Latest NEWS &
INSIGHTS

Valentiam Group: A Response to OECD Secretariat Proposal for a Unified Approach Under Pillar One

Valentiam Group: A Response to OECD Secretariat Proposal for a Unified Approach Under Pillar One

Preface

Today, the OECD released public comments that it received on the Secretariat Proposal for a "Unified Approach Under Pillar One." Among these public comments were those submitted by Clark Chandler of Valentiam Group. Clark's unedited comments are presented in full below.

Comments

I am submitting these comments in response to the OECD’s request for comments in its public consultation document Secretariat Proposal for a “Unified Approach Under Pillar One,”[1] In making my comments, I am accepting that the OECD has made the policy decision to adopt the proposed Unified Approach and will not address whether this is or is not appropriate. I do note that the proposed Unified Approach is a radical departure from the existing system, in that it combines a formulary system, which is used to determine the minimum taxing rights of market countries, with the existing transfer pricing rules, which will be used to determine the allocation of the right to tax the remaining profits.

1) Summary Description of Proposed Unified Approach

The OECD has laid out a Unified Approach that, as I understand it, would essentially split the right to tax corporate profits into three discrete baskets, as follows:

  • an “Amount A” which would be a share of excess profits allocated by formula to market jurisdictions, provided that these profits are based on sales to consumers. The formula that would be used to do this is:

    Amount A = total profits – hypothetical routine profits = residual profits * local source share of residual profits.

    The rationale for this taxing right presumably is that the MNE’s excess profits are dependent upon an MNE’s access to consumers in each local country. This taxing right will exist even if the MNE has no functions, assets or risks with respect to sales in the market jurisdiction.

  • an “Amount B” which would be the income that a local market economy would be entitled to for “… baseline marketing and distribution functions that take place in the market jurisdiction”. The OECD suggests that Amount B could possibly be determined using a fixed remuneration, which would in effect exempt it from existing transfer pricing rules.

  • an “Amount C” which at different points in the OECD document appears to be either the residual that is left after deducting Amount A and Amount B from total profits or as a “top up” that would be due to the market jurisdiction due to functions, assets and risks that are over and above its baseline marketing and distribution activities. Amount C would be determined under existing transfer pricing rules, which should be accompanied by a robust legally binding and effective dispute resolution system. In my subsequent discussion, I have distinguished between Amount C(t) which reflects the total amount left after deducting Amount A and Amount B from total profits and Amount C(m) which reflects the top-up that may be due to the marketing jurisdiction due to incremental functions, assets and risks carried out be legal entities located in the market jurisdiction.

A market jurisdiction would therefore be entitled to tax the sum of Amount A, Amount B and Amount C(m).

2) Selected Key Issues

I would like to make 5 basic key points as follows:

  • While the OECD suggests that the formula for computing the base for the new tax is akin to a residual profit split (“RPSM”) carried out in a traditional transfer pricing analysis, this is not correct. The share of profits attributed to each legal entity in a traditional RPSM is based (at least conceptually) upon the functions, assets and risks carried out by the legal entities involved. There are no functions, assets or risks that can be used to determine the share of profits attributable to Amount A – a traditional RPSM would assign zero profits to an entity with no functions, assets or risks, and therefore would assign zero profits to Amount A. As a practical matter the share of profits assigned to Amount A is simply a policy decision akin to a decision to set a tax rate at x% or y%.
  • The OECD’s proposed approach for computing Amount A appears to have attributes in common with both a tax on revenues and a tax on profits. Computing Amount A as, in essence, a profit-limited tax on revenues is likely to be substantially simpler than treating it as a tax on profits attributable to the market jurisdiction. While this may be a reasonable policy decision, such an approach is very different from a true tax on the profits earned by an MNE in each country and may have unintended consequences.
  • The OECD has suggested that Amount B could be determined as a fixed amount (e.g., a fixed cost plus or margin percentage). This would be equivalent to establishing a safe harbor amount. While there are many advantages to safe harbors, there are issues with how these formulary computations interact with existing transfer pricing rules when multiple countries have transactions with the same market economy. Moreover, past experience suggests that it is very difficult to get international agreement on such safe harbors.
  • The OECD’s approach to determining nexus appears to focus solely on the right of a market economy to collect Amount A; it completely ignores the question of who will be responsible for paying Amount A. Determining a right to collect a tax without determining who is responsible for paying the tax answers only half of the question that will be faced by MNEs that are subject to the tax and leaves the question of which parties will be involved in any transfer pricing disputes unanswered.
  • The OECD’s discussion of the dispute resolution mechanism appears to focus on the procedural aspects of such dispute resolution. While this is clearly an important topic, the two more basic questions are (a) which legal entities/countries will be involved in the dispute and (b) what substantive rules will be used to resolve the dispute in a principled way. The answers to both of these questions depend upon the interaction between the formulary approach used to determine Amount A (which essentially ignores existing legal entity structures) and the approach used to determine the profits that legal entities will earn from Amount C(t) (and possibly Amount B), which are dependent upon existing transfer pricing rules and therefore which are dependent upon existing legal entity structures.

