# 6. Profit allocation

496. The calculation and allocation of Amount A will be delivered through a formula that is not based on the ALP. This formula will apply to the tax base of a group (or segment where relevant), and will involve three distinct components represented in the steps below:

• Step 1: A profitability threshold to isolate the residual profit potentially subject to reallocation, and limit any interactions between Amount A and the remuneration of routine activities under conventional transfer pricing rules. To avoid complexity, this threshold would be based on a simplifying convention, which will be a PBT to revenue ratio.

• Step 2: A reallocation percentage to identify an appropriate share of residual profit that can be allocated to market jurisdictions under Amount A (hereafter, the “allocable tax base”). This will ensure that other factors such as trade intangibles, capital and risk, continue to be remunerated and allocated residual profit. To avoid complexity, this allocable tax base will be determined through a simplifying convention, which will be a fixed percentage.

• Step 3: An allocation key to distribute the allocable tax base amongst the eligible market jurisdictions (i.e. where nexus is established for Amount A). It will be based on locally sourced in-scope revenue determined by applying the rules on scope, nexus and revenue sourcing (see Chapter 4).

497. This three-step formula to determining the Amount A quantum could be delivered through two approaches: a profit-based approach or a profit margin-based approach. A profit-based approach would start the calculation with the Amount A tax base determined as a profit amount (e.g. an absolute profit of EUR 10 million), whereas a profit-margin approach would start the calculation with the Amount A tax base determined as a profit margin (e.g. a PBT to revenue of 15%). Both approaches would apply the three-steps of the allocation formula similarly, and hence would deliver the same quantum of Amount A taxable in each market jurisdiction. The administration of each approach may, however, present some variations (e.g. foreign currency exchanges), and these practical differences will inform the choice of the most appropriate approach to calculate and allocate Amount A. These two approaches are explored in more detail in Annex B.

498. As part of the comprehensive agreement still needed, it will be necessary to determine whether the formula should incorporate any “differentiation” mechanism. That is, whether the different components of this formula should apply similarly in all circumstances, or whether some variations (for example the profitability threshold under step 1 and/or the reallocation percentage under step 2) should sometimes be applied to increase (or decrease) the quantum of profit reallocated to market jurisdictions for certain business activities. No agreement has yet been reached on either the policy merits of these variations or their feasibility from a technical design perspective. There will also be some remaining issues around questions of regional and jurisdictional segmentation.

499. The Outline highlighted an important question as to whether the interactions between Amount A and existing taxing rights of market jurisdictions could, in some circumstances, result in a market jurisdiction being able to tax twice the residual profit of an MNE group: once under its existing taxing rights, and again through Amount A (the issue of “double counting”). The issue of double counting is expected to be addressed, at least partially, through the mechanism to eliminate double taxation (see Chapter 7). This is because where an entity is allocated significant residual profit in a market jurisdiction under existing profit allocation rules, this entity may be identified as a “paying entity” within the group for the purpose of eliminating double taxation which would bear a portion of the Amount A tax liability (resulting in a “netting-off” effect). However, some members of the Inclusive Framework suggest that, on its own:

• This may not fit with the overall rationale for Pillar One (and Amount A specifically) which has always been to adapt the income tax system where businesses have an active and sustained engagement in a market jurisdiction, but the existing profit allocation rules do not give that jurisdiction taxing rights over residual profits generated in that market. So, if Amount A did apply to businesses that already realise residual profits in the market, the problem Pillar One is trying to solve may not seem to be present.

• Applying the mechanism to eliminate double taxation to decentralised businesses that realise residual profits in a large number of entities and jurisdiction will be complex. Specifically, it may be difficult to calibrate this system to ensure that a full-risk distributor entitled to residual profit is identified as the paying entity for the Amount A allocated to the jurisdiction in which it is resident. For this reason, it may be preferable to develop a method that would reduce pressure on the elimination system, allowing this system to focus on more centralised businesses where it will be comparably easy to identify the paying entities.

500. The marketing and distribution profits safe harbour described below is an approach that seeks to address these issues related to double counting, as well as a number of other issues expressed by Inclusive Framework members and stakeholders. It would be an additional step in the Amount A formula to adjust the quantum of Amount A allocated to eligible market jurisdictions in specific circumstances. Consideration is also being given to other approaches to deal (or alleviate) double counting beyond the mechanism to eliminate double taxation, such as a domestic business exemption.

501. The premise of the “marketing and distribution profits safe harbour” is that Amount A should be allocated to a market jurisdiction that is not allocated residual profits under the existing profit allocation rules but should not be allocated to a market jurisdiction where (for its in-scope activities) an MNE group already leaves sufficient residual profit in the market. It would not be a traditional safe harbour, but would instead “cap” the allocation of Amount A to market jurisdictions that already have taxing rights over a group’s profits under existing tax rules. Conceptually, it would consider the income taxes payable in the market jurisdiction under existing taxing rights and Amount A together, and adjust the quantum of Amount A taxable in a market jurisdiction, on the basis of limiting it where the residual profit of the MNE group is already taxed in that jurisdiction as a result of the application of the existing profit allocation rules. Where an MNE qualifies under the safe harbour in the market jurisdictions where it operates, it would need to calculate Amount A, but would otherwise remain subject to the existing rules including on transfer pricing and the elimination of double taxation.

