of the product / service and this is not altered by the role of a third party in the distribution channel; but this does not mean that such third party itself is a consumer-facing business, as set out below. It also covers cases where the nature of the product is such that to be exploited it is licensed, rather than sold (e.g. rights to music), if it is otherwise of a type commonly licensed to consumers.

A service that is included as an ADS, being the more specific definition, is excluded from the definition of consumer-facing businesses. This avoids duplication and recognises the fact that a single classification is needed as it has implications for other aspects of Amount A, such as the appropriate nexus test.

An MNE would be regarded as being a “consumer-facing business” if the MNE is the owner of the consumer product / service and holder of the rights to the connected intangible property (including franchisors and licensors). This is the MNE whose “face” is apparent to the consumer. The term “owner” in this context does not refer to any party that at some point holds legal title to a consumer good, as there may be several legal owners in the supply chain. To be an “owner” in this context, the MNE must own the product and the related brands.

In addition, the “retailer” or other contractual counterparty of the consumer (if separate from the “owner”) would also be in scope, as they have a direct relationship to the consumer. In other words, they are perhaps the most obvious case of a “consumer-facing” business. The profit of the retailer will be different from that captured for the “owner,” thus there is no duplication caused by the inclusion of two different MNEs in the value chain. This category may also include franchisees and licensees that sell the consumer product directly, as described further below.

Other third party MNEs, such as manufacturers, wholesalers and distributors, which have no relationship with the customer – whether contractual or otherwise – and are therefore not in scope. While revenue and profitability thresholds may in any event exclude those types of intermediaries, they are already excluded as they do not have an active and sustained engagement in a market jurisdiction (beyond the mere conclusion of sales) without necessarily investing in local infrastructure and operations.

An MNE that meets the definition may, however, be (in whole or part) out-of-scope as provided in the specific exclusions, as set out further below. These exclusions are based on policy factors, such as where for specific reasons associated with a sector the residual profits are already captured in the market.

54. The test must be applied by the MNE. The test is to be applied with respect to the intended market. For example, where the MNE makes sales to a market that has unique tastes, preferences, regulatory or cultural requirements, and the product or service is not available in many other markets, it can still be a consumer-facing product.

55. For borderline cases, the early tax certainty process will be available to provide clarification and ensure consistent treatment across the Inclusive Framework.

56. The following sections discuss the application of the definition to the following particular sectors and business models: pharmaceuticals, franchising, licensing, dual use goods / services, and dual use intermediate products and components.

Pharmaceuticals

57. Broadly speaking, the pharmaceutical industry sells medications and medical devices.2 To the extent that medical devices are products of a type commonly sold to consumers, these will be in scope. This remainder of this section discusses pharmaceuticals (i.e. drugs).

58. In some jurisdictions, funding from governments and compulsory insurance schemes play a large role in purchasing pharmaceuticals, particularly for those medicines dispensed in hospitals and those available by prescription only. For example, on average, these schemes covered 58% of spending on retail pharmaceuticals in OECD countries,3 although the situation varies from one country to another.

59. It is noted that with respect to drugs, the government can regulate the prices at which drugs can be sold. It is recognised that pricing of drugs can be impacted by governments and insurance companies and are not exclusively based on consumer demand (e.g. where government directly acquires the drugs, or regulates the pricing). However, this does not change the fact that products are sold to and used by consumers (whether directly or indirectly), and that these products are generating (in some cases, e.g. when the drug is still patented) substantial profits for pharmaceutical MNEs. Therefore, even if the size of that profit may be influenced by the regulator, this regulation does not by itself take pharmaceuticals out-of-scope.

60. The analysis below contains two possible approaches for considering the extent to which pharmaceuticals would be in the scope of Amount A as a CFB:

• All drugs that are sold or administered to a consumer are in scope; or

• The scope is based on whether the drugs are sold over-the-counter (OTC) or by prescription.

61. These two options are discussed in turn.

Option 1 – All drugs in scope

62. The first approach starts from the notion that all pharmaceuticals are ultimately consumed / used by individual consumers. Taking this approach would mean that all drugs used by patients / consumers would be in scope, irrespective of whether they are prescribed and individually acquired by the consumer or are acquired in the course of seeking broader medical treatment. It would also apply irrespective of whether the products are bought by consumers themselves, or bought on their behalf by governments, hospitals or insurance companies, and irrespective of whether and to what extent the consumer paid for the drug (e.g. in some jurisdictions consumers do not pay for drugs at all, regardless whether they are OTC drugs or prescription drugs).

63. This approach would look to the more indirect engagement with the consumer, which, although in most jurisdictions does not happen by direct consumer engagement (such as by marketing), does occur through (in some cases, significant) marketing directed to the medical professionals, insurers and drug purchasing authorities which act on behalf of the consumer in prescribing, purchasing or funding the drug. Such marketing efforts and activities directed at regulatory compliance in a market (which can be very significant given the consumer safety issues associated with pharmaceuticals) could be seen as evidence of a sustained engagement with the market, representing significant expenditures for the pharmaceutical maker, and an important contributor to profitability.

64. This approach would also be consistent with the approach taken more generally of not limiting the notion of a CFB to cases where there is a direct contractual sale to the consumer, but a broader concept of how the MNE places its products in the market and engages with a consumer, including indirectly (such as the case for franchising and licensing, and through sales via third parties) – in other words, looking to the nature of the product, and not to the specific supply chain.

65. Pharmaceuticals are designed for personal use by individuals. Pharmaceutical drugs are accessible to consumers for personal consumption (e.g. being prepared in a package and with instructions for use, as opposed to being a bulk chemical compound), albeit with the support of the medical professional for safety reasons. Further, the MNE has developed the goods or services to be regularly, repeatedly, or ordinarily supplied to consumers, including by undertaking research as to consumer needs and being available for providing consumer support. Finally, it could be argued that pharmaceuticals are sold to consumers, indirectly as part of a broader delivery of a medical service noting that this could have wider implications for the CFB scope.

66. The approach may be simple to administer, as there would be no need to separate products based on the particular supply channel.

67. It would also have the benefit of being applied in a more consistent way across jurisdictions. For a range of reasons, jurisdictions make different decisions as to the designation of a given drug as prescription only or OTC. Drugs that are sold OTC in one jurisdiction can be prohibited altogether in other jurisdictions. There is no world-wide, general definition of OTC or prescription only drugs (and creating such a uniform list would be virtually impossible, both in the first instance and to update over time). Applying an inclusive approach whereby all drugs were included in scope would mean the same drugs would be in scope in every country, irrespective of a local regulatory designation. This would give each jurisdiction the possibility to receive an allocation of Amount A, irrespective of the regulatory approach taken.

68. In addition, Inclusive Framework members advocating the inclusion of prescription drugs within the definition of CFB take the view that the businesses that manufacture and distribute such products typically benefit from much higher profit margins than for OTC drugs because of the value of the legally protected intangibles often associated with prescription drugs. They believe that such high profit margins and mobile intangibles have facilitated significant base erosion and profit shifting out of market jurisdictions and into low tax jurisdictions. Proponents of including prescription drugs within the definition of CFB also consider that it would be counter-intuitive to exclude such products that have been the subject of aggressive international tax planning while including only OTC drugs, which pose a lesser risk of base erosion and profit shifting.

Option 2 – OTC drugs in scope

69. The second approach under consideration is that prescription drugs would be out-of-scope and non-prescription (OTC) drugs would be in scope, for the reasons that follow.

70. OTC drugs present features in common with other consumer goods, such as:

• They can be obtained at the customer’s choosing, as no prescription is needed;

• Advertising is permitted in many jurisdictions, and there are significant efforts to build sustained customer relationships including through targeted marketing and branding undertaken by pharmaceutical companies in respect of this type of product; and

• The acquisition of the drug is not generally a component part of a service of providing medical care.

71. This approach would exclude prescription drugs on the basis that, as compared with OTC drugs, they have specific features that do not fit as easily into the notion of a CFB. In particular:

• They are prescribed by physicians or other medical professionals based on their assessment of patients’ needs and delivered as a part of the service of health care ;

• The choice to consume a prescription drug is made by a physician, who may prescribe a different drug from the one known to the patient because it is better for them, or the one the patient knows about is not covered by their insurance;

• In most jurisdictions (other than the United States and New Zealand, for example), advertising prescription drugs to consumers is not permitted (although significant marketing may take place directly to physicians which in turn may influence the choices available to the patient); and

• The patient normally pays only a small proportion of the cost of the drug, with governments or insurance companies paying the rest.

72. These factors could be taken to suggest that the consumer relationship between the MNE producing the prescription drug and the consumer is more indirect than for other CFB, and as such, prescription drugs could be treated as out-of-scope. Applying the definition of CFB above, OTC drugs are “of a type commonly sold to consumers,” in that they are designed for individuals’ personal use, by being made available in ways capable of being for personal consumption (in individual portions, in finished form, and at purchase points accessible by an individual) and where the pharmaceutical MNE has developed the goods or services to be regularly, repeatedly, or ordinarily supplied directly to consumers (including through marketing and promotion), and the products are sold directly to consumers.

