Arm’s length principle mutations: control of risk in the OECD guidelines and variations in practice

By Andrew Hickman, UK consultant and former OECD Transfer Pricing Head

In this note, I reflect on interpretations of the 2017 OECD transfer pricing guidelines that have come to my attention during the four years or so since the guidance has been used in practice.

The reflections focus on the challenges facing the arm’s length principle in dealing with risk and the resulting emphasis on control of risk in the guidelines.

I suggest that mutated versions of the arm’s length principle are being adopted in practice and observe the rise in the use of control functions and concepts of value creation as relative profit-splitting factors to bypass transactional comparability.

It is to be expected that users of the guidelines will tend to be receptive to the guidance that favours their position and tend to overlook guidance that runs counter to that position (and I am not immune to the charge of selective readings).

The scope for different views may be an inevitable consequence of the non-prescriptive nature of the guidelines or may suggest that the guidance could have been better explained.

Nevertheless, some interpretations of the guidelines do not seem to be based on what the guidance says or seem to ignore what the guidance does say. 

I begin this note with a framing illustration of the different approaches. I then set out the differences between the version of the arm’s length principle I believe was envisaged in the OECD transfer pricing guidelines and the mutated versions being observed in practice.

In the final part of the note, I discuss whether the mutations of the arm’s length principle are responding to real or perceived difficulties in using the guidance in practice and question whether the arm’s length principle is adapting or whether these mutations should be eradicated.

A framing illustration

One example that frames the issues and summarises the differences I observe between how the OECD transfer pricing guidelines suggest control over risk should be applied and mutated versions relates to the treatment of research and development activities. 

The transfer pricing guidelines frequently discuss research and development activities because they epitomise the separation of capital and risk from activity, depending on the terms of the contractual arrangements, and the potential for significantly different levels of profits or losses being recorded by the transacting parties over time.

Research and development activities also exemplify tensions in applying the arm’s length principle in practice, if it seems that the arm’s length principle is not producing the ‘right” answer in terms of perceptions of relative functions performed by the parties or of relative value apparently created by the parties.

Assume that the contractual arrangements indicate that Company A engages Company B, an associated enterprise of the ABC Group, to provide specialist research services. The ABC Group generates profits from the sale of products based on intellectual property developed within the group and acquired from third parties.

Annual spend on research and development activities is significant, and the group makes regular acquisitions to supplement its in-house technology capabilities.  Responsibility for the group’s development strategy is delegated by its board to its technology committee. 

Of the six members of the technology committee, four are board members, of which three are employed by Company A, while one, the group chief technology officer, is employed by Company B.  

Company A acts as the group’s head office, employs most top management, provides support services to the group, and does not conduct research and development activities; in total, it employs 500 personnel. Company A licences its technology to two associated enterprises that share responsibility for global manufacturing and sales of products.

Company B is the largest of the group’s research and development centres and employs 1000 personnel, representing 60% of the group’s research and development staff. 

Assume further that under the strategic direction and capital allocation set by the board, the technology committee sponsors development areas and prioritises projects within each area, reviews project plans designed by the relevant specialists in the group’s research and development centres, and commits funding.

Committed projects are then managed and conducted with considerable discretion as to their design by the specialists in the research and development centres, who manage a range of their own operational risks, and whose expertise and performance affects the outcome of the development risk.

In seeking to apply the arm’s length principle to the pricing of the arrangements between Company A and Company B, the path that the OECD guidelines take is signposted “transactional comparability.”

In other words, the guidelines seek to delineate accurately the actual controlled transaction so that it can reliably be compared with similar uncontrolled transactions.  

This will involve an analysis of the economically relevant characteristics or comparability factors as summarised in Chapter 1, paragraph 1.36 of the OECD transfer pricing guidelines.

The economically significant risks assumed by the parties to the controlled transaction will typically be relevant to comparability with uncontrolled transactions since a difference in the assumption of risk will likely correspond to a difference in the pricing of the transaction.

In our example, the assumption of development risk by Company A will fall to be verified under the control framework summarised at Chapter 1, paragraph 1.60. 

In this contrived example, readers may recognize some features as causing potential divergence in applying the control over risk test, which is discussed further in this note.  However, the essence of the control over risk test is to focus on the two parties to the transaction, Company A and Company B, and to ask, is the contractual assumption of development risk by Company A supported by its capability and performance of decision-making to take on the development risk. 

If it is, then the guidelines are clear that it does not matter to the assumption of that risk under the control over risk framework if Company B also has capability, makes contributions to decision-making, and assumes and controls other risks.  

Having delineated the transaction, including assumption of the development risk, the guidelines describe how to apply methods based on comparable uncontrolled arrangements to determine the price of the transaction under the arm’s length principle (see Chapters II-III in particular).

Under this approach, consideration will likely be given to whether Company B operates in a manner that is replicated in the open market and if its pricing can reliably be compared, for example, with independent contract research organisations offering services with similar levels of specialism, expertise, and discretion.

This path is not always followed in practice. Instead, a mutated form of the arm’s length principle may be observed, which takes the path signposted “relative comparability” and seems to focus on the relative contributions of the parties with particular reference to their control functions and value creation.

Such an approach might typically draw on the following features of the example: research and development is critical to creating value for the ABC Group, and Company B is the main source of that activity; Company B employs many highly-paid individuals, including the chief technology officer, and participates in multi-layered management; Company B is responsible for the conduct of projects including decision-making on those projects and provides the operational management; Company B performs and controls the DEMPE functions (“development, enhancement, maintenance, protection, and exploitation,” see paragraph 6.48); Company B contributes significantly to the control of risks through design and management of projects, and provides extensive support to Company A. 

I observe that none of these features are incompatible with the possible conclusion under the guidelines’ approach that comparable transactions might reliably be found and applied.

However, under this mutated approach, there is a tendency to dismiss pricing methods based on transactional comparability, either by claiming that there are no reliable comparables or by claiming that Company B should have an enhanced share in the value created by the group and that a distorted version of a profit split is justified.

In some cases, in practice, the split is based on relative assessments of value creation or on relative assessments of decision-making and control functions. A notion of relative comparability between the parties is created. 

Thus, control functions have mutated from the stated purpose in the guidelines of verifying the assumption of risk as part of the process of delineating the controlled transaction to a new purpose of allocating profits.

A concerning aspect of the modified approach is that its proponents claim to find support in the guidelines. Yet, the guidelines do not encourage pricing based on relative evaluation of control functions and do not set out to explain value and how it is created. 