502. Under the safe harbour, where an MNE group has a taxable presence in a market jurisdiction conducting marketing and distribution activities connected to locally sourced in-scope revenue (either a resident entity or a permanent establishment), the group would determine the profits allocated to the market jurisdiction under existing profit allocation rules for the performance of these marketing and distribution activities (the “existing marketing and distribution profit”).1 The MNE group would then compare this with the “safe harbour return”, which would be the sum of two components:

1. Amount A, as computed under the Amount A formula; and

2. A fixed return for in-country routine marketing and distribution activities, which could include a regional, and industry uplift.

503. The safe harbour return represents the cap, by reference to which the quantum of Amount A allocated to a market jurisdiction would be adjusted. It would be applied by an MNE group separately to each market jurisdiction in which they operate and would give rise to three possible outcomes:

• Where the existing marketing and distribution profit is lower than the fixed return, the MNE group will not be eligible for the safe harbour;

• Where the existing marketing and distribution profits exceeds the fixed return, but falls below the safe harbour return, the quantum of Amount A allocated to that jurisdiction would be reduced to the difference between the safe harbour return and the profit already allocated to the local presence; and

• Where existing marketing and distribution profit exceeds the safe harbour return, no Amount A would be allocated to that jurisdiction.

504. In-scope MNE groups that for commercial reasons (given their particular business models) operate without an existing taxable presence in a market jurisdiction or only allocate a relatively limited return (e.g. on a cost-plus basis) to local marketing and distribution activities would not come under the safe harbour rule and thus would pay Amount A in the majority of market jurisdictions in which they operate. In contrast, more traditional CFB businesses, particularly those with decentralised business models and full-risk distributors, may already allocate profits to market jurisdictions that exceed the safe harbour return. Hence, though these businesses would need to calculate Amount A (to determine that they have met the safe harbour), they would in many instances ultimately not need to pay Amount A or apply the mechanism to eliminate double taxation.

505. This mechanism would exclude from the scope of Amount A profits derived by an ADS or CFB business in a market jurisdiction which can be seen as autonomous from the rest of the group, i.e. sale of goods or services that are developed, manufactured and sold in a single jurisdiction. As in this scenario residual profit is typically already allocated to the market, this exemption would prevent the risk of double counting in those circumstances.

506. Further work will be required to assess the effectiveness, the efficiency and the feasibility of the different options to deal with double counting, in close coordination with the work on the mechanism to eliminate double taxation and in light of the policy objective of Amount A (see Chapter 7). This will also include consideration of the interactions between Amount A and certain withholding taxes collected by market jurisdictions.

507. Amount A represents a simplified proxy of the portion of the residual profit of a business that can reasonably be associated with the sustained and significant participation of that business in the economy of a market jurisdiction. To isolate the residual profit of a business (group or segment where relevant) potentially subject to reallocation under Amount A, the formula includes a profitability threshold. This threshold is based on a simplified convention (i.e. a fixed percentage), and will apply to the Amount A tax base after the deduction of any available losses carried forward (see section 5.4).

508. One reason for introducing and using a fixed threshold, rather than a variable percentage based on facts and circumstances or related transfer pricing analysis, is to reduce complexity. The profitability threshold will materially reduce the scale of interactions of Amount A with conventional transfer pricing rules (e.g. remuneration of routine activities), and hence the complexity that these interactions create to eliminate double taxation (and double counting, see below Section 6.4). This threshold will not alter the allocation of profit derived from routine activities under the current transfer pricing rules (given that Amount A operates as an overlay to the existing profit allocation rules), but will simplify the identification and calculation of the residual profit subject to the new taxing right.

509. To achieve these results, the profitability threshold will be based on a simplifying convention (i.e. proxy). Consistent with the logic adopted for tax base determinations, and to facilitate both administration and compliance, the profitability of an MNE group (or segment) would be assessed through an Amount A PBT to revenue ratio (i.e. a percentage).2 The determination of this figure will not rest on any MNE-specific economic assessment nor necessarily correspond with underlying transfer pricing arrangements. The impact of different profitability thresholds is shown in the Table 6.1 below.

510. More data and analysis is available in the economic impact assessment.3 With these estimates, Inclusive Framework members will be able to take an informed decision when setting the threshold. The decision on the level of the profitability threshold will seek to combine different objectives, such as ensuring the amount of profit to be reallocated is modest but meaningful, proportionate to compliance costs and administrative burden, and that the number of groups impacted is kept at an administrable level. For the purpose of illustration, it is noted that based on a 10% threshold of PBT to revenue, the above estimates in Table 2.1 suggest that about 780 MNE groups potentially in scope of Amount A would have residual profits. This would represent about 35% of MNE groups subject to CbCR with a primary activity in ADS and CFB sectors. Further, the combined residual profit of these MNE groups would be USD 0.51 trillion.

