INSIGHT: OECD Financial Transactions Discussion Draft Revisits Longstanding Treasury Policies

The Tax Management Transfer Pricing Report ™ provides news and analysis on U.S. and international governments’ tax policies regarding intercompany transfer pricing.

By Danny Beeton

Danny Beeton examines the OECD’s discussion draft on financial transactions and concludes that the draft confirms what international case law has made clear, namely that longstanding treasury policies need to be reviewed in the light of the accelerating changes in the international transfer pricing conventions.


On 3 July 2018, the OECD launched a consultation on the transfer pricing of financial transactions by publishing the first draft of a new chapter of the OECD Transfer Pricing Guidelines for Tax Administrations and Multinational Enterprises. The consultation comments are invited until the end of the consultation period on 7 September 2018, after which the comments will be considered and a revised draft will be published. The OECD have requested comments on certain questions which are of particular interest to them, and which are highlighted in boxes throughout the text.

Potentially, a new chapter on financial transactions can help to fill a large gap in the Transfer Pricing Guidelines, which has resulted in high profile disputes in this area having to be settled by courts around the world on the basis of expert evidence on how independent parties approach such transactions.

The draft guidance covers a great deal of ground, and as a first attempt it sometimes reaches no conclusion or makes controversial suggestions. Comments have been invited on many issues, some of which merit a good deal of thinking and guidance in their own right. As the drafting team have warned separately, the finalization of the formal guidance on financial transactions looks set to be a multi-year process in several stages. The main benefit of this first draft is for some important controversies to be acknowledged so that a public dialogue can begin.

Identifying What Should be Treated as Debt for Tax Purposes

The first part of the discussion draft provides guidance on the situations in which loans can be recharacterised as equity, while the second part provides guidance on the pricing of financial transactions such as treasury services, loans, cash pooling, hedging, financial guarantees and captive insurance. (Notably, performance guarantees are not covered).

Paragraph 13 notes that the appropriate capital structure of a company will be influenced by its industry, the point in its lifecycle, government regulations and the availability of finance. Paragraph 14 suggests that as well as these external factors, the business strategy of a company’s group could justify a particular capital structure; however, it is doubtful whether the examples given here would persuade some tax administrations.

There is useful case law, for example in the United States, which deals with the features of instruments that point to them legitimately being viewed as debt rather than equity; paragraph 16 lists the factors, including a fixed repayment date, an obligation to pay interest, a right to enforce payment of principal and interest, the ranking of the funder compared to other funders, the existence of financial covenants and security, the source of interest payments, the ability of the recipient of the funds to obtain loans from unrelated parties, the extent to which the funds are used to acquire capital assets, and the failure of the recipient to repay on the due date or to seek a postponement. This is a helpful list.

Paragraph 17 gives an example of a borrower which clearly could not service its related-party loan. The suggestion is that the excess lending should not be treated as a loan, although question B.2 asks whether the entire amount of the loan should be reclassified as equity (which clearly it should not). Further detail would be helpful here: for example, some tax administrations accept that it is common in some industries for loans to be refinanced on maturity without the entire principal having been repaid out of the profits or assets of a borrower.

Paragraph 19 rightly introduces the concept of the other “options realistically available” to both parties, as is now standard practice when considering transactions involving intellectual property. Thus it should be asked whether the provider of the funds could have used them more profitably in another investment opportunity. (Some more guidance would be helpful in this respect—yes, it is simple to check whether a better interest rate might have been earned in the open market, but are we to use the funder’s weighted average cost of capital, or return on capital employed, as an indicator of what it could have achieved from other investment opportunities?). With respect to the recipient of the funds, paragraph 19 suggests that it should have a business need for the funds, and that it is necessary to confirm that it would have borrowed the funds given the impact of the new debt on its credit rating and its cost of capital.

Paragraph 20 considers the possibility of benchmarking which might confirm that similar companies took on similar amounts of debt, but does not provide any real guidance in this respect. For example, while reference is sometimes still made to gearing and interest cover in similar companies’ financial statements, it is now generally recognized that these ratios are the result of various loans and in developments in the borrower’s profits and retained earnings—it is much more appropriate to refer to the gearing and interest cover covenants in recent loan agreements.