 A concerning aspect of the modified approach is that its proponents claim to find support in the guidelines. Yet, the guidelines do not encourage pricing based on relative evaluation of control functions and do not set out to explain value and how it is created. 

The guidelines stick to the yardstick of observable prices in the open market for transactions that are comparable to the controlled transaction.

Transactional risk

The control of risk framework is intended, I believe, to serve a relatively narrow but critical purpose: to determine which party to a transaction assumes transaction-specific risks arising from that transaction. 

The process is similar to determining which party owns assets or performs functions so reliable comparisons can be obtained. Since the first version of the OECD transfer pricing guidelines in 1979, a functional analysis must identify which party assumes risk.  

Therefore, it is a necessary process for the healthy functioning of the arm’s length principle, as currently understood, to determine assumption of risk by the parties to the transaction.

For better or worse, the arm’s length principle is currently applied with a heavy emphasis on transactions and transactional comparability. 

A reader of Chapter I of the guidelines should find the emphasis on transactional comparability incontrovertible. Paragraph 1.33 states:

“Application of the arm’s length principle is based on a comparison of the conditions in a controlled transaction with the conditions that would have been made had the parties been independent and undertaking a comparable transaction under comparable circumstances.”

The bulk of Chapter I is then devoted to providing guidance on how accurately to delineate the actual transaction, a process which, according to paragraph 1.35, “requires analysis of the economically relevant characteristics of the transaction.”

Risks are consistently identified as an important aspect of understanding the transaction. Paragraph 1.57 states:

“Identifying risks goes hand in hand with identifying functions and assets and is integral to the process of identifying the commercial or financial relations between the associated enterprises and of accurately delineating the transaction or transactions.”

In turn, risks are important in making comparisons between transactions. As stated in paragraph 1.58:

 “[I]n making comparisons between controlled and uncontrolled transactions and between controlled and uncontrolled parties it is necessary to analyse what risks have been assumed, what functions are performed that relate to or affect the assumption or impact of these risks and which party or parties to the transaction assume these risks.”

The arm’s length principle described in the guidelines depends, therefore, on comparing the pricing of controlled transactions, once accurately delineated, with the pricing of similar uncontrolled transactions. 

That will usually involve finding uncontrolled transactions with a similar allocation of assets, functions, and risks to the party or parties being examined.  

How the parties assume economically significant risks under the terms of the transaction will tend to affect the pricing of the transaction. Therefore, determining the assumption of risk by the parties to a transaction can be critically important to making reliable comparisons with transactions involving a similar risk assumption so that appropriate pricing can be identified under the arm’s length principle.

A simple transactional example is that of a contract to provide debt-factoring services. A specific risk that will affect the pricing of the transaction is the bad-debt risk and whether that has been transferred to the service provider under without recourse terms. 

A comparison of the pricing of the controlled debt-factoring transaction with uncontrolled arrangements for the purposes of applying the arm’s length principle will be meaningless if it is not known which party assumes the bad-debt risk under the transaction. 

Therefore, application of the arm’s length principle requires examination of which parties to the transaction assume specific economically significant risks that would affect the pricing of the transaction at arm’s length so that reliable comparisons can be made.  But it does not require the identification of all possible risks and risk outcomes that could be encountered by the parties to the transaction.

Not all possible risk

 Once a risk is assumed in a transaction, and a price struck, then the consequences flow. Chapter 1, paragraph 1.63 of the OECD transfer pricing guidelines states:

“Risk assumption means taking on the upside and downside consequences of the risk with the result that the party assuming a risk will also bear the financial and other consequences if the risk materialises.”

Independent parties entering into, for example, a debt-factoring arrangement would likely analyse and price the bad debt risk by reference to the quality of the debt and may incorporate some macro-economic factors in the calculation. However, it is doubtful that wider risks, such as natural disasters, would be explicitly considered in striking a price. 

It is also highly doubtful that any debt factoring arrangement before 2020 took into account the suspension of the global economy in response to the COVID-19 pandemic.  But both natural disasters and a global pandemic, among other examples of adverse external events, could affect the outcome of the assumption of bad-debt risk in our example. 

However, the consequence of risk assumption is not an issue that should be determined under the arm’s length principle and is beyond the scope of what an MNE group can influence (see Chapter 1, paragraph 1.67).

This same point is made by the OECD in paragraph 8 of its publication dated 18 December 2020, “Guidance on the transfer pricing implications of the COVID-19 pandemic” (OECD 2020). That section states:

“[I]t may be determined that a party to a controlled transaction cannot influence the hazard risk associated with a pandemic, but nevertheless assumes other risks that have materialised as a result of COVID-19.”

Once it has been determined under the arm’s length principle which party assumes a risk arising in a controlled transaction and appropriate comparisons are made with uncontrolled transactions to determine the price of that transaction, then the consequences of that risk assumption are assumed to flow to the controlled and uncontrolled parties alike. The consequences of risk assumption are not determined by the controlled relationship.

It follows, therefore, from this transactional emphasis, that the guidelines do not require an analysis of all risks that the MNE group may encounter, and certainly not an assessment of all the possible consequences that arise from the assumption of a specific risk.

It follows, therefore, from this transactional emphasis, that the guidelines do not require an analysis of all risks that the MNE group may encounter, and certainly not an assessment of all the possible consequences that arise from the assumption of a specific risk.

 It would not be a fair test of the guidance on risk to examine how the risks associated with COVID-19 would be treated.  The pandemic is an example of an adverse external event that gives rise to a range of uncertainties affecting the outcomes of risk assumption.  But there is no requirement in the guidelines that such adverse external events should be included in an analysis of ex ante risk assumption in a transaction. 

Instead, the guidelines require examination of which parties to the transaction assume specific economically significant risks that would affect the pricing of the transaction at arm’s length so that reliable comparisons can be made. 

Paragraph 1.71 defines risk in the context of transfer pricing  as “the effect of uncertainty on the objectives of the business.”  

Some effects cannot be identified in advance of the risk eventuating, but this also means that such effects cannot be taken into account by independent parties and would not be priced into the terms of the transaction between them. 

Such effects of uncertainty may determine the outcomes of the assumption of identified risks. Still, if independent parties cannot take them into account in transactions between them, then the arm’s length principle would be wrong to insist that associated enterprises should do so.  The guidelines do not expect that associated enterprises would have perfect foresight denied to independent parties. 