511. Under Pillar One, only a portion of the residual profit of a group (or segment where relevant) is attributable to Amount A. This is because MNE groups perform a variety of activities unrelated to Amount A that generate residual profit, and hence a substantial portion of the group’s residual profit should continue to be allocated under existing rules to factors such as trade intangibles, capital and risk, etc.

512. The formulaic calculation of Amount A thus requires an additional step: the reallocation percentage. For simplicity, the share of residual profit that is attributable to the market jurisdiction will be determined by a simplifying convention (i.e. proxy) not based on the particular circumstances of the MNE group or the ALP. Such a convention could be residual profit multiplied by a fixed percentage. Consistent with the estimates provided above in Table 6.1, some estimates of the impact of different reallocation percentages (in combination with different possible profitability thresholds) on the amount of global residual profit allocable to market jurisdictions are shown in the Table 6.2 below.

513. More data and analysis is available in the economic impact assessment.4 With these estimates, Inclusive Framework members will be able to take an informed decision on the reallocation percentage. This decision will seek to combine different objectives, such as ensuring that activities and factors generating residual profit unrelated to Amount A would continue to be taxed under the existing ALP-based profit allocation system. For the purpose of illustration, it is noted that based on a 10% profitability threshold and 20% reallocation percentage, the above estimates in Table 6.2 suggests that using 2016 financials USD 98 billion would be allocated to market jurisdictions. Under this approach, 80% of the residual profit of an MNE group (or segment where relevant) calculated for the purpose of Amount A would thus continue to be taxed in accordance with the existing ALP-based profit allocation system, and the other 20% would constitute the allocable tax base for Amount A purposes.

514. Once the calculation of the allocable tax base for Amount A is completed, that profit needs to be allocated to the various eligible market jurisdictions based on an allocation key. This allocation is based on in-scope revenue derived from each eligible market jurisdiction (for revenue sourcing rules, see Chapter 4.2).

515. The application of this allocation key will require a clear definition of revenue. Under accounting standards, revenues are typically booked on a gross basis, net of certain types of taxes (including sales, use, value added and some excise taxes). Further work will be required to determine if the existing definitions of revenue provided by accounting standards and shown in financial statements (which are relied upon for CbCR) could be used to define revenue for the purposes of applying the Amount A formula (as well as to apply the threshold tests for scope, see section 2.3).

516. The application of the revenue-based allocation key will differ depending on whether the formula is implemented through a profit-based or a profit-margin approach (see section 6.2.4). Under a profit-based approach, the allocable tax base (a profit amount, i.e. PBT) could be multiplied by the ratio of locally sourced in-scope revenue to total revenue of an MNE group (or segment where relevant) used in computing the tax base, including revenue from ineligible market jurisdictions (where no nexus would be established for Amount A purposes) and potentially out-of-scope revenue. Under a profit-margin approach, the allocable tax base (a profit ratio, i.e. PBT / revenue) could be multiplied by locally sourced in-scope revenue.

517. Both of these approaches would ensure that Amount A profits attributable to revenue sourced in ineligible market jurisdictions are not allocated to other eligible market jurisdictions, and remain instead taxed under the existing profit allocation system (a so-called “throwback” system). In practice, given the likely level of the nexus revenue thresholds, (see section 3.2.1), the Amount A profit attributable to ineligible market jurisdictions is likely to be small. Both approaches would also ensure that where the calculation of the Amount A tax base (at the level of a group or segment) includes profit from out-of-scope revenue, the portion of the Amount A tax base that relates to out-of-scope revenue will not be reallocated to market jurisdictions.5

518. Another possible approach would be to allocate the entire allocable tax base (as determined by steps 1 and 2) between eligible market jurisdictions, and allocate Amount A profits related to revenue sourced in ineligible market jurisdictions to eligible market jurisdictions (the “throw-in” system). This would modify the application of the revenue-based allocation key described above in paragraph 516, and require determining the source of all in-scope revenue (including revenue sourced in non-eligible market jurisdictions).6

519. To summarise, a three-step process will be required to calculate the quantum of Amount A taxable in each eligible market jurisdiction. This process could be implemented by either using absolute amounts of profit (the “profit-based approach”) or, alternatively, profit ratios (the “profit margin-based approach”). A profit-based approach would start the calculation from the Amount A tax base determined as a profit amount (e.g. an absolute profit of EUR 10 million), whereas a profit-margin approach would start the calculation from the Amount A tax base determined as a profit margin (e.g. a PBT to revenue of 15%). Both approaches would apply the above-described steps without changes or variations, and hence would provide the same quantum of Amount A taxable in each market jurisdiction. As a next step, the Inclusive Framework will determine which approach will be used to implement the Amount A formula. The profit-based approach and profit margin-based approach are discussed in more detail in Annex B.

520. The technical work has considered whether the different components of the formula described above should apply similarly in all circumstances, or whether some variations are necessary to increase (or decrease) the amount of profit reallocated to market jurisdictions in some cases (the “differentiation mechanisms”). Such variations could have a significant impact to the general application of the Amount A rules.