3) Computation of the Share of Profits Assigned to Amount A

As presented, it appears as though the OECD proposal for determining Amount A could be determined relatively simply in a way that is linked to existing transfer pricing rules. The computation of Amount A could start with EBIT (or pre-tax?) profits as reported in an Annual Report/Form 10-K, deduct a notional measure of routine profits, and then multiply such residual profits by a set percentage of those profits. The OECD suggests that the formula for computing the base for the new tax is akin to a residual profit split carried out in a traditional transfer pricing analysis.

This is not really correct. A traditional transfer pricing profit split is based (at least conceptually) upon a set of variables related to functions, assets and risks carried out by the legal entities involved. There are no functions, assets or risks – or for that matter any other objectively measurable variables – that can be used to determine the allocation of profits between Amount A and Amounts B and C. This is clear from the fact that the use of a profit split method based on traditional transfer pricing principles would assign zero profits to an entity with no assets, functions or risks, and therefore zero profits to Amount A. Therefore, as a practical matter, the share of profits assigned to Amount A will be set globally based on a policy decision. While there may be nothing wrong with setting values based on policy considerations (tax rates are set in this way), there is no reason to expect a logical relationship between the factors used to determine Amount A and traditional transfer pricing considerations.

It is of course possible to view a country’s customers as an asset of the country rather than of a specific legal entity, much in the same way that natural resources or radio/TV bandwidths are a local asset. The OECD’s desire to impose a tax/royalty on the right to such access would normally be done through a tax on revenues, which are less difficult to measure than profits. The current digital taxes that are under consideration in various countries are, in this regard, generally structured as a tax on revenue. However, the level of tax imposed is relatively arbitrary to start with, and there is no straightforward way that I am aware of to translate such a tax on revenue into an equivalent tax on profits. [2]

4) Does Amount A Tax Revenues or Profits?

As noted above, the most common way of charging a tax or royalty for access to a resource – such as access to consumers – is to compute the charge as a percent of revenues. However, the OECD is presumably concerned that a revenue-based tax would be borne by local consumers rather than by the MNE, and therefore wishes to compute Amount A as a share of profits.

Profits are a much more complex tax base than revenues, however. Some of these additional complexities include:

  • Measuring profits is more complex than measuring revenues. The OECD has suggested that corporate global financial results (e.g., such as those reported in a Form 10-K for a US-based MNE) could be used to measure profits. This has the advantage of using an existing definition of profits and basing profits on numbers that are audited for financial statement purposes. However, profits as measured in financial statements are prepared for financial statement reporting rather than tax purposes, and therefore do not include some considerations that are common in tax reporting (e.g., treatment of NOLs) and may have other attributes that impact income in ways that may not be tax appropriate (e.g., the treatment of intangibles that arise as a result of a business acquisition.) They may also lack the detail needed to get accurate results by geography or line of business.
  • Linking profits to a specific source is more complex than linking revenues to a specific source. Companies often/generally track third party revenues at a fairly detailed level and can distinguish between revenues associated with game boxes and the game software used in such boxes. Profits, however, are dependent upon allocations of expenses as well as revenues, and may in fact not be tracked at that level if the two are viewed as an integrated business. There may also be synergies between sales made to consumers and to business customers. The use of profits rather than revenues may therefore make it more difficult to draw clear boundaries as to which income should be included in Amount A and which should not.
  • Profits are highly dependent upon the specific boundaries to an MNE’s business, and the way in which such boundaries are set often varies significantly from one company to another. As a simple example, Pharma Co. X holds a patent and the brand name for Compound X which it sells to consumers through a related distributor. Pharma Co Y holds a patent to Compound Y and sells it to unrelated Pharma Z, which in turn sells it to consumers. It would appear that, under the Unified Approach, all of the profits of Pharma X would be used in computing Amount A (e.g., both patent and brand value) but that for compound Y only the brand profits would be used in Computing Amount A.[3]
  • Profits are generally computed annually and can only be determined after the close of the fiscal and/or tax year. The information needed to compute a profit-based tax is therefore generally not available until after the fiscal year books have been closed.

Other complexities exist as well.

The OECD’s proposal for addressing these complexities is to suggest that Amount A could be computed based on aggregate profits, with the resulting aggregate amount of Amount A then allocated among different market jurisdictions based on revenues. This represents a hybrid approach that has attributes in common with both a tax on revenues and a tax on profits, and which is perhaps best thought of as a profit-limited tax on revenues with a rate that varies by taxpayer. In essence, the tax as a percent of revenues is equal to residual profits as a percent of revenues times the share of profits assigned to the market country times the tax rate of the market country. If residual profits are 20% of revenues, the share assigned to the market country is 25% of revenues and the tax rate is 20%, the tax rate as a percent of revenues would be equal to 20% * 25% = 5% * 20% = 1%. The tax rate would vary both by taxpayer (due to differences in profitability) and by country (due to differences in the tax rate).

The use of profits as the starting point in computing the tax base raises conceptual policy issue that would not exist with a tax based on revenues. One key question is whether a country’s right to collect Amount A will be dependent upon whether the MNE realizes profits in that country or simply profits in aggregate? It is relatively common for even highly profitable companies to realize substantially different profit margins by business line or country (or even high profits in one country while incurring losses in others). Starting with overall profits and allocating based on revenues will shift taxing rights on profits that arise due to sales in one country to another country; basing Amount A on the profits attributable to sales in a given country will obviously increase the amount of financial data needed and will require data that may or may not be tracked in an MNE’s existing records.

This is illustrated in Table 1 below. In Table 1, the MNE has equal sales in County X and Country Y but earns substantially higher profits in Country X than in Country Y.[4] Residual profits are determined by deducting routine profits of 5% of sales from residual profits. Amount A is assumed to be equal to 25%[5] of residual profits. If Amount A is computed based on overall profits, and then allocated evenly between Country X and Country Y based on sales, both Country X and Country Y will be able to tax an equal amount of profits. However, if the tax is based on country-specific profits, Country X will be able to tax the entire amount.

Table 1: Alternative Approaches To Computing Amount A

 

Country X

Country Y

Total

Sales

100

100

200

Profits

35

5

40

Routine Profit @5%

5

5

10

Residual Profits

30

0

30

Share Subject to Amount A

25%

   

Amount Subject to Tax

 

 

 

Based on Overall Profits

3.75

3.75

7.50

Based on Country Specific Profits

7.50

-

7.50

Computing Amount A as, in essence, a profit-limited tax on revenues is likely to be substantially simpler than treating it as a tax on profits. However, such an approach is very different from a true tax on the profits earned by an MNE in each country and may lead to a substantially different distribution of the profit that is taxable in the different market countries in which the MNE operates. There is also the fundamental question of whether it make sense for a profit-based tax to generate tax revenues in a country where the taxpayer is actually incurring low profits or even losses.