Some readers refer to the statement in paragraph 1.56 to support their reading that the transfer pricing guidelines require all risks to be identified. That section states that “[a] functional analysis is incomplete unless the material risks assumed by each party have been identified and considered . . .”

However, the rest of the sentence roots the phrase in a transactional pricing context, stating:“since the actual assumption of risks would influence the prices and other conditions of transactions between the associated enterprises.” 

As explored in the guidelines, the arm’s length principle is based on pricing a transaction taking into account all of its economically relevant characteristics, including the assumption of risk by the parties to the transaction. 

The risk of future contamination might be an example of an underlying hazard that could be difficult to anticipate and is an unanticipated effect of uncertainty that affects the risk assumed from investment in land. 

However, that risk is latent and not subject to analysis under the guidelines until it is material to a transaction. Contamination could be a risk intrinsic to a controlled transaction. For example, when an associated enterprise undertakes site preparation or remediation or undertakes a site’s development on behalf of the site owner.

The risk of land contamination may be highly relevant to the pricing of such transactions between independent parties. The party prepared to assume that risk might be expected to evaluate its likelihood and potential costs.  This type of transactional risk, where the risks are intrinsic to the pricing, require that the guidelines be considered.

I believe the guidelines encourage a narrow focus on transactional risk for purposes of appropriate comparisons with similar transactions. However, under the mutated approach, a census of all risks, the thousand natural shocks that a business may encounter, is taken, and the concentration of risk in every party is compared. 

Control over risk as an arm’s length concept

The guidelines provide that control over risk is the means of verifying assumption of risk in a transaction. Control is a defined term in the guidance that is derived from its definition of risk management. Paragraph 1.61 states:

“Risk management comprises three elements: (i) the capability to make decisions to take on, lay off, or decline a risk-bearing opportunity, together with the actual performance of that decision-making function, (ii) the capability to make decisions on whether and how to respond to the risks associated with the opportunity, together with the actual performance of that decision-making function, and (iii) the capability to mitigate risk, that is the capability to take measures that affect risk outcomes, together with the actual performance of such risk mitigation.”

According to paragraph 1.65, control over risk involves the first two elements of risk management. That paragraph states that “[i]t is not necessary for a party to perform the day-to-day mitigation, as described in (iii) in order to have control of the risks.”

The guidance further recognises that these day-to-day risk mitigation activities may be outsourced, stating:  

“However, where these day-to-day mitigation activities are outsourced, control of the risk would require capability to determine the objectives of the outsourced activities, to decide to hire the provider of the risk mitigation functions, to assess whether the objectives are being adequately met, and, where necessary, to decide to adapt or terminate the contract with that provider, together with the performance of such assessment and decision-making.”

The concept of control is not new to the OECD transfer pricing guidelines. The connection between risk allocation and control was made in the 1999 version (paragraph 1.27) of the guidelines and repeated in the 2010 version (paragraph 1.49).

The 2010 version included the new chapter on business restructurings, which further developed the connection.

It defined control in paragraphs 9.23 and 9.24 in similar terms to how it appears in the revised guidelines and reverse-engineered what was meant in paragraph 1.49.

Therefore, according to paragraph 9.22,  what had been meant in 1999 was that “[t]he question of the relationship between risk allocation and control as a factor relevant to economic substance is addressed at paragraph 1.49.”

From these beginnings, the 2017 guidelines affirm the connection between risk assumption and control and devote a significant number of paragraphs to guidance on dealing with risks related to a controlled transaction.

The guidelines seem sensitive to the perception of imbalance, stating in paragraph 1.59: 

“The detailed guidance provided in this section on the analysis of risks as part of a functional analysis covering functions, assets, and risks, should not be interpreted as indicating that risks are more important than functions or assets. The relevance of functions, assets and risks in a specific transaction will need to be determined through a detailed functional analysis.”

If control of risk serves the relatively narrow purpose of determining which party to a transaction assumes specific risks arising from that transaction, why, it may be asked, does it consume a significant chunk of Chapter I of the guidelines.

As the guidelines acknowledge in paragraph 1.59, part of the answer is that it is more difficult to determine where risks are assumed compared to where tangible assets are owned or where functions are performed. 

Part of the answer is also that there is perhaps less conformity in risk assumption within categories of business activities than conformity in functions and assets.

A manufacturing activity might be expected to conform to certain patterns of functions and assets, but its risk profile (or the capacity in which it carries out the manufacturing functions, as the 1979 guidelines would see it) might vary significantly compared to another manufacturer with similar functions and assets, with the potential for significant variation in pricing. 

Indeed, the same manufacturing activity could deploy the same functions and assets under two different transactions, each with different risk allocations between the parties and with consequential differences in pricing.

Such a lack of conformity in risk assumption has been exploited for transfer pricing purposes over the past thirty years or so as risks have been reallocated, limited, or stripped to support pricing outcomes.

Part of the answer is that ex ante risk assumption affects the pricing of transactions and can materially affect transactions’ outcomes, sometimes several years after the initial transaction. So, as suggested in paragraph 1.78, there is scope for significant controversy when there is disagreement about which party assumes risk.

The question of whether the control test reflects an arm’s length concept could be stated as follows: is it reasonable to presume that independent parties taking on an economically significant risk under a transaction would have the capability to evaluate the likely consequences of that risk assumption to determine the terms of that transaction?

I would contend that this is a reasonable presumption and that, therefore, the guidelines’ control test seems to be consistent with the arm’s length principle. 

In our debt-factoring example, the guidelines set out how to determine which party assumes the bad-debt risk under the transaction. 

The analysis would typically proceed along the following broad lines: bad-debt risk has been identified as an economically significant risk in this type of transaction, examination of the contract indicates the risk is assumed by Party A under without recourse terms, there is no evidence that in practice the cost of bad debts are covered by the other party, and Party A exhibits the capability to evaluate the bad-debt risk and the capability to respond to the risk together with the performance of that capability.

In this case, as the guidelines acknowledge in paragraph 1.87, the determination may be straightforward:

“Where a party contractually assuming a risk applies that contractual assumption of risk in its conduct, and also both exercises control over the risk and has the financial capacity to assume the risk, then there is no further analysis required . . .”

The analysis can be less straightforward when capability is lacking in the enterprise assuming risk under the contract and in practice.

If Party A in this example exhibits a lack of capability to evaluate the bad-debt risk and does not make up that capability by seeking relevant advice, it would have no basis for knowing whether the price it is charging adequately takes account of the assumption of that risk. 