521. Consistent with proposals formulated by some Inclusive Framework members, a differentiation mechanism could potentially be introduced to:

• Account for different degrees of digitalisation between in-scope business activities (e.g. based on the ADS definition used for scope), and increase the quantum of profit reallocated for certain types of business activities (hereafter, “digital differentiation”). This would seek to target businesses with lower marginal costs that are seen as deriving greater benefits from scale without mass.

• Account for substantial variations in profitability between different market jurisdictions, and increase the quantum of profit reallocated to market jurisdictions where the profitability is significantly higher than the average profitability of the segment (hereafter, “jurisdictional differentiation”). Such differentiation could be a simplified alternative to the jurisdictional segmentation examined in the context of tax base determination, which in many instances raises questions about feasibility and administration (see section 5.3).

522. Various mechanisms are available to provide for these types of differentiation. These include: (i) variations to the calculation of the allocable tax base, for example by increasing the reallocation percentage (step 2) in specific circumstances; (ii) variations to the allocation key used to distribute the allocable tax base between market jurisdictions (step 3), for example by weighting the amount of revenue derived from market jurisdictions; and (iii) other variations, such as adding a specific “routine return” for certain digital activities (e.g. ADS) that can be conducted without any physical presence in market jurisdictions.

523. The existing gaps between Inclusive Framework members on this policy issue will need to be resolved as part of the discussion of the quantum of Amount A.

524. In the discussions so far, Inclusive Framework members have different views on whether some form of “digital differentiation” is necessary in the design of the Amount A formula. Accordingly, various options are under consideration, including:

• No differentiation at all – The Amount A formula would apply to all in-scope business activities in the same way, and in all circumstances.

• Digital differentiation through adjustments to Amount A – A lower profitability threshold under step 1 of the formula, or a higher reallocation percentage under step 2, would apply to MNE groups (or segments where relevant) providing primarily ADS. Though intuitively simple, any such differentiation approach would also require exploration as regards its technical and conceptual feasibility considering issues related to businesses segmentation and implications for the respective scope rules.

• Differentiation of Amount A through a profit escalator – A progressive reallocation percentage under step 2 of the formula would be introduced based solely on the profitability of the group (or segment where relevant). This mechanism would not seek to directly account for different degrees of digitalisation between in-scope businesses, but would be premised on the assumption that higher returns reflect at some point a greater contribution of the market to the profitability of the MNE group (e.g. monopolistic rents). The allocable tax base would thus be determined by reference to one or more bandings (e.g. X% for profit margins between a-b%, X+Y% for profit margins between b-c%, X+Y+Z% for profit margins in excess of c%), but without any distinction based on the underlying nature of the in-scope business involved.

525. Further, some members consider that where ADS or CFB businesses make remote sales in a jurisdiction by using digital means to connect with customers, this jurisdiction should also receive an allocation of routine profits for the remote performance of marketing and distribution activities. In their view, once the new nexus threshold under Amount A is reached, it is unfair to deny market jurisdictions taxing rights over businesses that thanks to digitalisation are able to participate in the economic life of their jurisdiction remotely (i.e. making sales, marketing and distributing their products, collecting payments and addressing customer grievances). These members consider that in addition to Amount A, for businesses that operate in a market remotely certain marketing and distribution activities should be seen as taking place in the market jurisdiction and that in such circumstances profits would be allocated to the market jurisdiction. Such reallocated profits could be determined as an agreed percentage (e.g. 30%) of the deemed routine profit margin or the actual profit margin of the MNE group (or segment, where relevant), whichever is lower, subject to a minimum return (e.g. x% of sales) for the deemed routine marketing and distribution activities. Members that favor this approach recognizes that such allocation of deemed routine return for remote marketing and distribution activities needs to address the issue of double taxation. They are of the view, however, that such double taxation can be eliminated under the existing rules. Further, they believe it is important to reduce the incentives MNE groups face to conduct marketing and distribution remotely from a lower tax jurisdiction and to create neutrality between marketing and distribution conducted physically in a market jurisdiction and similar activities conducted remotely thanks to digital technologies. Other members take the view that there would be no case for reallocating both routine profit from distribution activities and residual profit.

526. The Outline identified the need to explore the rationale and technical feasibility of jurisdictional or regional segmentation as a way to account for variations in profitability across regions. For businesses that do not operate on a regional basis, regional segmentation would be technically challenging because, like segmenting between ADS, CFB and out-of-scope activities, it would require that potentially significant portions of central costs are apportioned among different regions using allocation keys. It is also recognised that regional or jurisdictional differentiation is mainly a question for CFB businesses and not for ADS businesses.7 Whilst further technical work will be conducted on this issue, the work so far suggests that regional segmentation may not be sufficient to account for significant variations in profitability across jurisdictions (which are particularly significant in the CFB sector). Consideration has been given to whether the Amount A formula could be weighted to allocate more profits to more profitable markets. However, again the challenges of calculating the profits attributable to a market, mean it can be difficult to accurately identify more or less profitable markets.