73. In addition, some Inclusive Framework members in favour of Option 2 are of the view that pharmaceutical drugs in general are acquired or used by consumers chiefly for their inherent characteristics.

74. However, the administration of this approach will be further considered, including based on further discussion with stakeholders. Given the different definitions of OTC and prescription drugs that vary from one jurisdiction to another, there could be significant difficulty for an MNE to apply the scope rule, and the Amount A allocation to a market will be smaller for those members that take a stricter regulatory approach to prescription drugs. In addition, some drugs are sold as both OTC drugs and as prescription drugs in the same country, for example where the dose is higher or when a specific drug is prescribed in combination with other drugs. This could lead to administrative difficulties to track and trace whether a particular drug was in scope in a given case and could also involve significant administrative costs for tax authorities and compliance costs for taxpayers in order to comply with the second approach. However, members in favour of this approach take the view that the compliance burden could be addressed, for example, by following the classification of a particular drug as prescription or OTC based on the approach in the majority of Inclusive Framework members or whether a substantial part of the revenue is from prescriptions.

Franchising

75. The proposed scope of Amount A would extend not only to businesses that sell goods and services directly to consumers, but also to those that sell consumer products indirectly through third-party resellers or intermediaries. In this sense, facing the consumer is broader than strictly contracting with the consumer, and as such, goods and services can be in scope irrespective of the particular distribution channel or selling agent.

76. For this reason, franchise models and licensing arrangements in respect of consumer goods and services are included within the scope. This means that the basic concept of CFB applies, even where the income is earned from franchising or licensing rather than direct sale to the consumer.

77. This is reflected in the definition of CFB above, which includes as a CFB one that licenses or otherwise exploits intangible property that is connected to the supply of goods or services that are otherwise of a type commonly sold to consumers. As franchising involves earning a return from making intangible property available to the franchisee (as described in the franchising models below), this language is intended to capture franchising arrangements. It does not mean that to be in scope, the consumer must be directly acquiring licensing or franchising rights. For example, where an international chain of restaurants operates through a franchise model, the local restaurant franchisee would be in scope as being a directly consumer-facing business. The MNE that is responsible for the franchising and earning franchise fees would also be in scope as a CFB, even though their revenue model is different from the local restaurant (or if the MNE also owns the local restaurant, that revenue would also be in scope as well as other franchising fees earned).

78. By way of background, there are broadly two franchising models: business format franchise and product distribution franchise. Product distribution franchising can be further divided into three sub-categories: manufacturer-retailer; wholesaler-retailer; and manufacturer-wholesaler. These categorisations are not mutually exclusive and some franchise arrangements may be covered by more than one category. However, the categorisations are helpful to understand the range of arrangements that are in scope. The analysis below describes a range of different formats for how franchising may take place. The question of scope does not focus on the legal form of commercial arrangements, but whether the MNE is supplying goods or services commonly sold to consumers, or licensing or otherwise exploiting intangible property that is connected to such goods or services. If the rights forming the franchise arrangement are connected to an underlying product or service of a type commonly sold to consumers as per the definition of CFB above, then the franchisor and franchisee will each be a CFB (provided that they are otherwise subject to Amount A, including with respect to the gross revenue threshold (see below) and profitability thresholds (see Chapter 6).

Business format franchising model

79. In the business format franchising model, the franchisor provides the ready-made business model, while the franchisee is responsible for day-to-day operations in accordance with the franchisor’s operational directions. This model is commonly used in the fast food and restaurant sector.

80. One of the key features of a business format franchising model is the level of authority the franchisor can exercise over the franchisee in respect of how the franchisee operates its business. In addition, the franchisee generally has continued access to intangibles owned by the franchisor (know-how, trademarks, etc.) and can obtain advice and assistance from the franchisor. The level of authority the franchisor exercises over the franchisee combined with the franchisee’s right to access the franchisor’s intangibles, assistance and advice ultimately provides a platform for the franchisor to engage with the consumer. Although there is no direct legal agreement between the consumer and the franchisor, there is significant engagement between them. Business model franchising in respect of CFB is therefore distinguished from other types of business to business arrangements where the supplying business does not have any engagement with the end consumer.

81. In this model, the franchisor’s revenue can include technical fees (for training), contributions to advertising, upfront fees, renewal fees and ‘franchise fees’ (including royalties) which are paid periodically as a percentage of total sales generated by the franchisee.4

82. Applying the general definition above, where the franchised rights are connected to a good or service commonly sold to consumers (e.g. burgers), the franchisor is in scope as the owner of the brand being monetised (via the intangible property made available to the franchisee), and the franchisee is in scope as the seller of the consumer product.

Product distribution franchising

83. Product distribution franchising is typically divided into three sub-categories:

• Manufacturer-retailer franchising where a franchisee sells the franchisor’s products directly to consumers. For example, manufacturer-retailer franchising is often used to sell new cars. Depending on the legal agreement, the franchisee may be permitted to sell competing products. Under these arrangements, the franchisee sells and operates under its own name although the franchisor’s trademarks are generally displayed. The franchisee sells the products using its own business systems and methods but the franchisor may require the franchisee’s store or showroom to be fitted-out to a particular specification. No special training on a business system will be required but the franchisee must be familiar with the product range, its capabilities and any follow-up services that are available, and training in this respect will be provided by the franchisor, in some cases on a paid basis. The franchisor may charge an initial fee for becoming a franchisee and may also charge on-going fees (often calculated as a percentage of sales) and contributions to advertising. Separately, the franchisee will purchase stock from the franchisor and may be required to place minimum orders. The franchisor may provide recommended retail prices to the franchisee.

• Wholesaler-retailer franchising, where the retailer as franchisee purchases products for sale to consumers from a franchisor wholesaler. Depending on the legal agreement, the franchisee may be permitted to sell competing and / or complementary products. Wholesaler-retailer franchising is often used as a model to sell car fuel and is also common where a cooperative of franchisee retailers form a wholesaling company through which they are contractually obliged to purchase (for example, hardware and automotive product stores). The franchisor may charge an initial fee for becoming a franchisee (in a cooperative model, this may be a subscription for shares or membership interests) and may also charge on-going fees (often calculated as a percentage of sales) and contributions to advertising. Separately, the franchisee will purchase stock from the franchisor and may be required to place minimum orders. The franchisor may provide recommended retail prices to the franchisee.

• Manufacturer-wholesaler franchising, where the franchisor manufactures an intermediary product that the franchisee assembles into the end product and distributes to customers.5 For example, many soft drinks businesses use manufacturer-wholesaler franchising, where franchisee bottlers mix, package and distribute the beverage. The franchisee will be required to assemble the end product in compliance with the franchisor’s prescriptive procedures and often will acquire equipment from the franchisor to facilitate assembly. Depending on the legal agreement, the franchisee may contract exclusively with the franchisor. The franchisor will manage the overall brand strategy and often will manage key global customer relationships. In some cases, the franchisee might pay the franchisor a franchise fee but in other cases the price paid by the franchisees for the intermediary product can be determined by reference to a number of factors, including, but not limited to, the franchisee’s revenue and pricing. Under the latter type of arrangement, different franchisees pay different per unit prices for the manufactured product and the franchisor’s return will fluctuate depending on the franchisee’s revenue.

84. Similar to business model franchising, under each of the product distribution franchising sub-categories the franchisor exercises a relatively high level of authority over the franchisee in respect of how the franchisee deals with the franchisor’s product. In addition, the franchisee generally has on-going access to specified intangibles owned by the franchisor (knowhow, trademarks, etc.) and can obtain advice and assistance from the franchisor (whether on a paid basis or otherwise), for which the franchisor earns a return. Similar to the business model franchise, the product distribution franchising provides a platform for the franchisor to engage with the consumer. As is the case for business format franchising, provided the franchised rights are connected to a good or service commonly sold to consumers (such as cars and beverages), the franchisor is in scope as the owner of the brand being monetised (via the intangible property made available to the franchisee). The franchisee may also be in scope if it sells directly to consumers, goods or services that are of a type commonly sold to consumers.

85. Further work is underway more generally with respect to how Amount A applies to franchising, such as how to apply the nexus test, what revenues should be regarded as in scope, interaction with withholding taxes, and so forth.6

Licensing

86. Although in some cases, licensing shares some of the characteristics of franchising, licensing generally covers a much broader spectrum of commercial agreements. On one hand, certain intangible property can itself be of a type commonly sold to consumers (such as a music) and which by its intangible nature must be licensed to be exploited rather than strictly “sold.” This is noted in the general definition above, in connection with the meaning of “sold.”

87. There is a second type of category, where an MNE has made available its intangible property to a licensee, to be incorporated in another consumer good or service. For example, providing the rights to a cartoon character or logo to appear on clothing or in a game. The proposal is that this will only apply to an MNE that otherwise supplies (or has in the past supplied) a consumer-facing good or service (such as entertainment or clothing), and makes the rights to intangible property connected with that good or service (such as a consumer brand) available through a licensing arrangement (although further consideration on marginal cases may be required, such as where a non-CFB (e.g. a maker of an industrial product) earns returns from licensing its brand to a CFB such as clothing).7 In these cases, the intangible is “connected” to the supply of the consumer-facing goods or services otherwise supplied by the licensor, and allowing it to be out-of-scope would distort the scope based on legal form rather than according to the relationship created with the consumer. This does not mean that to be in scope, the licensee must be a CFB. For example, the licensees may be providing ADS (such as digital content or online gaming in which the licensed cartoon character appears). The inquiry is whether the licensor has a consumer-facing product that is being licensed; if so, the licensor is in scope of CFB.