The guidelines make the inference that independent parties taking on an economically significant risk in a transaction would have the capability to evaluate the assumption of that risk in order to conclude the terms of that transaction.

Given the need to verify risk assumption to make reliable comparisons under the arm’s length principle, this seems to be a reasonable inference.

Decision-making is not competitive—who makes the most decisions is irrelevant

 Control over risk requires a consideration of decision-making. According to paragraph 1.76 of the guidelines:

“Control over a specific risk in a transaction focuses on the decision-making of the parties to the transaction in relation to the specific risk arising from the transaction.”

The revised version of the guidelines include the fund manager example at paragraph 1.70, which has its origins in the business restructurings chapter in the 2010 version (paragraph 9.25). 

In both versions, the example illustrates the principle that a risk (here, the risk that the value of the investment could go up or down) is not transferred to the party providing risk mitigation service (as defined in the guidelines as measures that are expected to affect risk outcomes).  

The fund manager’s investment activities are likely to be extensive and, as the guidelines acknowledge, will involve the fund manager’s own sub-stratum of risks.

The example may be reasonably representative of similar cases of service providers, each with their own specialist expertise and knowledge and their own population of risks that may not be apparent to the person hiring them. 

The investor is clearly not evaluating and deciding on all the risks involved in managing the investment, but the question that the guidelines seek to address is whether the investor assumes the specific investment risk in its transaction with the fund manager or whether the fund manager should be regarded as assuming the investment risk and, as a result, should be allocated the upside and downside outcomes of the investment. 

For the fund manager to be allocated the upside and downside outcomes of the investment, the investor would need to be in a position similar to that of Company A in the example of investment in a tangible asset set out in paragraph 1.85 of the guidelines, and the fund manager would need to be in position similar to that of Company B. 

In other words, the investor would have no capability to evaluate the potential investment, to assess how much to invest, and how that investment should be managed, whereas the fund manager would have made decisions about the size and nature of the investment that it would manage. The test in the guidelines does not refer to the relative concentration of risks between the parties.

There are wider implications to the example that touch on some of the deviations from the transfer pricing guidelines observed in practice.

There will likely be far more people involved in fund management activity than in investment activity. There will also likely be a far greater number of decisions made by fund management activity than made by investment activity, and a far greater number of risks managed by the fund manager in performing its activities than in the investment activity.

The example touches on a source of increasing tension in cases where one party does all the work and another party provides the funding and direction for that work to be done.

Namely, significant risks are apparently assumed by a relatively small number of people compared to the number of people performing activities that affect the outcome of those risks. Further, there may be a relatively narrow decision-making process by the party assuming risk compared to the extensive number of continual decisions made by the party mitigating the risks. 

The OECD guidelines’ introductory guidance on conducting a functional analysis makes a cautionary comment on scale:

“While one party may provide a large number of functions relative to that of the other party to the transaction, it is the economic significance of those functions in terms of their frequency, nature, and value to the respective parties to the transactions that is important.”

This statement, in paragraph 1.51, indicates that scale that is represented, for example, by extensive activity by one party and more modest activity by another is not, in itself, a determinative factor. What matters is the economic significance of the respective activities’ contribution, tempered by the contractual framework in which they take place. 

The stated purpose of delineating activity in the guidelines is not to allocate relative value to concentrations of functions and risks as the mutated interpretation to the guidance would suggest, but to identify comparable uncontrolled transactions.

The stated purpose of delineating activity in the guidelines is not to allocate relative value to concentrations of functions and risks as the mutated interpretation to the guidance would suggest, but to identify comparable uncontrolled transactions

A common application of the wider implications of the fund manager example in practice is the investment in research and development. 

A scenario similar to the framing example at the beginning of this note, might be the appointment by the parent or other principal enterprise of an associated enterprise to conduct research and development activities under a transaction that requires the parent/principal to fund all costs.

Typically, the research affiliate is highly specialised with significant capability and expertise, including its own management team that will likely have considerable discretion to design and direct the research programmes. 

Such programmes are unlikely to be rigid but will be adapted, curtailed, or expanded as results are analysed.

Typically the research affiliate will not have any ownership interest in the intellectual property that might be generated by those programmes (other than being licensed rights to conduct its research) and will not assume the development risk (the risk of upside or downside outcomes compared to the financial outlay). The research affiliate will instead have a raft of risks of its own relating to its conduct of research and associated with, for example, safety, regulatory, data-management, confidentiality, and maintaining performance capability.

It may also be the case that detailed research programmes designed by the research affiliate may not be specifically evaluated by the funding party, and the raft of risks encountered by the research affiliate may not be apparent to the funding party. 

A key transfer pricing issue to address in these situations is which party assumes the specific development risk and, as a result, should be allocated the upside and downside outcomes of the development. 

The control of risk framework in the transfer pricing guidelines provides a way to address the issue, based on the criterion of control.

Does the parent/principal that contractually assumes development risk decide how much cash to make available and how the research programmes should be funded? (The scenario is effectively that of example 1 in paragraph 1.83.) If it does, then its contractual assumption of the development risk is supported.

It does not matter to the assumption of risk if the research affiliate helps make that decision, if it makes far more numerous decisions about the conduct of the research it undertakes, or if it controls a concentration of risks about which the parent/principal may have little awareness.

The determination of the assumption of development risk and other economically relevant characteristics sets the scene for pricing comparisons with uncontrolled transactions.

What is irrelevant to the question of the assumption of development risk is the enumeration of people involved in decision-making between the two parties and an enumeration of decisions. I do not read the guidelines as encouraging a census of decision-making in an MNE group.

In practice, there seems to be a bias in transfer pricing audits favouring the party with more people, with more extensive activities and associated decision-making, as the appropriate party to be allocated the risk.

Decision-making is wrongly being treated as competitive between the parties in the quest for risk assumption under this mutated version of the OECD transfer pricing guidelines.

There is nothing in the transfer pricing guidelines to equate risk assumption with the number of people or concentration of decision-making. In any event, such enumeration and recording do not assist in determining the price of a transaction.

There is no recognised currency of decision-making, and I can find no statements in the guidelines that might be read to suggest that decision-making should be priced.

Contract research organisations make continual and numerous decisions about the conduct of the research services they provide. They are not, to the best of my knowledge, remunerated by decision.

However, an uncontrolled transaction involving a contract research organisation that does not assume the client’s investment risk could help price the services of a controlled provider of research services where the associated counter-party assumes the development risk.