 Finally, given the lack on any consideration of intercompany transactions, there are some cases in which a profit limited tax on revenues generates results that are simply bizarre. Take an MNE that operates in two market countries, Country X and Country Y. There are no intercompany transactions between Country X and Country Y. The MNE realizes residual profits of 10% of sales in Country X but earns no residual profits in Country Y. However, Amount A is computed by adding the residual profits earned in all countries, which in this case is equal to the residual profits earned in Country X, and then allocated between the two countries based on revenues. If the countries are of equal size, Country Y gets to tax an Amount A that is equal to 5% of its sales, and Country X gets to tax only half of the residual profits generated in its jurisdiction. There seems to be no logical reason for allowing the tax authorities in Country Y to collect taxes on profits that were generated in Country X given that these profits are in no way linked to customers located in Country Y. While this is an extreme example, it suggests that computing Amount A based on aggregate profits and then allocating the right to tax Amount A among market countries based on relative revenues may lead to unexpected outcomes in the case of products where there are only limited intercompany transactions or where IP is not centrally owned or managed.

5) Computation of Amount B

The OECD suggests that Amount B could be determined based on a “fixed” remuneration, possibly one that varies by business. Such a fixed remuneration is largely equivalent to a proposal to establish a safe harbor, which has a lot of appeal from the perspective of administrative simplicity and certainty. Moreover, as has been suggested in earlier comments to the OECD, Amount B could be adjusted to reflect differences in the profitability of an MNE, thereby allowing the market country to share in residual profits.

However, an Amount B generated under a formulary approach would still have to interact with existing transfer pricing rules. First, to the extent that the sales giving rise to Amount B are the product of intercompany transactions, then Amount B will, as a practical matter, be reached through adjustments to transfer prices. If there is a single controlled counterparty with sufficient profits, this may not present an issue. The situation is more complicated if this single counterparty does not have sufficient profits to pay the Amount B implied by the formulary calculation without incurring losses – particularly if these losses are caused or are increased by an upward adjustment to Amount B based on the MNE’s global profitability. Finally, if there are multiple counterparties (e.g., one that supplies products and one that charges a royalty), then there is the question of which transfer price is adjusted in order to obtain the Amount B mandated by the formulary calculation.

There is also the question of whether the formulary calculation of Amount B would be limited to situations in which there is an intercompany transaction, and whether it would apply in cases where the intercompany transaction has only a limited impact upon profits. This could happen, for example, if an MNE owns (1) a highly profitable set of apparel stores in country X that buy from unrelated suppliers and sell under brand X and (2) a set of loss-making apparel stores in Country Y that also buy from unrelated suppliers and which sell under Brand Y. If there are no intercompany transactions, there is no transactional mechanism for creating the Amount B computed under the pricing formula. While the extreme cases of this may be unlikely, the question of whether an MNE should be responsible for losses in a country that have little or nothing to do with intercompany transactions is relatively common, and therefore there is the open question of whether a formulary approach to computing Amount B would be applied in such cases.

My other concern with this proposal is a practical one. The OECD has solicited substantial comments on safe harbors in the past but has not been able to implement widely accepted safe harbors. The OECD is likely to face the same difficulties in determining a fixed amount for Amount B as it has had in setting safe harbors. Adding profit-based adjustments would add to difficulty in obtaining consensus on the formula for determining Amount B. Given the difficulties in reaching agreement on safe harbors, would it be more practical to reach a consensus that Amount B should be determined based on current transfer pricing rules?