88. There are a number of commercial arrangements by which the MNE licensing the intangible property may derive revenue. At one end of the spectrum are licence agreements that effectively replicate franchise agreements, where the licensor licenses intangibles to the licensee in return for a royalty that is based on the level of sales achieved by the licensee. Trademarks and brand names are often licensed in this way to sell consumer goods. Typically, the licensor will retain strict control over how the licensee uses the licensed mark or name and the quality of products produced by the licensee using the licensed intangibles. For example, a number of luxury brands license their trademarks and logos to perfume and eyewear manufacturers. The licensor will generally impose strict quality controls on the licensees, may obtain rights to audit the licensee’s use of the licensed materials and will often impose restrictions on how the intangibles may be used. Royalties for such licenses are typically commensurate with sales or returns earned by the licensee.

89. At the other end of the spectrum are rights that are licensed to licensees for a fixed fee. In those circumstances, the licensor has no entitlement to share in the revenues earned by the licensee and is entitled to a fixed fee regardless of how successful the licensee is in commercialising the licensed material. Film rights and television rights may be licensed in this way for a fixed period in a particular territory or territories. The licensor will typically have limited rights to control how the licensee operates in respect of the licensed material (although there will be strict controls preventing the licensee from altering that material).

90. In other cases, a licensee may agree to pay a royalty to the licensor based on revenues earned by the licensee. Although the licensor may not have significant authority to prescribe how the licensee operates in respect of the licensed material, the licensor is incentivised to ensure success of the licensee in commercialising the material (as this enhances the licensor’s return). In those cases, it would not be unusual for the licensor to market the licensed material in the jurisdictions covered by the licence. Music, films or other content provided to a platform are often licensed in this way, with royalties tied to the number of views / streams / downloads.

91. Although the degree of control over the licensee’s operations and the revenue models may vary, the principle is if the licensor has made available intangible property that is connected to a CFB, then it should be in scope even if it is leveraging the licensee to have that distributed rather than doing this within the same MNE group. If the licensed product / service meets the definition of a consumer-facing good or service, then the licensing arrangement is in scope.

92. Therefore, as with franchising, although licensing arrangements involve a different set of legal arrangements in how the goods are made available to the consumer, these methods of exploiting a consumer-facing good or service through the controls over the licensed item being exploited by the licensee would be in scope of CFB. Further consideration will be given to how this approach interacts with other elements of Amount A, such as withholding tax on royalty arrangements.8

Dual use finished goods / services

93. Once a good or service meets the test of being of a type commonly sold to consumers, all sales of goods or services of that type of product will be entirely within scope (subject to the analysis below on dual use intermediate products / components). This remains true, even if the product is sold to a business customer. These are “dual use” products that can be sold to both consumers and businesses. For example, passenger cars, personal computers and some medical products (such as blood pressure monitors) fall in this category. Applying the general definition above, an unusual or infrequent transaction would not re-characterise the goods or service, such as where a product otherwise intended to be B2B is unusually acquired by an individual.

94. There are three reasons for taking this approach.

95. First, at a general level, it is difficult given the nature of these dual use goods to say which parts are consumer-facing and which are business-facing. In the case of the passenger car or personal computer, there is nothing conceptually to distinguish one car or computer from another because it is already a finished product designed for consumers. The design features of the car are the same, whether sold to a consumer or to a business.

96. A more practical concern is that tracking revenue according to the nature of the purchaser for every given transaction may be very challenging – and even impossible – for businesses to do. Apart from the administrative burden of recording the consumer purchasers versus business purchasers, that very distinction is not necessarily always apparent. The lines between personal and professional use are becoming less distinct, with a changing economy that allows more business activity to take place with personal items, such as working from home on a personal laptop or participating in the sharing economy. In addition, these goods are marketed to consumers, and ultimately used by an individual. This may influence the individual’s future choices and therefore build a relationship with that person, even if they originally are not the purchaser (e.g. where a person uses a laptop provided by their employer, becomes familiar with it, and therefore chooses that same brand when it comes time to purchase a laptop for personal use). It would therefore not always be possible on a practical level to distinguish the line of business according to who makes the purchase on a given occasion.

97. Third, if dual use goods were only in scope to the extent of the amount of consumer purchases in practice, it could give rise to a need for look-through or anti-abuse rules to address cases where sales were intentionally not made directly to consumers, which would add complexity for business and tax administrations.

98. For these reasons, where the good is of a type commonly sold to consumers, the full amount of sales that are of this type of product / service would be in scope, irrespective of whether a given purchaser is a business or an individual.

Dual use intermediate products and components

99. The Outline notes that businesses selling intermediate products and components that are incorporated into a finished product sold to consumers would be out-of-scope. Examples include wood pulp, steel rods, bulk fabric, and electrical parts within a laptop or phone. These do not meet the test of a CFB – neither when considering the nature of the MNE’s relationship with the consumer (which will be one of an intermediate manufacturer of component parts), or the type of product itself (which will not be of a type commonly sold to consumers).

100. However, some intermediate products / components are dual use. That is that they are of a type that is also designed for use by consumers. If such intermediate products / components meet the general definition above, they would be in scope. Possible examples include a car tyre, some replacement parts, batteries for consumer products, and some types of medical products such as bandages. However, they would be in scope only to the extent of the sales to consumers (unlike the analysis above for dual use finished products, which are entirely in scope).

101. This approach would give the most accurate result for two reasons.

102. First, the nature of an intermediate product / component has differences whether being sold to businesses or to consumers. For example, the packaging, pricing and distribution channels will be different (e.g. where packaging for consumers contains branding and instructions for consumer use, where packaging is in smaller individual lots rather than in bulk, or where price points are higher for retail sale, or where there is a separate distribution channel directed to consumer sales). These features make it inherently a consumer product and not a component; but the sales to business retain features of a component part.

103. Furthermore, including only the sales actually made to consumers is a much simpler compliance approach. The alternative would be to include in scope even the sale of component parts to businesses on the basis that sometimes they can be directly acquired by a consumer. This would be a significant compliance burden, and would require difficult revenue sourcing rules to be developed. This is because a component part when sold to a business would then need to be traced through what may be a complex supply chain. For example, in the case of a car tyre, this may be shipped to the location of the factory where all tyres are stored, to the factory where the cars are assembled and further manufactured, and through the distribution channel to locate the place where the tyre is ultimately sold to a consumer as part of the car. Instead, the obligation would only be to source those sales made to consumers, which should be feasible because of the different nature of the product when sold to consumers, as noted above.

104. However, the compliance burdens and benefits associated with this approach are being further considered in order to inform the approach, as well as whether simplification measures should be included and whether further guidance can be provided on delineating dual use finished goods from dual use intermediate products / component. It is noted that some Inclusive Framework members would, in the interests of addressing the compliance costs, favour removing all component / intermediate products from scope (even those that are otherwise consumer products); while one member favours including all component / intermediate products that are dual use, irrespective of sale to business or consumers in scope.

2.2.3. Exclusions and carve-outs

105. This section sets out the types of activities that are proposed to be specifically excluded from Amount A. These are: (i) natural resources; (ii) financial services (iii) construction, sale and leasing of residential property; and (iv) international airline and shipping businesses.

106. Many of the goods and services sold in these sectors would already be out-of-scope of Amount A because they are neither consumer products nor ADS. For clarity, specific exclusions are set out for the whole or, as appropriate, parts of these sectors.

107. This section sets out the rationale for exclusions by reference to the ADS and CFB criteria indicated in the above sections of this chapter, and delineates how far the exclusion applies to each sector.

108. The scope exclusions would apply on a segment basis, which means that for an MNE group with multiple business lines, some may fall within a scope exclusion, and some may not.

Natural resources

109. The term “natural resources” encompasses non-renewable extractives (such as petroleum and minerals), renewables (such as agricultural, fishery and forestry products) and renewable energy products and similar energy products (such as biofuels, biogas, green hydrogen).

110. As stated in the Outline, extractive industries9 and other producers and sellers of raw materials and commodities, including commodity traders, will not be within the consumer-facing business definition, even if those materials and commodities are incorporated further down the supply chain into consumer products. Taxes on profits from the extraction of a nation’s natural resources can be considered to be part of the price paid by the exploiting company for those national assets, a price which is properly paid to the resource owner.

111. Natural resources, such as agricultural and forestry products, are generally generic goods which are sold, and whose price is determined, on the basis of their inherent characteristics, rather than on other factors such as marketing. Where premium prices exist (which could lead to residual profits), they tend to be a function of aggregate supply and demand dynamics for the commodity. As natural products of the earth, all are closely linked with the place of production and quality is highly dependent on location-specific factors. Due to local environmental impacts including the sustainability of the resource itself, production is usually regulated in the local jurisdiction.