Decision-making and multi-layered management

 Decision-making can be hard to locate. The transactional nature of transfer pricing, a consequence of taxation by reference to individual legal entity profit, can be a source of conceptual and practical difficulty. 

MNE groups seek to reduce transactions and may operate through divisions based on activities that ignore and effectively merge legal entities. Tax requirements mean that transactions must be reinstated and profits calculated for each legal entity. 

The tension in making such reconfiguration is most acute, perhaps, in locating decision-making since the actual, original decision is the goal, not a reconfigured one.

In practice, MNE group decision-making may not be compartmentalised within neat legal entity structures that have little bearing on how the MNE group is managed; senior management may be employed by several group companies and organised in executive committees under delegated authority from the main board. 

It could be the case, for example, that the group’s chief technology officer, or similar, is not employed by the party contractually assuming the development risk, and that decisions about development opportunities are made by an executive committee comprised of senior managers formally employed by several companies.

These matrix management and collaborative approaches should, it seems, provide the context in which to evaluate whether the enterprise contractually assuming the risk has the capability, through the decision-making structures in place in the MNE group, to evaluate taking on the risk.

Referring to decision-making, paragraph 1.51 of the guidelines acknowledge that “it may be helpful to understand the structure and organisation of the MNE group and how they influence the context in which the MNE operates.”

It would have been useful, however, if further guidance on the location of decision-making in multi-layered management structures was attempted. 

If the MNE group uses matrix management and this provides the decision-making forum transfer pricing considerations should not require the establishment of a different, legal-entity based management structure.

At the same time, the tax fiction of a separate, functioning legal entity must be maintained under current articulations of the arm’s length principle.

For the control test, the question is how the enterprise assuming transactional risk participates in that matrix management to obtain access to the necessary capability and in order to be involved in the performance of decision-making, not only with respect to the decision to take on a risk, but for decisions regarding responding to risk. 

This may involve considering whether the party assuming risk employs any of the matrix managers, or has its own management capable of evaluating the information it obtains.

In such matrix management structures, it may be the case that both associated enterprises in the transaction have access and involvement. However, this would not mean that the enterprise contractually assuming risk could not meet the control test in its own right.

Under the mutated form of the arm’s length principle, the existence of matrix, multi-layered management is a reason to assert that, since many parties are involved in decision-making, all should share in the resulting profits.

Under this approach, the transactional way that the parties deal with one another, where they assume different risks and have the capability to evaluate those risks, is eroded. Thus, the application of the arm’s length principle by reference to the pricing of comparable uncontrolled transactions is ignored.

Operational decision-making is not more relevant than decision-making by top-management

I do not observe any bias in the transfer pricing guidelines regarding whether the control test is directed towards top management, line management in business segments, operating entities, or functional departments. 

Tax authorities of jurisdictions that host the group’s top management team may feel that their activities are not appropriately rewarded and may be most interested in applying the decision-making analysis at the level of top management.

Tax authorities of jurisdictions that do not host the group’s top management may be most interested in applying the decision-making analysis at the operational level.

They may suggest that the following extract from the guidelines in paragraph 1.76 advances such an interpretation:

“Line management . . . may identify and assess risk against the commercial opportunities, and put in place appropriate controls and processes to address risk and influence the risk outcomes arising from day-to-day operations. The opportunities pursued by operational entities require the ongoing management of the risk that the resources allocated to the opportunity will deliver the anticipated return.”

Such an interpretation of this extract is without foundation. A proper reading in the context of the entire paragraph shows that the guidance means the opposite: management levels are illustrative of activities that do not determine control over risk.

The paragraph refers to a wider-policy setting that cannot be regarded as decisions to take on the specific inventory risks in the example of the product supply transaction and is pointedly even-handed in referring to top management policy-setting and line management policy-setting. 

Wider policy setting has two illustrations in this paragraph: determining overall levels of working capital tied up in inventory, and co-ordination of appropriate minimum stocking levels across markets to meet strategic objectives. 

I read the illustrations to be agnostic about levels of management, although working capital policies sounds rather like a top management objective and appropriate stocking levels rather like an operational management policy. 

However, I think the main point of the paragraph is to encourage the analysis to concentrate on the parties to the supply transaction that give rise to inventory risk for them, and not on the various ways in which the outcome of that risk may be affected, mitigated or increased, by policy-setting at various levels of management in the MNE group (that, by implication, are not party to the transaction). 

To my mind, the point emphasises the focus in the guidelines on the parties to the transaction that give rise to risk for them, and the guidance does not encourage a census of all decision-making.

I cannot read into the guidelines any under-estimation of the importance of risk management decision-making performed by top management and any over-estimation of the importance of executional processes on the part of line management. 

The most critical judgements involving the greatest risks typically are made by top management, and I do not find any under-estimation in the guidelines of top management. 

Decision-making in the guidelines is not linked to levels but is linked to competence appropriate to the risk. Paragraph 1.66 notes that “[d]ecision-makers should possess competence and experience in the area of the particular risk for which the decision is being made and possess an understanding of the impact of their decision on the business.”

Critical decisions around, for example, R&D investment may require judgements about how much cash should be made available to R&D and therefore not available to be used for other purposes, and then how the R&D cash should best be targeted. 

That might involve evaluating which trends in the commercial and external environment might be anticipated to produce new opportunities, which potential projects should be prioritised, and which should be postponed or discarded. It might also involve judgements about balancing returns in the shorter term with longer-term bets, and about how to achieve the objectives through acquisitions, in-house capability, or collaborations. 

All of these sound like top management decisions and may be relevant in determining which associated enterprise in a transaction involving investment in R&D assumes the investment risk.

Critical decisions around, for example, inventory levels may require judgements about anticipated demand, cost forecasting, market-price changes, production scheduling, storage capacity, product life-cycle, and monitoring of working capital requirements. They sound like line management decisions and may be relevant in determining which associated enterprise in a transaction involving the transfer of products assumes the inventory risk.

The guidelines demonstrate awareness that different risks are likely to be controlled by different levels of management. Paragraph 1.71 states:

“[A] different flavour of ice-cream may not be the company’s sole product, the costs of developing, introducing, and marketing the product may have been marginal, the success or failure of the product may not create significant reputational risks so long as business management protocols are followed, and decision-making may have been effected by delegation to local or regional management who can provide knowledge of local tastes.  However, ground-breaking technology or an innovative healthcare treatment may represent the sole or major product, involve significant strategic decisions at different stages, require substantial investment costs, create significant opportunities to make or break reputation, and require centralised management that would be of keen interest to shareholders and other stakeholders.”