527. Conceptually, incorporating jurisdictional differentiation within the Amount A model is particularly challenging, because it is inconsistent with the overall approach which is to calculate the profits allocable to a market jurisdiction on a group or segment basis. However, there are a number of features of Amount A that will help ensure that its introduction does not result in profits from more profitable market jurisdictions, being reallocated to less profitable market jurisdictions:

• Mechanism to eliminate double taxation. The mechanism to eliminate double taxation (see Chapter 7) could include an activities test and a market connection priority test. These two tests would significantly reduce the likelihood of the profits of in-market distributors being reallocated to other markets as a result of the introduction of Amount A.

• Domestic business exemption. As explained in more detail below, a “domestic business exemption” could be developed to exclude profits derived from the sale of goods or services that are developed, manufactured and sold in a single jurisdiction from the Amount A tax base. However, by requiring that profits from domestic businesses are excluded from the Amount A tax base, this exemption would add significant complexity to the determination of the Amount A tax base.

528. Amount A will be allocated as an overlay to the existing income tax system. This means interactions between Amount A and the existing income tax system are inevitable.8 The interactions between Amount A and the existing taxing rights of market jurisdictions on business profit, including withholding taxes, is conceptually challenging and an area where members have expressed different views. An important issue identified by the Outline is whether some of these interactions (i.e. between Amounts A and existing ALP-based profit allocation rules) could result in a duplicative taxation of the same profit of an MNE group in a particular market jurisdiction, which could be inconsistent with the policy intention of Pillar One (the issue of “double counting”). 9 The concern is that the market jurisdiction may get to tax the same item of residual profit twice: once through an existing taxable presence under transfer pricing rules, and again through Amount A. The issue of double counting is expected to be addressed, at least partially, through the mechanism to eliminate double taxation (see Chapter 7). This is because where an entity is allocated significant residual profit in a market jurisdiction under existing profit allocation rules, this entity may be identified as a “paying entity” within the group for the purpose of eliminating double taxation which would bear a portion of the Amount A tax liability (resulting in a “netting-off” effect). However, some members of the Inclusive Framework suggest that this approach on its own has a number of drawbacks:

• It may seem counterintuitive. As explored below, the mechanism to eliminate double taxation may address this issue through a “netting-off” effect, but fundamentally Amount A would still be allocated to a market jurisdiction that already has taxing rights over an MNE group’s residual profits.

• It may be inconsistent with the overall rationale for Pillar One (and Amount A specifically) which has always been to adapt the income tax system where businesses, both ADS and CFB, have an active and sustained engagement in a market jurisdiction, but the existing profit allocation rules do not give that jurisdiction taxing rights over residual profits generated in that market. So, if Amount A did apply to businesses that already realise residual profits in the market, the problem Pillar 1 is trying to solve may not seem to be present.

• Applying the mechanism to eliminate double taxation (see Chapter 7) to decentralised businesses that realise residual profits in a large number of entities and jurisdictions will be complex. Specifically, it may be difficult to calibrate this system to ensure that a full-risk distributor (already allocated residual profit) is identified as the paying entity for the Amount A allocated to the jurisdiction in which it is resident. For this reason, it may be preferable to develop a method that would reduce pressure on the elimination system, allowing this system to focus on more centralised businesses where it will be comparably easy to identify the paying entities.

529. These challenges are variations on the same theme. Amount A can be easily rationalised when it is applied to businesses (both ADS and CFB) that realise residual profit in a handful of jurisdictions, but may be more difficult to rationalise and hence design when it is applied to businesses with less centralised business models that already leave residual profits in the market. It is also true that businesses have consistently pointed out that the ability to leverage off their existing systems that support their current in country distribution activities would seem simpler and would be very welcome. At the same time, discussions are ongoing on the issue of double counting, including on whether marketing and distribution profit allocated to a market jurisdiction under the ALP in excess of a fixed return may be seen as duplicative with Amount A profit. This will include assessing the implications of considering the income taxes payable in the market jurisdiction under existing taxing rights and Amount A together.

530. Consequently, consideration is being given to different options to deal with issues of double counting beyond the mechanism to eliminate double taxation, such as the marketing and distribution profits safe harbour and the domestic business exemption.

531. The elimination of double taxation process is an important element in dealing with any potential double counting in the market jurisdiction, or at least materially reduce it. This is because where an entity is allocated significant residual profit in a market jurisdiction under existing profit allocation rules, this entity may be identified as a “paying entity” within the group for the purpose of eliminating double taxation (see below Section 7.2). Identifying this entity as a “paying entity” for Amount A purposes will, in turn, result in a “netting-off” effect: the residual profit allocated under existing rules to the market jurisdiction will, in effect, be reduced by the method used to relieve double taxation from Amount A (including Amount A allocated to other market jurisdictions). There is a question however as to whether this framework can deliver such netting-off effect in all cases, and hence whether it should be the sole basis to deal with double counting issues.