112. Other policy considerations may also support the conclusion that these generic and undifferentiated natural resource products should be out-of-scope of Amount A. These include the close connection with the place where the resources are found and transformed (reflected in current tax treaty treatment of immovable property); the capital intensity of their production; and the high volume of the products used as raw materials in other industries.

113. The vast majority of natural resources would already be outside of the scope under the general CFB and ADS definitions. These materials are not of a type commonly sold to consumers, nor are they automated services provided digitally.10

114. At the other end of the production chain are products that incorporate natural resources, such as jewellery and chocolates, which would be within the general definition of CFB.

115. Between these two points are borderline cases, which are discussed below.

Non-renewable resources (extractives)11

116. Some final products derived from hydrocarbons are sold to consumers. Products like natural gas, liquefied petroleum gas and kerosene have many industrial uses but also may be sold to consumers, e.g. for heating and cooking. Similarly, fuels such as petrol (gasoline), diesel, and aviation gasoline are used by both industry and consumers for transportation purposes.

117. Despite being commonly sold to consumers, however, they have features which distinguish them from other types of consumer products and may bring them within the scope of the exclusion for natural resources and commodities. While these products have undergone a degree of processing, the processing is largely standardised, meaning that the products may be considered generic, undifferentiated and sold on the basis of their inherent characteristics. Indeed, given their standardised characteristics, some of these products such as natural gas and petrol are sold on commodity exchanges. As such, to a consumer, even if sometimes these products are sold under a brand name, there is generally little that distinguishes one MNE’s product from another’s. This seems to imply that there is limited scope for an MNE to build consumer relationships and interactions and create consumer-facing intangibles to increase value and sales. In determining the importance of firm-specific intangibles, it could be relevant to consider such factors as the degree of retail competition, whether there are material prices differences across supplier, the size of retail profit margins, etc. Fuel retailing also relies on extensive local infrastructure, meaning profits are typically realised locally. However, given questions about the role of branding, advertising and level of processing, in such cases as automobile lubricants, petrol or diesel, further work is required.

118. Likewise, precious gemstones (diamonds, emeralds, non-mineral crystals such as opals) are sometimes sold to consumers. Cut and polished gemstones could be seen as “dual use intermediate products / components” in that they are most often sold to jewellery manufacturers but are sometimes sold directly to consumers. Consideration is being given to whether they ought to be treated as falling outside the natural resources exclusion and instead within the general rules for CFB (for example, having regard to the marketing effort sometimes associated with consumer sales of gemstones, including online sales channels).

Renewable resources and renewable energy products

119. Most bulk agricultural, fishery and forestry products do not meet the CFB definition.12 Following further processing and packaging, however, some agricultural, fishery and forestry products become consumer products. However, if the product remains generally generic, undifferentiated and sold on the basis of its inherent characteristics, there is a case that it should be excluded from the scope of Amount A. While work continues to delineate more precisely how far the exclusion extends, further work will be required on the role of relevant factors such as the degree of processing, packaging and branding.13

120. Renewable energy products also raise questions about the scope of the resources exclusion. Electricity from water, sun and wind power exploits a renewable natural resource. Even when sold to consumers, the product is homogeneous and purchased for its inherent characteristics. As discussed below under infrastructure services, electricity distribution is already closely tied physically to the market due to the physical infrastructure used. Biogas is very similar to natural gas except that it is produced from non-fossil sources. Renewable transportation fuels (biofuels like ethanol) are substitutes for petroleum products like gasoline and diesel. While the origin is distinct, they have similar characteristics, including retail branding, suggesting a case for parallel treatment.

121. Further work will be undertaken, taking into account the above considerations, to provide greater clarity defining the exclusion for generic and undifferentiated natural resource products and commodities.

Financial services (FS)

122. FS business comprises banking, insurance and asset management. As the issues raised by each of these three sectors in relation to the new taxing right (particularly in relation to CFB) are different, the discussion below considers each FS sector separately in relation to CFB. However, in relation to ADS, the position is more uniform across the three sectors and accordingly is dealt with immediately below. No separate regime is envisaged for finance companies.

123. Digital functionality is widely used in each of the banking, insurance and asset management sectors. Nevertheless, it is generally used to automate what people used to do (and often still do), whether in conducting conventional FS business or in risk-management functions. As such, human intervention and judgement is normally a feature of the use of digital functionality in FS business. For this reason, and subject to certain detailed work that is outlined below, FS business should not generally involve business of the sort that is properly regarded as ADS business for the purposes of the new taxing right.

124. With regard to CFB, very significant parts of FS business are not consumer-facing. However, there are also significant parts of the FS business that involve CFB. Though the analysis of CFB raises different issues across the three sectors, the central rationale for the exclusion of FS business from the new taxing right is common to each of the three sectors. That rationale stems from the highly-regulated nature of FS business. However, it should be emphasised that this central rationale is not premised on the mere fact of regulation but rather is based on the effects of that regulation. More specifically, the regulations governing the relevant business in each of these three sectors, that may involve CFB activity, generally require that appropriately capitalised entities are maintained in each market jurisdiction to carry on business in the market concerned. Due to this factor, the profits from CFB activities that arise in a particular market jurisdiction will generally be taxed in that market location with the result that there is no further need for any Amount A re-allocation. Additional factors relevant to the analysis of banking, insurance and asset management business are considered below.

Banking Business

125. In banking, there is substantial local regulation, which materially affects the local booking and taxation of banks’ consumer business. Accordingly, the central rationale for the exclusion for FS business is clearly relevant to the banking sector. Almost without exception, local regulators will implement banking rules based on the Basel II & III framework with amendments to suit its local environment. Because consumer-facing banks need to have local approval and a local licence, they generally provide their services only to local customers. This results in a taxable presence in the locations where they face their customers. Given the dependence of an economy on the financial sector, most developing economies have also implemented regulations such as exchange controls, which add to the existing compliance burden of conducting cross-border financial transactions. Further, banks in many jurisdictions are the primary transmission mechanism for monetary policy through their borrowing and depository transactions with their central banks, and must abide by the requirements of those central banks.

126. Certain additional technical and practical factors further support the case for an exclusion of the banking sector from the new taxing right. One important consideration is that the use of a profit to sales ratio (which would normally be used in non-FS sectors in the calculation of Amount A) would likely prove unworkable in the banking sector. Operating profit margin (operating earnings/revenue), or profits before taxes, is not an appropriate basis for comparing banks with other industries because: (i) interest expense and cost of inventory, which in effect is equivalent to the cost of goods sold for banks, is not included in the definition of operating income; and (ii) banks report trading revenue net of trading costs, and interest revenue net of interest expense. While reporting revenues net of trading and interest costs artificially decreases the revenue generated by a financial institution, grossing up the revenues by trading costs would cause a typical FS group never to have an operating margin above 1%. The volumes, in financial terms, of bank trading operations are typically in the USD trillions and, if sales were grossed up, (i.e. not netted with the cost of inventory) as they are in other industries, the denominator for computing operating profits would be in the trillions of dollars. These differences materially distort bank operating margins by increasing the numerator and decreasing the denominator in the calculation of the operating profit margin. Failure to accurately account for these costs of doing business in both the numerator and denominator of the calculation would result in a disproportionate taxation burden on the FS industry. Corrected for these amounts, either through modifying the current proposal or developing an alternative metric for the FS sector, banks’ profit margins would be dramatically lower.

127. The nature of bank regulation provides a further “positive” and a “negative” reason to support the exclusion. From a positive perspective, where an industry, such as the banking sector, is subject to robust non-tax governmental regulations consistent with recognised industry standards (e.g. Basel requirements), due regard should be given to the constraints those regulations impose upon them. As a result of such regulations, profits from retail banking operations often arise in the jurisdiction in which those operations are conducted. And the consequences of regulation (rather than the mere fact of regulation itself) mean that those jurisdictions will often be the ones that tax the resulting profit. From a negative perspective, there is potential for a clash between the way that Amount A works and such non-tax regulatory regimes because the regulatory objectives may seek to protect the local regulated entities, e.g. from the insolvency of affiliates. This might mean that regulators could refuse to accept the existence of intercompany payments made under Pillar One, or require that an affiliate reimburse the local regulated bank, or require that additional capital be held. This would result in an undermining of the agreed regulatory regime and potentially also in inefficient allocations of capital and liquidity, resulting in increased costs for ordinary financial transactions.

128. Consideration of the application of an exclusion in the context of the banking sector has led to the review of a number of additional matters, specifically the status of Fintech and private banking business and the practical operation of EU passporting.

129. Fintech is broadly finance enabled or provided via new technologies, meaning the application of new technologies to FS. A Fintech business may be developed either by banks or dedicated Fintech firms. Most of the Fintech firms are intermediaries between banks and customers. Some Fintech firms are lightly regulated or may not be subject to any regulatory regime under their national law. However, Fintech firms that wish to provide regulated FS have two options: (i) they can conclude a partnership with a bank, or (ii) they can request a banking licence, in which case they are subject to the same regulations that apply to banks, and thus to the same obligations such as minimum regulatory capital, reporting requirements, cross-border activities limitations or infrastructure obligations that require them to have a physical presence in the local market jurisdiction. Banking regulation is mandatory for four types of Fintech activities: e-money, payment, credit, and full banking licences. Other products and services supporting financial activities provided by Fintechs in some areas (hardware, software, middleware) remain excluded from banking regulation. That is because these Fintechs are only providing technology to the financial sector to perform activities that otherwise would be performed by individuals. Given that regulated Fintech firms have a specific status, similar to that of the regulated banking sector, their exclusion appears consistent with the policy rationale outlined for FS. These firms should therefore receive a similar treatment to regulated banks. However, as the rules in many jurisdictions governing Fintech are at an early stage of development, the issue will be explored further. The logic of the approach would not be applicable to unregulated Fintech activities.