For the reasons given above, I do not believe that the guidelines position decision-making at the business line level within the operating subsidiaries instead of top management. 

The appropriate level of competence depends on the risk.

My position developed in this note is that the framework in the guidelines of control of risk is applied to and is rooted in a transaction between the parties solely to verify the assumption of risk arising from that transaction.

The ultimate aim is to delineate all economically relevant characteristics of that transaction, of which risk assumption is one element in order that the pricing may be compared with the pricing of uncontrolled transactions that have similar characteristics. 

If the parent is a party to the transaction, then its functions and its risk assumption will be relevant to delineating the transaction and to the pricing of that transaction. 

If the parent is not a party to the transaction, then it is not affected by the pricing of the transaction or by the risks assumed by the parties to the transaction. 

However, there is a further transfer pricing matter for an associated enterprise or enterprises providing top management (not necessarily limited to the parent, particularly in matrix management structures or in structures where certain functions are centralised) which is not a party to the transaction being analysed and which does not, therefore, assume risk arising from that transaction. 

The transfer pricing matter is whether top management is providing a service under a separate transaction, which requires to be evaluated under the arm’s length principle.

The guidelines currently do not offer much guidance about the transfer pricing issues involving services provided by top management to associated enterprises assuming risks, and the programme of work to revise Chapter VII may help fill the gap.

Control of risk is not relative, is not designed to value control functions, and the rash of profit allocations

 It is unsatisfactory if taxpayers feel compelled and tax administrations feel encouraged to introduce so-called profit split models based on the decision-makers’ location because of differing interpretations of the control of risk framework in the guidelines. 

The justification for independent parties to share their profits arising from a transaction between them based on their respective numbers of decision-making employees eludes me (and is not discussed in guidance on profit splits in the guidelines).  

The sole purpose of the control of risk framework is to verify transactional risk assumption based on the capability and performance to take on the risk of the parties to the transaction. 

However, the growing rash of profit splits, which seems to be encouraged by tax administrations during audits, is noted in practice, and I suggest that it stems from three misconceptions.

The first misconception is to interpret the control test as a relative one. 

My spirits sink when transfer pricing examiners announce that it is necessary to map out control functions within the MNE group since experience shows that this leads to a side-stepping of the transactional approach of the guidelines and a self-fulfilling conclusion that pricing based on comparability cannot be pursued and a profit split must be used.

 My spirits sink when transfer pricing examiners announce that it is necessary to map out control functions within the MNE group since experience shows that this leads to a side-stepping of the transactional approach of the guidelines and a self-fulfilling conclusion that pricing based on comparability cannot be pursued and a profit split must be used.

The control test under the transactional approach in the revised guidelines seeks to determine the legal entity assuming risk, and does not allocate risk to every party that might have an involvement in control.  It is, therefore, unlike the approach adopted under the OECD AOA for attributing profits to different parts of a legal entity based on certain people functions.

The purpose of the guidelines’ control test is not to allocate risks to each party that might have some involvement in the exercise of control of a specific risk, irrespective of whether they are involved in the transaction in question.  Rather, the control test aims to challenge the contractual assumption of risk by a party to the transaction that is not involved in the control of that risk.

There is no concept of partial control, or a certain level of control, that needs to be reached to justify the assumption of risk.  If the party contractually assuming risk exercises control, then that party’s assumption of risk is confirmed.

If a party to the transaction contractually assuming risk exercises control over that risk, then the risk remains with that party and no further analysis is required (see paragraph 1.87). 

The fact that other parties can contribute to control does not complicate the analysis if the party contractually assuming risk is the only party that actually exercises control through capability and functional performance (see paragraph 1.93).

Importantly, where the associated enterprise assuming risk controls that risk, no further analysis is required even if other enterprises also exercise control over the risk (see paragraph 1.94). 

These statements recognise that in an MNE group several associated enterprises may be involved in risk management and may participate in the exercise of control (see for example paragraph 1.76, where various parties in the MNE group may be involved in managing the inventory risk, used as an illustration).

Notwithstanding management and control by several associated enterprises, if control is demonstrated by the enterprise contractually assuming risk, there is no need to consider re-allocation. 

Two or more parties to the transaction would only share assumption of risks under the OECD  guidance if those parties contractually agree to do so, and that joint assumption of risk is borne out in their conduct (a matter relevant to the transactional profit split method, as discussed further, below) (see paragraph 1.95).

There is no concept of relative control determining which party to the transaction assumes risk except in one defined, and probably unusual, circumstance. That circumstance is where the party contractually assuming risk does not control the risk, and multiple associated enterprises do control the risk.  In that case, paragraph 1.98 states that “the risk should be allocated to the associated enterprise or group of associated enterprises exercising the most control.”

There are no other circumstances in the guidance where risk is allocated on the basis of relative contribution to control. 

In all other circumstances, the test is a simple one: does the party to the transaction contractually assuming risk control the risk.  If it does, the existence of other parties exercising control through their decision-making is irrelevant to the assumption of risk.

The second misconception stems from a misreading of one line in the guidelines.  The guidance provides in paragraph 1.105 that compensation to a party that controls risk will usually derive from the consequences of being allocated risk since that party “will be entitled to receive the upside benefits and to incur the downside costs.”

Where a party contributes to the control of risk, but does not assume the risk, the guidelines, in paragraph 1.105  state:

“[C]ompensation which takes the form of a sharing in the potential upside and downside, commensurate with that contribution to control, may be appropriate.”

That guidance was intended, I believe, to be interpreted as referring to incentivised pricing structures that may be seen between independent parties under which a party might be entitled to an additional fee (or suffer a penalty) when it meets (or fails to meet) certain pre-determined targets that may impact the risk consequences of the principal party assuming the risk. Those targets might relate to, for example, time taken, cost variability, or the quality and effectiveness of the delivered product or service.

Unhappily, the guidance seems to be interpreted as meaning a transactional profit split method may be applied. 

However, this erroneous interpretation has been corrected by the revised guidance on profit splits, which states in paragraph 2.137, that “the mere fact that an entity performs control functions in relation to a risk will not necessarily lead to the conclusion that the transactional profit split is the most appropriate method in the case.”