6) The Determination of Nexus Should Include an Identification of the Legal Entity (ies) Responsible for the Payment

The OECD appears to be focused on simply determining whether the MNE has nexus in a given market jurisdiction and does not provide any discussion of which parts of the MNE group will have such nexus. While I understand the point that the obligation to pay tax on Amount A is based on the benefits/profits of the MNE group, the income subject to tax will be income that is recorded by a specific legal entity or combination of entities. The provision of specific guidance on the attributes of the specific legal entities that are liable for tax on Amount A, and what happens if (for example) these entities do not have the requisite income, is a necessary first step both for the practical implementation of the proposed unified system and in the design of any dispute resolution mechanism. As a simple example, can a legal entity that owns IP that is limited to Asia have nexus with respect to an MNE’s sales to a market jurisdiction located in Europe? The issue of which legal entities are potentially subject to tax has to be addressed at some point, and this appears to be a basic question of nexus.

This issue arises because the proposed Unified Approach is a mix of a formulary approach that essentially ignores legal entity structure/intercompany business arrangements and an arm’s length approach that is highly dependent upon legal entity structure/business arrangements. Therefore, while it may be possible to determine Amount A for any given market country without any reference to the legal entity structure of the taxpayer, generating an actual payment for Amount A requires determining the identity of the payor.

I have developed a simple example to illustrate the issue. The first key question is to identify the location of profits before the computation of Amount A. As I understand it, there is no proposal to change the role of current transfer pricing in this regard – e.g., there is no suggestion that all profits over and above Amount B profits will be reported by the parent company and then allocated first between Amount A and Amount C(t), with Amount C(t) then being allocated to various legal entities. Given this, current transfer pricing rules will still be used to determine which specific legal entities report operating profits.

With this in mind, we can go through a simplified set of steps to split an MNE’s total profits between Amounts A, B and C(t). This is done in Table 2.

Table 2 starts with third party revenues, which are assumed to be earned in two market countries, one of which has a legal entity with substance and the second of which has customers, but no local substance under current rules. Profits are earned in four of the five countries: ParentCo in Country A, a legal entity in Country B that owns IP, a manufacturing entity in Country C, and the marketing entity with local substance in Country D. For simplicity, I am assuming that Amount A is calculated in aggregate by deducting notional routine profits of 5% of sales from actual profits, with 25% of these residual profits allocated to Amount A.

Table 2: Example Showing the Calculation of Amount A

 

Country A

Country B

Country C

Country D

Country E

Total

 

ParentCo

IPCO 1

(MfgCo)

Dist w Substance

Dist w/o Substance

 

Sales

 

 

 

100

100

200.0

Profits

20

20

15

5

0

60.0

Routine Profits for Computing Amount A (5%)

 

 

 

 

 

10.0

Residual Profits

 

 

 

 

 

50.0

Share Subject to Amount A

 

 

 

 

 

25%

Amount A

 

 

 

 

 

12.5

Amount B

 

 

 

5

0

5

Amount C(t)

 

 

 

 

 

42.5

Once Amount A has been calculated (and regardless of how the right to tax Amount A is split between the two market jurisdictions), the key question is where does Amount A come from? In Table 2, there are three (really four) options: the ParentCo, the IPCO, the MfgCo, and potentially the distributor with substance. While I have used terminology that suggests that Amount A should come from IPCO, it is certainly easy enough to imagine scenarios where the answer is not obvious – ParentCo may own the brand while IPCO owns technology; arguably the amount of profit is a result of synergies among the intangibles owned by ParentCo, IPCO and MfgCo, and the tax authorities of IPCO may therefore contend that Amount A should be paid pro-ratably by each of the three legal entities.[6] It is also possible that the full amount of Amount A should logically come from IPCO, but that IPCO does not have sufficient profits to pay Amount A.[7]

There are at least two broad options for determining which legal entities will have a nexus/a taxable presence in a given market jurisdiction. One option is to adopt a formulary approach that essentially says that Amount A will be paid pro-ratably by legal entities based on their share of residual profits. An alternative option would be to look at which legal entities benefit from access to a specific market economy, and to link the payment obligation to benefits. The first option has the benefit of simplicity and the disadvantage of arbitrariness – why should a legal entity that owns IP related to product sales pay tax on an Amount A that is based largely on profits generated through advertising revenues based on the use of a search engine? The latter has the advantage of linking tax obligations to benefits/the taxable income that gives rise to such tax obligations but has the disadvantage of requiring criteria for identifying benefits. In some cases, this may be relatively simple (e.g., a legal entity that owns Asia-specific IP benefits from profits in Asian market countries but not European or Americas market countries). In other cases, the determination may be quite complex or even subjective – if one legal entity owns trademark and marketing IP and another owns patents and technology IP, which benefits more from access to customers in a market country?