130. Private banking/wealth management (PB) business is an activity conducted within regulated banking groups that falls on the borderline between retail banking and institutional banking, given that clients include wealthy individuals as well as charities, family offices, trusts, etc. PB is often conducted through entities subject to banking regulation both globally and locally. However, PB may in part be conducted through regional hubs, meaning that the policy rationale discussed above may not be fully applicable. Nonetheless, various factors support the application of the exclusion to PB. First, without a locally regulated presence, a bank generally cannot have an active and sustained participation in a local market; publicly offer services (including digital services); or widely advertise or solicit business. Second, many PB clients are business owners, and PB services may also be provided to the business rather than the individual. Third, in the event that PB business were to be included in the scope of Amount A, there would be certain practical difficulties in identifying and isolating private banking profits. These include: (i) the practical difficulties in segmenting the PB business as a substantial portion of customers will be charities, foundation, etc. Banks would also need to distinguish between services provided to individual clients, and services provided to businesses owned by their clients; (ii) segmenting the in-scope income streams given that the services and products provided to PB customers involve a variety of activities and income that may be booked across the different divisions/entities within a banking group, and isolating and identifying these highly fragmented income streams would be difficult; and (iii) identifying the market jurisdiction as many contractual relationships (especially institutional client relationships) are centralised. On the basis of the above discussion, it is proposed to include PB business within the exclusion for FS.

131. Consideration has also been given to EU passporting (which enables FS firms that are authorised in one EU member state to trade freely in other member states) to assess whether its practical operation might be in conflict with the central policy rationale supporting the exclusion for FS business. However, due to various regulatory and commercial factors, FS businesses generally enter into CFB transactions in local markets through entities that create a physical presence in those markets and thus that are taxed locally on those profits. Banks operating in the EU thus do in practice have a taxable presence in individual EU Member States for the provision of customer-facing activities. It is therefore considered that the existence of EU passporting does not conflict with the policy rationale discussed above.

Insurance Business

132. The analysis of CFB in the context of the insurance sector is similar in many respects to the position in the banking sector. In the insurance sector, there is, as in the banking sector, extensive regulation of consumer-facing business (retail insurance business) leading to the same effect, namely that the profits from CFB activities that arise in a particular market jurisdiction will generally be taxed in that market location, again with the result that there is no further need for any Amount A re-allocation. Regulation of the industry – especially conduct regulation – makes it very unlikely that consumers could purchase insurance without the insurer having a local presence. In many jurisdictions it is illegal to offer it. The regulatory requirements therefore mean that income from retail insurance business is generated and taxed in the market jurisdiction from which it is derived. The prevalence of insurance regulation, dating back decades, is evidenced in studies and surveys of regulation of insurance regulation in different jurisdictions and regions published by the OECD and other organisations14. Since the 1990s, in a similar manner to the Basel Committee in banking, insurance supervisors have developed global standards for adoption into national frameworks. Despite global recognition of the importance of regulating the insurance sector, regulation continues to be executed primarily within the local jurisdictions. For example, jurisdictions do not, generally, recognise licensing outside their own borders, and particularly in matters such as access to domestic markets. Insurance companies operating in Latin America, India and China, for example, are subject to both prudential and conduct regulation.

133. The “positive” and “negative” factors that stem from the highly regulated nature of the business (and that are discussed above in the context of banking business) are also relevant in the case of insurance business. There are also concerns that the measurement of profits in the insurance sector is not comparable to the approach outside the financial sector. Insurers measure income and costs differently than other industries so traditional profit measurements might inaccurately result in excess profits that do not in reality exist. Most industries incur costs such as labour, raw materials, etc. at an early stage in the business cycle, with the corresponding revenues generated later in the process. The opposite is true of insurance: insurers collect premiums upfront but incur unpredictable costs later – sometimes much later – when they pay claims. The effect of these types of losses on the insurance industry is unique. The insurance industry’s role is to assume risk over many years, with an uncertain realisation and timing of the insured event. This exposes the industry to volatile profits and losses. Premium rates change over the cycle depending on the availability of capital. During a soft market (when capital is plentiful), competition reduces premium rates. But as the market hardens (when capital becomes scarce, typically after a major catastrophe), premiums rise. This creates a multi-year business cycle unique to the insurance industry. Current year profits are often based on insurance reserve estimates that reflect losses that may (or may not) occur and are based on complex actuarial modelling techniques. The ebb and flow of the insurance cycle makes the determination of normal returns for the industry difficult. As noted above, for the insurance industry, a local market presence is generally required to sell retail insurance products. This is for the regulatory reasons already covered. This means the relevant local entity will already be taxed in the relevant local market. Together with the some of the technical challenges noted in computing insurance industry profits means it makes little practical sense to include the insurance industry in the CFB definition.

134. Consideration has also been given to the impact of EU passporting in the insurance sector with the same result that is discussed above for the banking sector. There is therefore no conflict with the rationale for the exclusion of the insurance sector.

Asset Management Business

135. The asset management sector contains three main types of participants: fund vehicles, financial intermediaries, and investment managers. This section explains why most (but not all) jurisdictions believe that it is appropriate to exclude the sector from the scope of Amount A.

136. Funds are not active businesses. Their purpose is to pool investor capital, which is then invested in accordance with the fund’s investment strategy. The tax neutral and passive status of funds is widely recognised under jurisdictions’ domestic law, international tax principles and the OECD’s own guidance. Funds are therefore considered to be outside the scope of CFB.

137. Under the predominant retail industry operating model, the indirect model, financial intermediaries comprise third-party banks, financial advisors, or similar businesses that advise investors and offer them investment products. Although interactions between financial intermediaries and consumers may on first principles be within the scope of CFB, they are, as for retail banking, subject to local country regulation. Customer interactions would therefore normally occur locally in the market jurisdiction, and the associated profits would be realised locally. Under the less common direct model, an investment manager distributes investment products through an affiliate that is a broker or similarly regulated entity. But under either model, regulatory requirements focus on investor protection, suitability and transparency and result in a required local presence for financial intermediary business (in so far as it concerns retail investors). This means such business is, owing to its broad-based regulatory regime, in the same position as retail banking and retail insurance business. The activities of financial intermediaries are therefore covered by the central policy rationale for the exclusion of financial sector business.

138. In the context of Amount A, investment management services raise more complex questions. Investment managers normally deal with financial intermediaries rather than directly with underlying individual investors in both the indirect and the direct business model. In the direct model, the financial intermediary that has the direct contact with the investor would be the investment manager’s affiliated broker, broker-dealer or similarly regulated entity, which serves the role of the financial intermediary. Additionally, the investment manager is normally engaged by the fund or investment vehicle, rather than investors themselves. The engagement of an investment manager by the fund vehicle generally occurs before any individuals invest in the fund. Whether the business is a CFB therefore depends on how interactions between investment managers and intermediaries are construed and in particular whether they are considered to be direct business-to-business transactions or some form of “indirect” service to consumers, through the medium of the financial intermediary.

139. Based on the analysis of the role of the investment manager, on balance the investment manager’s services are properly seen as a component of the service the financial intermediary offers its clients. Businesses selling intermediate products and components that are not commonly sold to consumers (but which are incorporated into a finished product sold to consumers) would be out-of-scope of Amount A. Where intermediate products and services are commonly sold to customers they may be within scope (e.g. where they are branded). But this does not seem to apply for investment managers, as the investment management services are incorporated into the service the intermediary offers, which is then provided, as a component, to consumers. On this basis, they would not be in scope of CFB. The conclusion that the services of investment managers are an intermediate product is supported by the fact that the financial intermediary makes a meaningful difference to the nature and value of the investment management service as a result of (i) the important role of the intermediary and (ii) the fact that investment manager interacts directly (from a legal and practical standpoint) with the fund, similar to other service providers (e.g. custodian, transfer agent, fund administrator, etc.). These matters are discussed further below. On (i), the financial intermediary performs important functions for the consumer, providing substantial economic value. It uses its knowledge of investment opportunities to provide the consumer with bespoke advice about the best products to buy (typically managed by unrelated investment managers) based on the consumer’s needs and risk profile. The intermediary is remunerated by either asset-based or transaction fees, separate from the fee earned by the investment manager. These fees can be substantial. On (ii), the investment manager provides its services (legally and factually) to the funds themselves, not to the investors, and its fees are paid by the funds. A bank or other service provider providing custodial, transfer agency or administrative services to a fund works in the same way. The comparison is relevant as there seems to be no basis to think that these custodial, transfer agency or administrative services amount to services to consumers, given how indirect the benefit is to them. The funds to which the investment managers provide services have a fiduciary duty to act in the best interest of unitholders. Those services are guided by the fund’s investment strategy and objectives. The services have no regard to investors’ objectives, nor to the inflow or outflow of investors, nor to investors’ preferences about timing, diversification, liquidity, currency or anything else. These preferences are managed by financial intermediaries, as described above.