The misreading of paragraph 1.105 to support profit splits contains a logical inconsistency.  The purpose of the control test is to identify the party to a transaction assuming risk, and an important consequence for transfer pricing flows from the finding. 

The important consequence is that the profits or losses arising from risk outcomes rest with the party assuming the risk (see paragraph 1.106), and that a party not assuming risk will not be entitled to returns or bear losses associated with that risk.  

Having spent the effort in applying the control of risk framework to identify the party assuming risk, it is logically inconsistent to read paragraph 1.105 as saying, ‘forget about all that analysis, if there is a party that does not assume, risk it can nevertheless share in the risk outcomes.’

It seems inconceivable that this paragraph could be taken to mean that a party that does not assume risk should have a pricing mechanism that reflects the outcome of assuming the risk that it has been determined it does not assume.

The revised guidance on profit splits is entirely consistent with the view that profits that reflect the playing out of risks should be split only between the parties assuming those risks.

The guidance endorses the application of a profit split where both parties share the assumption of risks in relation to the transaction, as envisaged in paragraph 1.95, and where the split of actual profits arising from the transaction will reflect the playing out of the risks assumed by each party (see paragraphs 2.139-142).  

The guidance on profit splits is unequivocal on this point, noting at paragraph 2.159 that the splitting of profits “would only be appropriate where the accurate delineation of the transaction shows that the parties either share the assumption of the same economically significant risks associated with the business opportunity or separately assume closely related, economically significant risks associated with the business opportunity and consequently should share in the resulting profits or losses.”

The guidance at paragraph 1.105 cannot properly be read to mean that a party not assuming a risk should share in a split of the profits which incorporate risk outcomes, or to signify an endorsement of profit splits that is contrary to the extensive qualification of profit splits in Chapter 2 of the guidelines.

The third misconception arises from a fundamental misunderstanding, or a deliberate ignoring, of the way in which transfer pricing is envisaged to be applied in the guidelines.

The question addressed by the guidelines is what conditions would independent parties have agreed in comparable transactions in comparable circumstances. 

The guidelines focus on identifying the parties, the transaction, and its circumstances, but do this so that reliable comparisons with uncontrolled arrangements may be made. 

The guidelines do not depend on determining prices by analysing value creation.

Unhappily, the BEPS campaign slogan of aligning transfer pricing outcomes with value creation seems to have been adopted as a guiding principle of transfer pricing. 

Unhappily, the BEPS campaign slogan of aligning transfer pricing outcomes with value creation seems to have been adopted as a guiding principle of transfer pricing. 

A taxpayer that responds to the audit challenge from a tax administration to demonstrate that profits are taxed in accordance with value creation by querying, quite reasonably in my view, what is meant by value and how is it created, and, by the way, how is this question relevant to the transfer pricing law, will likely be deemed obstructive and uncooperative.

A transfer pricing audit casts itself adrift from the guidelines if it insists that value creation concepts can be used, not just for providing insights into how the business operates, but also to assign profits to parties contributing to value creation. 

The guidelines offer no framework for determining how value is created or how to discriminate between value created by different functions, assets, and risks. The guidelines studiously avoid the slogan of aligning pricing with value creation, and tend to refer to value creation only in relation to helping to understand the business context in which transactions take place. 

Despite what the guidelines state, however, there has been a tendency ever since the slogan was adopted to mutate the OECD approach and apply transfer pricing as if it does need to determine the components of value creation in generating profits, rather than seeking evidence of how independent parties price comparable transactions.

This tendency can lead to frustrating debates about the relative value of raising and allocating versus deploying capital, of labour, of assets and rights, and how any such value is created, and whether there is value in decision-making or leadership. 

The debate is frustrating because there are no agreed yardsticks to measure value, and countries will take very different views about value and how it is created.

It is a debate that the guidelines do not encourage and which they are not designed to resolve.  The guidelines stick to the yardstick of observable prices in the open market for comparable transactions in comparable circumstances.

Critically, the adoption of a value creation approach can lead to decision-making being assumed to have a value and possessing an undefined role in value creation. Instead, it should be viewed as an activity that can be relevant in defining the parties’ circumstances, particularly the assumption of risk in the transaction, to make reliable comparisons with uncontrolled transactions. 

The risk control framework is not designed to value control functions but to delineate the transaction. In particular, it verifies risk assumption so that in subsequent steps, the transaction’s price can be determined through appropriate comparisons.

The risk control framework is not designed to value control functions but to delineate the transaction. In particular, it verifies risk assumption so that in subsequent steps, the transaction’s price can be determined through appropriate comparisons.

A similar point can be made about the so-called DEMPE functions that may be relevant in delineating a transaction involving intangibles.

A value creation approach can lead to DEMPE functions being assumed to create value and further assumed to create equivalent units of value, rather than as functions that may be relevant to varying extents (including no extent—see paragraph 6.49) in defining the circumstances of the parties and the transaction, and consequently relevant to comparability. 

The mapping of DEMPE functions does not identify how profits should be allocated; one is still left with the requirement under the arm’s length principle as set out in the guidelines of delineating the transaction that the functions relate to, and finding how similar transactions are priced in uncontrolled arrangements. 

Nevertheless, in its mutated form, the arm’s length principle is being applied as if DEMPE functions and their control are units of independent value.

Once an auditor chooses to analyse value creation, and to categorise decision-making and DEMPE functions as contributors to value creation, the lack of yardsticks to determine value and to discriminate between claimed contributions, and the lack of established procedures about how to convert the heatmaps of value into profits from transactions, mean that the audit is inevitably pushed in the direction of splitting the profits. 

The profit split is a contrived solution to a problem that did not need to be created. The guidelines should not be blamed for this faulty mutated approach, which is leading to a rash of profit allocation approaches. 

It will be interesting to see how the courts respond to such approaches in jurisdictions whose transfer pricing laws incorporate the OECD guidelines.

The framing illustration once more

A the start of this note, I offered a framing illustration that summarises the divergences in applying the arm’s length principle.

The guidelines offer a version of the arm’s length principle based on transactional comparability under which it is necessary to determine the characteristics of the transaction, including the assumption of risk arising from the transaction, to make reliable pricing comparisons. 

The mutations offer a different version of the arm’s length principle based on comparisons of risk concentrations, introducing a currency of decision-making across levels of management.  The mutations thrive on relative concepts of value creation that bypass transactions and transactional comparability.

The mutations offer a different version of the arm’s length principle based on comparisons of risk concentrations, introducing a currency of decision-making across levels of management.  The mutations thrive on relative concepts of value creation that bypass transactions and transactional comparability.