532. As illustrated in Annex C (see Box C.3.), the netting-off effect can be easily identified when it is applied to an MNE group with a centralised business model (both ADS and CFB) that allocates residual profit to a limited number of jurisdictions, but is more difficult to assess when applied to a group with a decentralised business model that leaves residual profit in multiple market jurisdictions (see Annex C, Box C.4.). For example, the proposed mechanism to eliminate double taxation may not identify an in-market full-risk distributor entitled to residual profit under existing rules as a paying entity (or allocate to it the full responsibility for relieving double tax) if it does not meet the profitability test articulated in section 7.2.3, while other group entities (in different jurisdictions) would satisfy such test. Views differ as to the appropriate result in this situation. The efficiency of this approach on its own needs therefore to be further tested when applied to diverse situations, alongside with other approaches that could be developed in combination with this elimination of double taxation process to deal more effectively with double counting issues.

533. The “marketing and distribution profits safe harbour” would start from the premise that Amount A should be allocated to a market jurisdiction that is not allocated residual profits under existing profit allocation rules, but where a group already allocates and actually earns residual profit in the market on in-scope revenue then there should be no Amount A allocation. This would mean an MNE group would have to compute Amount A under the above rules, but would not allocate it to a market jurisdiction to the extent it already allocates and earns residual profit in that jurisdiction. The marketing and distribution profit safe harbour seeks to deliver this outcome. It would not be a traditional safe harbour, but would instead “cap” the allocation of Amount A to market jurisdictions that already have taxing rights over a group’s profits under existing tax rules. Conceptually, it would consider the income taxes payable in the market jurisdiction under existing taxing rights and Amount A together, and adjust the quantum of Amount A taxable in a market jurisdiction, on the basis of limiting it where the residual profit of the MNE group is already taxed in that jurisdiction as a result of the existing profit allocation rules.

534. Under this approach, the basic mechanics of Amount A would be retained, and the formula itself would remain unchanged. A safe harbour return would be determined which would combine the residual profit that an MNE group would be expected to allocate to a market jurisdiction, with an additional fixed return to compensate the local marketing and distribution presence (more below). The safe harbour would recognise that there are two ways that residual profits relevant to Amount A could be allocated to the market jurisdictions. All MNE groups would calculate Amount A and would then either benefit from the safe harbour or pay Amount A through the new Amount A system.

535. Where a group has a taxable presence in a market jurisdiction conducting marketing and distribution activities connected to locally sourced in-scope revenue (either a resident entity or a permanent establishment), the group would determine the profits allocated to the market jurisdiction under existing profit allocation rules for the performance of these marketing and distribution activities (the “existing marketing and distribution profit”).10 The MNE group would then compare the existing marketing and distribution profit with the “safe harbour return”, which would be the sum of two components:

• Amount A, as computed under the Amount A formula; and

• A fixed return for in-country routine marketing and distribution activities, which could include a regional, and industry uplift.

536. The safe harbour return represents the cap, by reference to which the quantum of Amount A allocated to a market jurisdiction will potentially be adjusted. It would be applied by an MNE group on a market-by-market basis and would give rise to three possible outcomes:

• Where the existing marketing and distribution profit is lower than the fixed return, the MNE group will not be eligible for the safe harbour;

• Where the existing marketing and distribution profit exceeds the fixed return, but falls below the safe harbour return, the quantum of Amount A allocated to that jurisdiction would be reduced to the difference between the safe harbour return and the profit already allocated to the local presence; and

• Where the existing marketing and distribution profit exceeds the safe harbour return, no Amount A would be allocated to that jurisdiction.

537. In situations where an MNE group is eligible for the safe harbour in a market jurisdiction (second and third scenario above), it should be noted that it is possible that an entity resident in that jurisdiction may still be identified as a paying entity for Amount A allocated to other jurisdictions under the mechanism to eliminate double taxation if the other rules and requirements articulated in the Section 7.2, including potentially the market connection priority test, are met.

538. Further, in-scope MNE groups that, for commercial reasons (given their particular business models), operate without an existing taxable presence in a market jurisdiction or only allocate a relatively limited return (e.g. on a cost-plus basis) to local marketing and distribution activities, would not come under the safe harbour rule and thus would pay Amount A in the majority of market jurisdictions in which they operate. In contrast, more traditional CFB businesses, particularly those with decentralised business models and full-risk distributors, may already allocate profits to market jurisdictions that exceed the safe harbour return. Hence, though these businesses would need to calculate Amount A (to determine that they have met the safe harbour), they would in many instances ultimately not need to pay Amount A or apply the mechanism to eliminate double taxation. An example of its application is outlined in Annex C (see Box C.2.).

539. The safe harbour may be particularly relevant for decentralised businesses that realise residual profits in a large number of entities and jurisdictions, where it is conceptual challenging to identify the entity or entities within the group that should bear the Amount A tax liability. The adoption of the safe harbour may reduce in some cases the pressure of the mechanism to eliminate double taxation arising from Amount A and could allow the mechanism to eliminate double taxation to be developed with a focus on businesses with more centralised operating models that are less likely to be impacted by the safe harbour.