140. In addition, there are in any event practical policy reasons to exclude investment managers from Amount A. These reasons are significant in this case as the conclusion that the investment manager’s services are a component of the service the financial intermediary offers its clients may be debated. These practical policy reasons concern access to relevant information and the practicability of the exclusion for FS activities. On information access, in the indirect model, intermediaries often cannot disclose information on underlying investors to an investment manager because of data privacy and regulatory restrictions and the nature of proprietary customer lists. The investment manager therefore often cannot access the information needed to identify market jurisdictions. In addition, there are the practical problems caused by the constant change in the identity of investors in a fund, and the fact that the investor base will often be in a state of constant flux as regards the percentage of retail investors (consumers) and other professional and institutional investors. The process of tracing the identity and balance of the investor mix will be further complicated by the fact that funds commonly invest in other funds. Further, if there were no exclusion for investment management activity, banks and insurance companies (assuming they are out-of-scope) could face complex issues. This is because they also conduct investment management business, and would therefore need to isolate their profits from their investment management services. Another concern would be the extensive practical difficulties in establishing and enforcing a definition of in-scope investment management business. The fungible nature of financial products means there can be economic equivalence between investment management advice and the embedded features and performance of a structured product offered by a bank, like an indexed certificate of deposit or a swap, and a similar investment offered in an insurance wrapper, such as an annuity, by an insurance company. For all these reasons, it is concluded that the activities of investment managers do, and should, fall outside the scope of CFB. However, those jurisdictions which do not support exclusion for asset management sector from the scope of Amount A, believe that the asset management sector is very lightly regulated, which (unlike retail banking) may not ensure that the major part of residual profit is captured in market jurisdictions.

Infrastructure and general construction business

141. Infrastructure construction is the provision of physical structures and facilities such as buildings, roads, bridges, tunnels, power plants, and airports. Infrastructure projects can be broadly categorised into B2G and B2B projects, where the government or a commercial enterprise acquires the infrastructure for use by the public (e.g. a road) or by the business (e.g. a port).

142. General construction can be broken down into commercial and residential. Commercial real estate includes the provision of office buildings, hotels, factories and warehouses. These are developed and sold or leased for use by other businesses.

143. Neither infrastructure construction nor general commercial construction is within the scope of ADS or CFB based on first principles. They would not fall within the definition of ADS, not being a service delivered over the Internet but by physical delivery, and requiring significant human intervention to provide the services. Construction of commercial real estate would also not fall within CFB. It is inherently of a scale beyond an individual’s private need or enjoyment. Although an individual may make use of the infrastructure or building (e.g. driving across a bridge or working in an office building), it is not sold to or otherwise acquired by the individual as a product for their personal use. They are therefore not of a type commonly sold to consumers. Therefore, under the general definition, these are not in scope.

144. The third segment, residential real estate, refers to the construction of buildings for the customer’s personal use, such as an apartment or house. Applying the general definition of CFB, MNEs that sell residential real estate would be in scope, given that the nature of the product is of a type commonly sold to consumer. Consumers acquire real estate for personal use; the real estate is designed for sale to consumers (being made available in a form for consumption by individuals, where the MNE has set itself up to supply the real estate to consumers); and it is sold (including leased) to consumers. The MNEs that would be the CFB would be the “owner” of the brand and the direct “retailer” of the real estate (which in both cases may be the developer).15

145. However, this outcome does not align with the policy rationale for Amount A. That is (i) MNEs can engage with consumers in a market jurisdiction from a remote location in a way that represents a sustained engagement in the market jurisdiction; and (ii) the current taxing rules, being based on physical presence, do not adequately capture this new engagement.

146. For residential real estate, the first element – the consumer relationship – is not one of remote engagement (such as through the use of marketing) to the same extent as in other CFB. This is because even if the construction business has a well-known brand, ultimately the consumer’s decision to buy will be determined by the characteristics of the real estate specific to that market jurisdiction, such as its size, its location (e.g. in a desirable neighbourhood), and its other features (e.g. garage, security, etc.). In addition, given the nature of the relationship to the product, it is difficult to practically extract consumer-facing intangible property and separate it from the jurisdiction of the sale.

147. With respect to the second element – lack of physical presence – a company selling or leasing its interest in residential real estate will require a substantial physical presence in a market to earn its revenue. Existing tax rules attach to this physical presence, including through rules governing the creation of a permanent establishment through construction activities (Article 5 of the Model Tax Convention), and rules governing the income and gains from the rental and sale of immovable property (Articles 6 and 13(1) of the Model).

148. As such, market jurisdictions are already well placed under existing corporate tax rules and treaty provisions to impose taxation on the construction MNE given that there will necessarily be a substantial and taxable physical presence in the market. Finally, were the construction sector and sale of immovable property to be included in the scope of Amount A, the requirement to invest in physical infrastructure in the market jurisdiction makes it in any event unlikely that there would be a material re-allocation of profits under the principles of Amount A, given the revenue sourcing rules would in any case allocate Amount A to the jurisdiction where the real estate is “delivered” which in this case would be where the real estate is physically located.

149. For these reasons, the construction, sale or letting of residential dwellings by businesses that have engaged in the development and construction of the residential real estate is proposed to be excluded from the scope of Amount A.

150. For the purposes of this exclusion, residential dwelling means the structure and land intended to be used or occupied, in whole or in part, as the home of one or more persons. The definition of residential dwelling means that the offering of temporary accommodation, such as hotels, remain in the scope of Amount A. While such businesses will use real property in their business, they are not selling or letting an interest in that property to consumers, but rather providing a service to consumers, and which remains in scope of CFB.

151. The exclusion is limited to those businesses that have engaged in the development and construction of the residential real estate (and therefore have the physical presence in the jurisdiction), and derive their revenue from its sale or leasing. As such, this exclusion does not include businesses that derive commission or service fees that are not based on the sale or leasing of an interest in land, and do not depend on a physical presence in the market location of the real estate. For example, online intermediation platforms intermediating offers of real estate would be within the scope of Amount A. The exclusion would likewise not cover professional services related to construction (such as those provided by architects, designers, lawyers, consultants), or professional services or intermediation services facilitating the sale or leasing of property (such as those provided by real estate agents, mortgage brokers, financial institutions), although they may be otherwise excluded from the scope under the general rules.

152. Further technical work will be conducted to address the link between the type of income that a franchising business may earn (which could include rental income), and the analysis above on commercial real estate, to ensure that revenue from franchising is appropriately included in the scope of Amount A.

153. Apart from these, there are certain sectors related to the operation of infrastructure which is provided as a service to consumers. Further work will be done to consider an exclusion for the operation of infrastructure businesses, such as:

• electricity generation and distribution,

• natural gas and water distribution,

• certain telecommunications (except perhaps satellite phones, which do not require physical presence in a jurisdiction);

• railways, highways, ports and airports; and

• public transportation.

154. While there is both public and private provision of infrastructure services, businesses are increasingly involved, sometimes as part of public-private partnerships (P3).

155. There appear to be good reasons for such an exclusion. Infrastructure services businesses are by definition closely tied physically to the market where the activity is carried out. The source and market are generally the same, so arguably there is no need to apply re-allocation rules, since substantial profits are already allocated to the market. With the possible exception of telecommunications, branding generally is not an important component of profitability for these businesses. In fact, many of these businesses are “natural monopolies” because it is often not efficient for different suppliers to build competing networks in the same place. For this reason, infrastructure businesses are often subject to price regulation to protect consumers. This limits the ability to earn residual profits, providing an additional rationale for exclusion. Further, a general exclusion for all electricity generation and distribution would be consistent and neutral with respect to the exclusion for renewable energy production.

International air and shipping businesses

156. This section sets out the justification for the Inclusive Framework’s statement that it would be “inappropriate to include airline and shipping businesses in the scope of the new taxing right”.16

157. It has long been recognised that the characteristics of international air and shipping businesses give rise to special income tax considerations. Unlike other types of enterprises of one jurisdiction doing business in another jurisdiction, the earnings of international air and shipping enterprises are based upon the use of vessels in operations between multiple tax jurisdictions, much of the time conducted outside any tax jurisdiction – that is, on or over the high seas – raising the prospect of either multiple taxation or considerable income allocation challenges.

158. For this reason, there is a longstanding international consensus that the profits of enterprises operating ships or aircraft in international traffic should be taxable only in the jurisdiction in which the enterprise has its residence. This special treatment, which is applied regardless of whether such an enterprise carries on business through foreign permanent establishments, is reflected in Article 8 of both the OECD and United Nations (UN) Model Tax Conventions17 and in the vast majority of the 3,500+ bilateral tax treaties currently in force. It reciprocally exempts in each jurisdiction the profits of non-resident international air and shipping enterprises, removing the compliance and administrative burdens (and associated prospect of disputes) that would otherwise arise, especially over the attribution of profits.