Mutations in the arm’s length principle: adoption or rejection?

Are these mutations of the arm’s length principle opportunistic aberrations that may not survive, or are they responding to real or perceived difficulties in using the guidance and indicate ways in which the arm’s length principle needs to adapt to survive?

My first observation is that, while there will likely always be selective readings of the guidelines to support a particular position, I do not see the mutations in the arm’s length principle in practice as merely opportunistic.  They seem too prevalent and conform to repeated patterns. 

Rather, I suggest that the mutations respond to a feeling of unmet expectations, particularly on the part of tax administrations. 

The BEPS project did not replace the reliance on the arm’s length principle and did not modify its application, which continues to be based in the guidelines on the often challenging task of finding comparable uncontrolled transactions. 

Even the modified guidance on profit splits continues to stress, perhaps even more insistently than before, the transactional nature of the method. 

The transactional approach is key to the current version of the arm’s length principle since it allows access to third-party open-market benchmarks, but at the same time requires some leaps of faith and ingenuity to interpret and adapt how associated enterprises are actually managed and how they actually interact with one another to construct the basis for such comparisons with independent enterprises. 

One of those leaps is to determine risk assumption by a legal entity when the MNE group, (unless its business deals in risk) may not perceive the allocation of risk to legal entities as relevant to how it manages risks. 

The concept of a legal entity in the MNE group deciding whether or not to take on risk may seem nonsensical to its CEO, but is an essential fiction to be adopted under a version of the arm’s length principle based on transactional comparability. 

In a sense, therefore, the revisions to the guidelines, in providing further guidance on how to verify risk assumption, serve also to emphasise the frailties of the arm’s length principle.

I suggest, therefore, that tax administrations are looking to find ways of doing transfer pricing with reduced emphasis on transactional comparability.

Two alternatives to the transactional approach are emerging, both of which combine profits of legal entities, reduced in some cases to take into account standard functions to leave a residual profit or loss.  I refer to one as a people approach and to the other as a value creation approach.

A people approach is based very loosely on the significant people functions framework used to attribute assets and risks between permanent establishments of a single legal entity under the AOA.

This mutated version seems to thrive on the notion that if less reliance is being placed on the very features that distinguish the separate legal identity of associated enterprises, and more reliance is placed on allocating profits associated with asset ownership and assumption of risk to the important functions associated with assets, intangibles, and risks, then strong parallels emerge to the approach for attributing profits to parts of a single enterprise.

This approach tends to focus on those people that take decisions, and so can be seen to elevate the focus on decision-making from one of determining assumption of transactional risk in the guidelines to one of allocating profits and bypassing comparability.

The level of decision-making is typically subjectively determined, and the approach implies that decision-making drives profits, that decisions are equal in value, or can be reliably weighted, and that location of relative decision-making is an appropriate way to allocate profits. 

From the perspective of the tax administrations, there are considerable shortcomings in this approach given the high mobility of people and the ignoring of actual ownership of assets, rights, and the making of historic investments.

Problems with this approach include the AOA’s unsatisfactory treatment of intangibles, particularly marketing intangibles, and the AOA’s bias towards operational management

However, if a people approach to associated enterprises transfer pricing takes root in practice, it might be necessary to explore why tax administrations are adopting the approach, its potential advantages and disadvantages, and whether there are circumstances in which it could be applied in a way that reduces the scope for subjective analyses that give rise to contentious disputes, or whether the approach should be rejected.

A second approach is based on value creation.  This approach seeks to determine the factors that contribute to value, may give different weightings to different factors, and then maps those factors to jurisdictions in which the MNE group operates to provide the basis for relative allocations of profits. 

The approach is sometimes claimed to resemble a contribution analysis described in the guidelines (see paragraph 2.150) but rarely meets the indicators for the appropriate use of a profit split or the insistence in the guidelines that splitting factors should be verifiable (see paragraph 2.166). 

The main problem with this approach is that there is no yardstick to determine what contributes to value, how very different factors can meaningfully be compared even with weighting, and the location of that factor may be contentious. 

The relationship between value, whatever that might ultimately be defined to mean and which in practice may be confused with importance, and profits is debateable. 

Currently the application of a value creation approach to allocate profits tends to be frustrating because of the absence of agreed yardsticks to measure value. Thus, tax administrations in different countries will likely take very different views about value and how it is created.

However, if a value creation approach to transfer pricing is taking root in practice, then it might be necessary to explore why tax administrations are adopting the approach, its potential advantages and disadvantages, and whether there are circumstances in which it could be applied, perhaps in the context of a contributions analysis profit split, whether contributions to value could be agreed in theory and the scope for contentious disputes reduced, or whether the approach should be rejected.The necessity of exploring the advantages and disadvantages of these mutated approaches is driven by recognition that they are proving costly and contentious in audits in the absence of common understanding of how they should appropriately be applied. 

If tax administrations do adopt the mutated approaches even more widely, and taxpayers are compelled to adopt their approach accordingly, then it would also be necessary, I believe, for tax administrations to be clearer about where in the guidelines they see endorsement for relative control, conflating decision-making and control functions with value, or tabulating relative control functions to determine the allocation of profits in a contrived value creation profit split. 

If there is no such endorsement, and my view is that this is the case, then the authority for such approaches is doubtful, and the guidelines should be revised to provide authority, or the mutated approaches should be abandoned. 

The biggest issue is whether tax administrations feel that the approach based on transactional comparability, the clear standard set out in the guidelines, is worth keeping in all cases and whether they want to make it work in practice. 

This will likely require re-affirmation of the guidelines and a continued willingness to make those leaps of faith and ingenuity to reconcile how the business of the MNE group is actually conducted and decisions actually taken with the fiction of separate legal entities entering into transactions.

If transactional comparability is worth keeping, then it seems essential that there is a mechanism to determine which party assumes significant risks so that pricing comparisons can be made. 

If there are failings in the current guidance then these should continue to be discussed and improvements made.

Acknowledgement:

The catalyst for my note is R. Collier & I. Dykes, The Virus in the ALP: Critique of the Transfer Pricing Guidance on Risk and Capital in the Light of the COVID-19 Pandemic, Bulletin for International Taxation, 2020 (Volume 74), No.12  I am grateful to Richard and Ian for their comments on an earlier draft of this note.  We do not agree on everything, but share the view that debate of the issues covered in the two papers is important given the variations in practical application of the guidance.