Finally, there may be a circular relationship between the transfer pricing payments and the obligation to pay Amount A. Consider a scenario in which a manufacturing entity pays a royalty to an IP entity that is based on a 50:50 split of residual profits. If the IP entity is responsible for paying all of Amount A, then the tax authorities of the IP entity will be shortchanged – they will have received a royalty that is only half of total residual profits, but will have had to pay all of the tax on Amount A collected from that income, and they presumably will have no right to tax the profits allocated to Amount A. This issue could presumably be resolved by having a royalty payment equal to 100 percent of residual profit, but this may be inconsistent with the functions, assets and risks of the manufacturing and IP entities. The example presented in not a particularly complex one but does illustrate the point that allocating a share of profits to Amount A based on a formulary computation is likely to have significant knock-on effects on the detailed application of existing transfer pricing rules, particularly those that use profits as a starting point for determining intercompany pricing.

7) Dispute Resolution

The OECD’s discussion of the dispute resolution mechanism focuses on the procedural aspects of such dispute resolution. While this is clearly an important topic, the two more basic questions are (a) which legal entities/countries will be involved in the dispute and (b) what rules will govern the principled resolution of the dispute? The answers to both of these questions depend upon the interaction between the formulary approach used to determine Amount A (which essentially ignores existing legal entity structures) and the approach used to determine the profits that legal entities will earn from Amount C(t) (and possibly Amount B), which are dependent upon existing transfer pricing rules and therefore which are dependent upon existing legal entity structures.

The only disputes that the OECD discusses specifically are those that may arise if tax authorities in a market jurisdiction concludes that, based on the functions, assets and risks of the taxpayer in its market, that the market jurisdiction is entitled to a share of Amount C(t). The participants to such disputes, and the transfer pricing rules governing such disputes, will presumably be based on specific transactions between legal entities and are similar to those that arise under existing transfer pricing rules.

However, transfer pricing disputes may also arise when it is unclear as to which legal entity is obligated to pay Amount A. As a simple example, the taxpayer may pay Amount A from legal entity X in Country X because legal entity X owns marketing intangibles. The tax authorities of Country X may determine that Legal Entity Y in Country Y should have paid some or all of Amount A because Legal Entity Y owns the relevant technology, and the taxpayer has reported that its profits are highly dependent upon technology.

The need to identify a specific legal entity(ies) that are responsible for paying Amount A raises two key questions that have to be addressed under the Unified Approach that are different from the questions that have had to be addressed under current transfer pricing rules:

  • who are the parties to the dispute? If the dispute is over the size of Amount A, one or both of the two marketing entities will be involved. But if Amount A is computed based on a formulary approach that is not subject to challenge, then the two distribution entities may have no interest in the dispute and (referring back to Table 2), the potentially interested parties will be ParentCo, IPCO and MfgCo.[8] Moreover, as the payment of Amount A is not linked to any specific transactions, there is no particular presumption that the MAP dispute would be limited to just two countries.[9]
  • What conceptual “rules of the road” would guide the settlement of the MAP disputes? Is there a presumption that Amount A would be paid by the owner of marketing intangibles? Of digital technology intangibles? Or is there a presumption that all parties earning residual profits would be equally responsible for paying Amount A? What is the role of geography and business lines – is an IPCO holding Asian IP rights liable for the payment of Amount A to non-Asian marketing jurisdictions?