159. International air and shipping businesses commonly maintain a physical presence, which would constitute a permanent establishment, in multiple jurisdictions. The exemption enjoyed by these enterprises under Article 8 is therefore a deliberate policy choice, reflecting the unique considerations briefly outlined above, and does not result from increasing scale without mass or other factors arising from the digitalisation of the economy.

160. With the possible exception of the provision of online marketplace platforms, the activities of international air and shipping enterprises (i.e. income to which Article 8 applies) should be outside the proposed definition of ADS as those services are not delivered over the Internet. The nature of those services has not changed as a result of digitalisation and accordingly, the Inclusive Framework sees no reason to adjust how the existing rules apply.

161. Air freight and cargo shipping activities, including mail services, parcel delivery18 and both bulk and liner shipping, will, as almost exclusively B2B services, be outside the definition of CFB. But passenger transportation services, including pleasure cruises, foot and car ferry crossing services, passenger air travel, and associated services (including those delivered online) are within the definition of CFB.

162. The question therefore is whether in-scope activities of international air and shipping businesses should be taken out-of-scope, on policy grounds.

163. In the special circumstances applying to international air and shipping businesses, it was the allocation of taxing rights, under those normal rules, to the multiple jurisdictions in which physical operations are conducted that gave rise to the particular policy problem to which the consensus solution was and has remained exclusive residence state taxation. Those same positions continue today. Given the present policy choice with respect to this sector, members agree that air and shipping businesses be carved out-of-scope for the purpose of application of new taxing rights.

164. Profits that would fall within Article 8 of the OECD Model Tax Convention would therefore be excluded from the scope of Amount A. This would be the case whether or not a bilateral tax treaty exists between the jurisdictions in question.

2.2.4. Alternative to Activities Based Tests: Applying Amount A on a Safe Harbour Basis

165. As described above, the Outline proposes to base the scope for Amount A on an activities test to include automated digital services and consumer facing businesses, with defined carve outs. The intent is to have Amount A apply to a broad range of businesses that generate profits through active and sustained engagement in market jurisdictions. Although considerable technical work has been devoted to the practical, administrative and definitional issues of such an approach and good progress has been made, this approach has to date not achieved political consensus. The need to define ADS and CFB, together with the need for facts and circumstances determinations to apply such definitions, adds complexity, in particular with respect to CFB which covers a wider and more diverse range of business models. In this context, the United States has proposed that Pillar One be implemented on what it refers to as a safe harbour basis.

166. A safe harbour basis would recognise that greatly enhanced tax certainty and robust mechanisms for the resolution of cross border tax disputes have always been critical components of the design of Pillar One. Such enhanced tax certainty in many cases could make Pillar One attractive to electing MNE groups notwithstanding a marginal increase to their global tax liabilities resulting from allocations of Amount A.

167. Under such a safe harbour implementation, MNE’s could elect to have all the components of Pillar One apply to them on a global basis, including the Amount A allocation, the Amount B fixed margin mechanism and the mandatory binding dispute prevention and resolution procedures. By allowing any MNE to elect to apply Pillar One, the need to resolve contentious scoping issues, including the definitions of ADS and CFB, would potentially be reduced. Election procedures could be provided to require that an MNE’s election be made on a global, multi-year basis. In the view of the United States, implementing Pillar One on a safe harbour basis could also avoid the political challenges of mandating changes to longstanding international tax principles.

168. The United States contemplates that such a safe harbour implementation of Pillar One would be part of an overall agreement that replaces digital services taxes and similar unilateral measures, with the effect that the benefits of increased administrability, greater tax certainty, and mandatory binding dispute resolution procedures would be the primary motivations for MNEs to avail themselves of Pillar One as a safe harbour.

169. Many other jurisdictions have expressed scepticism about an elective approach. In addition to noting the outstanding questions around how an elective regime would operate, they have argued that the objectives of Pillar One would be frustrated in cases where MNEs do not elect into the new rules, and have noted the perverse incentives it could create for jurisdictions to introduce unilateral measures, or increase the breadth and rate of existing unilateral measures, in order to protect against that outcome. They believe that an optional basis would create inconsistency across similar businesses and lead to distortions. They also take the view that an elective regime would undermine the policy justification for reallocating profit to market jurisdictions, threatening its coherence/sustainability and reducing the likelihood of international consensus. Some have further questioned the legal effect of a safe harbour approach, given that such a safe harbour can only be provided in a coordinated manner by jurisdictions if they have the relevant taxing rights in the first place, which will not be the case for all in a safe harbour approach. Further, others emphasise that the tax challenges of the digitalising economy specifically result from their interplay with longstanding international tax principles which therefore need to be addressed with a view to overcoming political challenges.

2.2.5. Next steps

170. The Inclusive Framework will continue its efforts to bridge the remaining differences as regards scope, recognising the importance of coming to a common understanding, and in a spirit of compromise where needed, reflecting its continuing commitment to working together towards a consensus-based long-term solution. This includes taking into account the divergent views expressed above, and remaining open to simplifications and innovative approaches that help to achieve the tax policy objective while minimising compliance and administrative costs.

2.3. Threshold tests

171. In defining the MNEs that are in scope of Amount A, there is a recognition that below a certain overall size threshold, a cost-benefit analysis does not justify the imposition of the rules required to apply Amount A. The implementation of Amount A is likely to be associated with additional compliance and administrative costs, especially in the short term, as taxpayers and tax administrations implement the new rules. Large businesses are more likely to possess the financial and human resources and systems in their tax function to manage and implement new rules and bear the additional compliance costs inherent in complying with the new taxing right, whereas smaller MNEs may struggle to assemble the necessary resources. Equally, the Amount A that would be available to be potentially allocated to market jurisdictions from smaller MNEs will be limited in absolute terms, resulting in significant compliance costs in exchange for insignificant benefits.

172. Similarly, even for larger MNEs that may be above a defined size threshold, if they have only a small amount of foreign revenue that is in scope the cost of compliance associated with Amount A may still exceed benefits.

173. The above observations are particularly true when considering not just the compliance cost of the MNE but also taking into account the wider administration cost for tax administrations that would be required to process and verify compliance for a large number of taxpayers, whether as part of the early tax certainty process or otherwise.

174. Against this backdrop, this Blueprint contains two thresholds designed to achieve this:

• a global revenue test; and

• a de minimis foreign in-scope revenue test.

2.3.1. Global revenue test

175. Gross revenue19 would seem to be the easiest metric to use for determining size. A gross revenue test would allow excluding “smaller” MNEs on the basis of the annual consolidated group revenue, as shown in its consolidated financial statements. This then raises the question of where to set the level of consolidated revenue. Considering the costs and benefits, there may be little advantage in using a threshold below the current EUR 750 million threshold that is used for the purposes of Country-by-Country reporting (CbCR).20

176. First, the impact assessment shows that there is very little increase in the residual profit that would be allocated to market jurisdictions from using a lower figure. Second, it substantially increases compliance and administrative costs. A lower threshold brings in a large number of additional taxpayers that need to determine whether they have revenue in scope, and if they do, comply with the new system. This results in burdens not only for those businesses, but equally for the tax administrations of all Inclusive Framework members to manage the additional compliance and administrative costs.

177. Furthermore, going lower than EUR 750 million will substantially increase the number of privately held groups in scope that currently are not required to prepare financial statements, or if they do, they are preparing them on the basis of purely local generally accepted accounting principles (GAAP). Requiring such privately held groups to prepare financial accounts on the basis of a different standard (or adjusting to that standard) solely for the purposes of Pillar One (and one that may then not be subject to audit requirements) would be difficult and costly for taxpayers, and resource intensive for tax administrations to verify.

178. Finally, a tax certainty process with anywhere near the number of taxpayers that a threshold below EUR 750 million would generate may well go beyond the capacity of tax administrations to operate.

179. Generally, and to avoid tax administrations being overwhelmed with the operation of a new system, including as it relates to tax certainty, it will be important that Amount A is limited to a manageable number of MNE groups. Further, it is difficult to foresee the tax certainty process being applied to review the computation of Amount A for thousands of MNE groups from a standing start. Rather, it is reasonable to assume that the new process will require some time to ramp-up, to allow for identifying and addressing operational challenges before any expansion.

180. At the same time, the threshold needs to ensure that it does not exclude a material part of residual profits otherwise in scope. In light of this, and to minimise compliance costs and keep the administration of the new rules manageable for tax administrations, further work will explore different approaches. One approach could be to implement the threshold on a phased approach. This could start with a higher global revenue threshold that could either be gradually reduced over a number of years or be applied for a longer period and then only start the reductions after a post-implementation review has been undertaken. Other approaches will also be explored.

181. The table below contains some economic analysis on the number of MNE groups that could potentially fall in scope of Amount A at different global revenue levels. (see Table 2.1 below).

Table 2.1. Estimated number of MNE groups above potential global revenue thresholds

Global Revenue Threshold (EUR m) 1	Estimated Number of MNE groups after applying global revenue Threshold	Estimated Number of MNE groups with a primary activity in ADS or CFB sectors after applying the global revenue threshold
750	~8,000	~2,300
1,000	~6,800	~2,000
2,000	~4,100	~1,300
5,000	~2,000	~620
10,000	~1,000	~350

1. See also Table 6.1, which reflects the number of MNE groups above a residual profit threshold after applying a global revenue threshold of EUR 750 million.