540. At the same time, the safe harbour would maintain the need to calculate Amount A, while introducing new rules to implement and administer the capping mechanism. Further, the application of the safe harbour would need to take into account any subsequent transfer pricing adjustments that changed marketing and distribution profits allocated to a market jurisdiction under the existing ALP-based profit allocation rules. For example, if a market jurisdiction made an upwards adjustment to the profits it was allocated for marketing and distribution activities, it would need to recalculate whether a MNE group would be eligible for the safe harbour, with any additional tax due under the existing profit allocation rules being offset against the tax that would no longer be due under Amount A.

541. Under the marketing and distribution profits safe harbour, the formula for calculating Amount A would remain unchanged. However, it would be necessary to determine a fixed return for in-country routine marketing and distribution activities. This will raise a number of technical and design challenges that will need to be addressed within the context of the safe harbour.

542. The fixed return would not necessarily seek to replicate an arm’s length return, nor would it limit the profits allocable to marketing and distribution activities. Jurisdictions that are entitled to a higher return for the performance of marketing and distribution activities under the ALP would continue to be entitled to this return. Instead, this fixed return (which would only be relevant for large in-scope MNE groups) would act as a test to identify situations when allocating Amount A to a market jurisdiction would give rise to double counting.

543. For example, if the fixed return were set at a return on sales of 4%, it would mean that any profits allocated to a market jurisdiction in excess of this return would be deemed to duplicate the return allocated to a market jurisdiction under Amount A. So, if a market jurisdiction were allocated a 3% return under existing profit allocation rules it would receive a full allocation of Amount A, but if it were allocated a 5% return, its allocation of Amount A would be reduced by 1% (the difference between the return it receives under the existing profit allocation rules and the fixed return). If a lower fixed return were agreed, the safe harbour would apply to cap the allocation of Amount A more frequently than if a higher return were agreed.

544. The fixed return could be computed in a variety of different ways, but perhaps the simplest approach would be to agree a single fixed return on sales that would be applied to in-scope locally sourced revenue (as determined under the revenue-sourcing rules). This approach would draw on an existing component of the Amount A formula and would therefore avoid some of the challenges that would arise under a different approach, specifically in defining the base (e.g. sales or costs) to which a margin or mark-up would need to be applied.

545. Although the fixed return would not seek to be consistent with the ALP, it could be agreed that the fixed return could vary by region or industry (but probably not based on functions). For example, it may be argued that as pharmaceutical distributors are typically allocated a higher return under the ALP, the fixed return for pharmaceutical businesses could also be higher under the safe harbour. This would mean that a pharmaceutical business would need to allocate a higher return to a market jurisdiction under the existing profit allocation rules than a comparable business in another sector to benefit from the safe harbour. However, for simplicity it could also be agreed that there should be a single fixed rate applicable across all regions and industries.11

546. Hence, further consideration of this safe harbour will require additional work on a number of challenges, including for defining the fixed return for routine marketing and distribution activities.

547. The development of a “domestic business exemption" to reduce the instances of “double counting” is also considered, together with the mechanism to eliminate double taxation and the safe harbour. There are two potential types of domestic business exemptions. The first, and simplest, would exclude from the scope of Amount A large, domestically-focused business with a minimal level of foreign income. This would be implemented through the exemption of groups whose foreign source in-scope revenue falls below an agreed threshold from the scope of Amount A (see above section 2.3.2).

548. The second, more complex exemption, would seek to exclude from Amount A part of a group’s business that is primarily or solely carried on in a single jurisdiction. This may occur for instance, where a group acquires a business operating in another jurisdiction that it does not subsequently integrate within its broader operations. In this scenario, it may be difficult to justify why Amount A should apply to this portion of a group’s business, as it could result in the residual profits from this business, which are demonstrably only derived from one jurisdiction, being reallocated to other jurisdictions around the world. It could also result in residual profits generated from other parts of the business being allocated to the jurisdiction in question; despite the fact that the said jurisdiction already has taxing rights over the residual profits (to the extent they arise) associated with the relevant sales. Another possible view is that this is simply a logical consequence of the formulaic nature of Amount A applicable to the group as a whole.

549. The “domestic business exemption” would address some instances of double counting by excluding from Amount A profits derived from the sale of goods or services that are developed, manufactured and sold in a single jurisdiction.

550. There are two challenges that would need to be overcome to develop this domestic business exemption. First, it would be necessary for a taxpayer to isolate and segment out the profits of this standalone domestic business from the other activities of the group. This would require a remodelling of the segmentation framework that would increase complexity and the associated compliance costs and administrative burden. That said, on the assumption that the business is operated on a standalone basis, it may be relatively easy for the taxpayer to perform this additional segmentation.

551. Second, it is unlikely that there are many examples of MNE groups with completely standalone domestic businesses. This is because in most instances these businesses will be integrated to some extent in the broader activity of the group, whether through shared development of intellectual property (IP), intragroup financing activities, or other central services. For large CFB in particularly, royalty payments in relation to IP may be a significant expense in many market jurisdictions. Even where groups manufacture and sell goods in a single jurisdiction using local IP, these goods may include input purchased from a related party in another jurisdiction or produced using manufacturing know-how for which a licence fee is paid.