Note that the fundamental questions of who will be involved in the MAP disputes and what conceptual rules will govern the settlement of the MAP disputes have to be resolved before the procedural issues identified by the OECD come into play. To cite one example, it is it hard to determine who the parties to an APA should be until we know which parties are likely to have competing claims on right to tax specific income. If Amount A and Amount B are determined by fixed formulas, the potential disputes will be among countries that are liable to pay Amount A, and not among countries that tax Amount A. Therefore, there is no need to involve countries who only have the right to tax Amount A and/or Amount B. Moreover, determining which countries may have competing and overlapping claims on Amount C(t) may be difficult to determine given that the dispute will be around which country owes Amount A, and this is determined without reference to specific transactions among different legal entities.[10] Is it, in fact, even possible to negotiate an APA between two countries when there are no transactions between these two countries?

As a second example, while binding Arbitration will in theory guarantee relief from double taxation, the results of such arbitration will be inherently arbitrary unless the conceptual rules governing how the resolution should be determined are clearly defined. Absent guidance, one arbitrator may decide to use a proportional allocation to compute the share of Amount A owed by different legal entities while another may conclude that the owner of one type of IP is primarily responsible for the payment of Amount A. The interested parties to the arbitration will have no basis for reaching a reasonable settlement unless they know the rules that will be used by the arbitrators.

Download Now: A Work Plan For Meeting OECD BEPS Requirements

Endnotes:

[1] These comments reflect my views alone, not those of Valentiam.

[2] Traditional valuation approaches used in financial statement valuations often do assign a share of profits to customer intangibles. However, this is most commonly done by computing the value of customer intangibles as a residual, and such an approach requires a detailed economic analysis of the facts and circumstances giving rise to customer intangibles.

[3] The OECD has indicated that it would like to get consistent results regardless of the business model, and so in theory may want to include the profits earned by third-party suppliers/IP holders in its computation of Amount A. However, these profits could not be used by the company that is selling the product, as it would not have access to data on the profits of a third-party supplier/IP company. Alternatively, the third-party supplier/IP holder could have a separate nexus and would compute a separate and additive Amount A. This, however, assumes that the third-party supplier has a detailed knowledge of which countries the unrelated buyer of its product is selling into, and the specific value of such sales. But this information is generally not shared by the buyer.

[4] This result could occur because of either country-specific differences in profitability for sales in a single business line or because profits vary by business line with differences in the relative importance of business lines by country.

[5] I am using this purely for illustrative purposes; I am not recommending the use of this or any other specific percentage.

[6] If Amount A is computed in aggregate and then allocated among the different market countries based on sales, it is even possible to imagine a scenario in which some market countries may be responsible for paying tax on Amount A to other market countries. In a simple two-country example where both countries are market countries and where one has substantial residual profits and the other has losses, an approach that involves first determining Amount A on an aggregate basis and then allocating it based on revenues would leave the country with losses entitled to tax some of the profits earned by the taxpayer in the other country.

[7] This situation would occur whenever IPCO’s profits were less than $12.5 in the example shown.

[8] Moreover, if one distribution entity was highly profitable and the other was not, it is even possible that Amount A would require a reallocation from the high profit marketing entity to the low profit marketing entity.

[9] I have relatively little background with tax treaties, but I have to believe that this raises a number of issues with respect to such tax treaties, starting with the basic question of how Competent Authorities get together when there may be no transactions between/among their countries, and how would the simply mechanics of a MAP agreement be handled when the issue is that Country X collected $100 in taxes from a CFC in Country Y but the Competent Authorities of Countries Y and Z agree that Country X should have collected just $30 in taxes from the CFC in Country Y and $70 in taxes from the CFC in Country Z?.

[10] Simply requiring a tax payment based on Amount A without stating which legal entity(ies) is responsible for this payment is the worst of all possible worlds as it is impossible to get double tax relief through a MAP process without identifying the entities involved. This is acknowledged in paragraph 36.

Topics: Transfer pricing, Digital services tax, Pillar One, OECD, BEPS, international tax, tax, Pillar Two, BEPS 2.0, Action 1, Secretariat Proposal, Unified Approach, DST, Formulary Apportionment, Inclusive Framework