2.3.2. De minimis foreign in-scope revenue test

182. The second threshold relates to MNEs that exceed the gross revenue threshold above, but only have a small amount of foreign source in-scope revenue. In such cases, the total profit to be allocated under the new taxing right would not be material relative to the costs to businesses and tax administrations arising from the application of the Amount A rules. First, for MNEs with less in-scope revenue that than the global revenue test the potential profits allocable to market jurisdictions under Amount A will be relatively small. This reflects the fact that an MNE with in-scope revenue of, for example, EUR 250 million and a profit margin of 20% (which is relatively high) will only have profit before tax (PBT) of EUR 50 million that would be in scope of Amount A. Further, the profits allocable to market jurisdictions under Amount A will be further constrained through the application of the profit allocation formula (i.e. the profitability threshold and reallocation percentages).21 Second, for MNEs that primarily earn in-scope revenue in a single jurisdiction, applying Amount A is likely to have a limited tax impact, because it is likely that the Amount A formula will allocate taxing rights over an MNE’s residual profits to the same jurisdiction that already has taxing rights under the current ALP-based profit allocation rules.

183. For this reason, a secondary de minimis foreign in-scope revenue test would be applied to determine the MNEs in scope of Amount A. This threshold for de minimis foreign in-scope revenue would be set at an absolute number, rather than being relative to the size of a given MNE’s domestic business. This has the benefit of achieving the same outcome for MNE groups that engage in foreign market jurisdictions to a similar extent, irrespective of whether that foreign market engagement is significantly smaller than the domestic business or not (and which may otherwise lead to different outcomes depending on the relative size of the domestic market).

184. This test would have two steps. First, an MNE would apply the activities test to determine whether the group earned more than the de minimis foreign in-scope revenue threshold amount from ADS or CFB activities. MNEs that do not earn more than the de minimis foreign in-scope revenue threshold amount from in-scope activities would be excluded from scope on this basis. Second, the MNE would then need to determine whether it earned more than the de minimis foreign in-scope revenue threshold amount from “foreign” in-scope activities. This will first require an MNE to identify its domestic or home market using a standardised definition, for example, where the group is headquartered or where the ultimate parent entity (UPE) is tax resident. Further work is required to define an MNE Group’s domestic or home market in order to prevent manipulation. The MNE would then need to determine whether it derived more than the de minimis foreign in-scope revenue threshold amount in in-scope revenue from jurisdictions outside its home or domestic market, based on the Amount A revenue sourcing rules (see Chapter 4). This would be simple for MNEs that derives all its in-scope revenue from its own domestic market, at which this second step is primarily targeted. But it could also be relevant for other MNEs that generate relatively small amounts of revenue outside their home market. Though the application of the Amount A revenue sourcing rules may be perceived to create complexity for MNEs seeking to apply this test, doing so would allow eligible MNEs to avoid the compliance costs associated with the other aspects of applying Amount A. Further work will be undertaken on this threshold including exploring whether any thresholds should be introduced on a phased basis.22

2.3.3. Next steps

185. As a next step, a decision will be necessary to agree the definitive thresholds, including whether a phase in/transition period is appropriate and, if so, its design.

Notes

← 1.  see CTPA/CFA/WP2/NOE2(2020)10).

← 2.  Other sectors are in the medical equipment business, such as selling consumer goods (bandages, blood pressure meters); medical equipment not purchased by consumers (e.g. hip replacement materials); or medical equipment for professional use (e.g. MRI machines, surgical tools). These are outside the scope of the pharmaceutical industry, and the ordinary rules applying to determine the scope of CFB, including the approach to dual use goods and components, will be used to address these sectors.

← 3.  Retail pharmaceutical refers to those provided outside hospital care, such as those dispensed through a pharmacy or bought from a supermarket.

← 4.  The income can also include rent (where the site of the franchise operation is leased to the franchisee). However, an exclusion from scope from Amount A is being considered for commercial real estate. If agreed, the interaction between that exclusion and franchising fees may require further clarification.

← 5.  This is separate from the notion of an intermediary product / component sold B2B, as discussed below. In a product distribution franchising arrangement, the company that sells the intermediate product has a significant degree of control over the manner in which the franchisee company assembles the finished product to the consumer (and the franchisee can only use that intermediate product to assemble the final product as instructed by the franchisor); and the franchisor derives revenues from the sales of the end product. Whereas in a sale of a component part B2B, the seller of the component has very limited control in how the buyer uses the component for assembly in its own finished products (and which may be used in a variety of different types of the buyer’s products), and the seller of the component earns revenue from the actual sale of the component as opposed to revenue from the finished product.

← 6.  See also section 6.4 and section 7.3.

← 7.  It therefore does not include the licensing of trade intangibles or manufacturing know-how.

← 8.  See also section 6.4, and section 7.3.

← 9.  Extractive businesses are those engaged in the exploration for, and extraction from the earth’s crust of, non-renewable natural resources such as hydrocarbons and minerals, the processing and refining of those resources into usable commodities, and the sale of those commodities.

← 10.  While natural resources businesses may make use of digital tools, the product or service they offer is not a digital service itself.

← 11.  Non-renewable resource include hydrocarbons and minerals. Upstream hydrocarbons extracted from the earth such as crude oil (petroleum) and natural gas are out of scope even under the general CFB definition, as they are not sold to consumers. After processing, intermediate products such as ethane, pentanes, asphalt, etc., are commonly sold as inputs to other industries such as chemical production and construction, and likewise do not meet the test of being sold to consumers. Minerals, including base metals (copper, zinc) and precious metals (gold, silver) are extracted from the earth before being refined and smelted into metals and other mineral commodities. Most mineral products (steel, aluminium, gypsum) are not commonly sold to consumers and so are not within the scope of CFB. Likewise, sale of rough or raw gemstones and those for industrial use would be natural resources that are excluded from scope.

← 12.  For example, crops (a sack of wheat or green coffee beans), a haul of caught fish (whole cod), and a stack of timber.

← 13.  Application of the factors, however, will require some nuance. For example, there is a spectrum of processing, packaging and branding when considering agricultural products, ranging from raw fruit (which may carry a brand label for retail sale, but where the product is untransformed, unpackaged and generic), through fruit that has undergone simple processing (e.g. cutting and peeling) before being placed in a branded can, to highly processed and branded food products that are combined with other ingredients like fruit-based confectionary that would be the products in scope of conventional CFB. Other examples that demonstrate the need for nuance include bottled water and speciality cheeses. A similar analysis could be applied to fishery and forestry products. For example, raw logs undergo simple processing into dimensional lumber, paper and cardboard (where value is primarily attributable to the raw material and is determined chiefly by their physical characteristics) but which can also be a type of product commonly sold to consumers, compared with further transformation into wooden furniture (which derives its value mainly from manufacturing and sales) which would fall under the CFB rules.

← 14. . For example, the United Nations (UN) for the Trade and Development Conference (1994) noted Genoa, Barcelona, Bruges, Brussels and Antwerp regulated insurance as early as the fifteenth and sixteenth centuries and that today “in most countries, if not all, there are specific regulations concerning the insurance business.” See https://unctad.org/en/PublicationsLibrary/unctadsddins6_en.pdf.

← 15.  A third party construction company contracted by the developer to build the residence would be the “manufacturer,” without a relationship to the consumer, and would be out of scope applying the general definition.

← 16.  Paragraph 32 of the Outline.

← 17.  For shipping companies, the UN Model contains an “alternative B”, which awards limited source state taxing rights over “an appropriate allocation of the overall net profits derived by the enterprise from its [non-casual] shipping operations”. It is little-used.

← 18.  Where a parcel is delivered to a consumer, the parcel company’s customer is invariably the company sending the parcel – it is not the consumer that receives it. Parcel companies receive virtually no revenue from individuals sending parcels abroad.

← 19.  No decision has yet been taken by the Inclusive Framework on the amount or design of the revenue thresholds for the application of Amount A. Further work will be required to determine if the existing definition of revenue provided by accounting standards and shown in financial statements (which are relied upon for CbCR) could be used for the purposes of applying the global revenue test (as well as to determine the Amount A tax base, see section 6.2.3).

← 20.  CbCR also refers to Country-by-Country Report in this document.

← 21.  If it were assume that for the Amount A formula it is agreed that 20% of the MNE’s profits in excess of a 10% profit margin would be allocated to the market. Under Amount A, the MNE’s residual profits would be EUR 25 million, of which EUR 5 million (20%) would be allocated to market jurisdictions under Amount A. At a 25% corporate tax rate, this would equate to EUR 1.25 million in additional CIT or EUR 125,000 if this amount were split equally between 10 market jurisdictions.

← 22.  An issue has also been raised about whether it is equitable under the approach outlined in section 2.3 that a larger firm with a small volume of in-scope revenue should be in scope of Amount A, when a smaller firm with a much higher volume of in-scope revenue may be out-of-scope depending on the amount of the global and de minimis foreign in-scope revenue thresholds. Given this issue, consideration will also be given to a threshold based on in-scope revenues.