553. Setting aside these challenges, it is also important to emphasise that the “domestic business exemption” would only reduce the occurrence of “double counting”. For example, it would not address situations where a market jurisdiction had taxing rights over the residual profits arising from the activities of a distributor not eligible for the “domestic business exemption”. Therefore, the “domestic business exemption” could only be developed in combination with another mechanism to address double counting.

554. As a first next step, drawing on the data and the analysis prepared as part of the impact assessment of different percentages for the profitability threshold (step 1) and the reallocation percentage (step 2), a decision of the Inclusive Framework members will be necessary to determine the quantum of Amount A, including whether the formula should incorporate any “differentiation mechanism”. Relevant considerations in this discussion will include, for example, the amount of residual profit to be reallocated (including proportionality to compliance costs and administrative burden), the possible impact on this amount of residual profit of accounting for profit shortfalls (see section 5.4.6), and the number of MNE groups impacted. There will also be some remaining issues around questions of regional and jurisdictional segmentation.

555. In addition, further work will be required to assess the different options to deal with double counting (and their possible combinations), in close coordination with the work on the mechanism to eliminate double taxation (see Chapter 7). This will also include consideration of the interactions between Amount A and certain withholding taxes collected by market jurisdictions. For the safe harbour, issues where further work is required include:

• Assessing the implications of considering the income taxes payable in the market jurisdiction under existing taxing rights and Amount A together (e.g. whether marketing and distribution profit allocated to a market jurisdiction under the ALP in excess of a fixed return may be seen as duplicative with Amount A);;

• Defining the fixed return for routine marketing and distribution activities, including determining whether this return should vary by industry and/or region;

• Identifying and isolating the profit from marketing and distribution activities in the market jurisdiction that are covered by the safe harbour; and

• Developing a mechanism to deal with transfer pricing adjustments, lagged threshold permanent establishment claims or denial of deduction for shared costs that are made after the safe harbour has been applied in the market jurisdiction.

• Considering how common/prevalent the double counting issue is, and whether the practical and administrative challenges in designing the safe harbour are commensurate with this double counting issue; and

• Clarifying the treatment of withholding taxes collected by the market jurisdiction.

## Notes

← 1. Where a market jurisdiction is allocated profits for other activities, e.g. manufacturing, or marketing and distribution activities relating to out-of-scope revenue this would not be taken into account for the purposes of the safe harbour.

← 2. For the purpose of calculating and applying Amount A, further work will be required to define the term “revenue” (see paragraph 515).

← 3. see CTPA/CFA/WP2/NOE2(2020)10).

← 4. see CTPA/CFA/WP2/NOE2(2020)10).

← 5. For example, if a segment had total in-scope revenue of 80 (all of which was sourced to eligible market jurisdictions under the nexus rule) and 100 total revenue (including out of scope revenue), the allocation key would mean that only 80% of the Amount A tax base would be allocated to market jurisdictions under Amount A. The 20% would remain unallocated because it relates to revenue derived from out of scope activities.

← 6. The calculation of Amount A already requires MNE groups to determine the total revenue used in computing the Amount A tax base (and where relevant attributable to each segment). Hence, only data on revenue sourced in one eligible jurisdiction would need to be verified to calculate a market specific Amount A tax liability under a throwback system. In contrast, under a throw-in system, the calculation would require verification of the revenue sourced in all market jurisdictions.

← 7. Business models for MNE groups in ADS industries typically entail early and ongoing centralised development of intellectual property (IP) and the incurring of other costs aimed at developing the service offering as a whole, which then may be rolled out to new markets at limited marginal cost. Consequently, significant costs are centralised and the variation of profit between regions and jurisdictions is materially affected by the allocation of those costs to entities that operate in the various market jurisdictions and that benefit from the initial and ongoing development. This implies that regional or jurisdictional differentiation will be less relevant for ADS businesses.

← 8. As noted above, the profitability threshold of the Amount A formula is designed to limit interactions between Amount A and existing taxing rights of market jurisdictions based on the ALP.

← 9. It could also be argued that the interactions between Amount A and some withholding taxes could give rise to double counting, i.e. that a market jurisdiction would tax twice the same item of profit if they were allocated Amount A on top of certain existing withholding tax liabilities. For example, assume an entity in the MNE group outside the market jurisdiction is receiving a royalty payment from an entity in the market jurisdiction for the use of branding or licensing rights in respect of sales in that jurisdiction. Assume the royalty is subject to withholding tax in the market jurisdiction. If the royalty is contributing to residual profits, then the market jurisdiction can be seen as already taxing a share of this residual via the withholding tax. It is important to emphasise that members have different views on this issue, some consider that this interaction could give rise to double counting, whilst others argue that it does not.

← 10. Where a market jurisdiction is allocated profits for other activities, e.g. manufacturing, or marketing and distribution activities relating to out-of-scope revenue, this would not be taken into account for the purposes of the safe harbour. Further rules may be required to determine the existing marketing and distribution profit (e.g. book-to-tax adjustments).

← 11. The pharmaceutical industry typically has higher returns than most other industries and so even under a single fixed return approach, the total safe harbour return (i.e. Amount A plus the fixed return) would, when compared to other industries, still be relatively high for most pharmaceutical groups.