In paragraph 24 it is suggested that whether funding should be respected as a loan or not should be influenced by the extent to which the provider of the funds performs the usual functions of a lender, i.e. determining creditworthiness and assessing the risks, determining the appropriate terms of the funding, organising and documenting the funding agreement and then “reviewing” the loan (which would benefit from an explanation). This short discussion might also benefit from references to “key risk decision-making” rather than “functions”, in order to be consistent with the rest of the 2017 Transfer Pricing Guidelines. The paragraph notes that the intensity of these functions would not need to be as deep as usual for parties as well-known to each other as are members of the same group of companies.

Paragraph 25 suggests further that the beneficiary of funding should also perform borrower functions if the funding is to be respected as a loan. These include ensuring that the funds are available with which to repay the principal and the interest at the time specified, providing collateral (if needed), and monitoring and fulfilling any other obligations created by the funding agreement.

Paragraph 26 notes that the functions of a lender and a borrower may be undertaken by a central treasury company.

Paragraph 28 lists some of the features of a funding agreement which would point to it being a loan, including the amount of the loan, its maturity, the repayment schedule, the nature and purpose of the loan, the level of seniority and subordination, the currency, the collateral provided (if any), the presence and quality of any guarantee, and whether the interest rate is fixed or floating.

Identifying the Arm’s Length Interest Rate

Paragraph 31 notes that for benchmarking purposes, the common practice of considering a run of years is not appropriate for financial transactions because financial markets are so susceptible to events.

Paragraph 32 expresses strong reservations about reference to financial transactions in other currencies because these may be influenced by different economic factors such as growth rate, inflation and exchange rate volatility. It also warns against reference to transactions in other countries even in the same currencies, because of such local factors as interest rate or exchange rate controls and foreign exchange restrictions.

Box B.4 states that if the funder does not make the relevant risk decisions (but does provide the finance) then it will be entitled to no more than a “risk-free” rate of return—that is, the return on a government security in the same currency with the same maturity. If it does make the risk decisions, then it is entitled to earn a “risk-adjusted return”, which can be identified from comparable loans or bonds, or constructed by adding a market risk premium to a risk-free rate. (Here—and again in paragraph 86—it would be helpful to include a discussion of the differences between interest rates and bond yields, in particular the illiquidity premium which is embedded in interest rates compared to bond yields where the borrower has a poor credit rating).

Box B.4 also refers to the possibility of building up an interest rate by adding a risk-appropriate “profit margin” to the lender’s cost of funds. In paragraph 89 this calculation is set out in more detail, namely, the cost of funds to the lender in raising the funds to lend, plus its expenses incurred in arranging and servicing the loan, plus a risk premium, plus a profit margin. It is important that a comment be added here to the effect that all this will calculate is the minimum interest rate which the lender would charge, and not the maximum interest rate which the borrower would pay (unless the calculation of the “profit margin” is specified in a certain way). In paragraph 91 it is suggested that when financing is through “back-to-back” loans this method may not be appropriate, as only agency or intermediary services functions may be being performed and hence a mark-on on the cost of the “agency” function might then be a more appropriate transfer pricing method than an additional interest rate margin.

Centralised Treasury Services

Section C of the draft guidance, (starting at paragraph 37), deals with group treasury services. Paragraph 39 notes that these can involve such functions as optimising liquidity management (that is, ensuring funds are available in the form required). Paragraph 40 notes that a further function of the group treasury centre may be to develop longer-term financial risk management strategies (which can reduce the risk of the whole business and hence its cost of capital) and to plan investment decisions. Paragraph 41 notes that other treasury functions can include raising debt and equity, and managing relationships with the group’s external bankers and with ratings agencies. Paragraph 43 suggests that treasury services are usually “a support service” for which the general guidance on intra-group services can be applied. Paragraph 45 supports this view by noting that group management will set the policies on target rates of return, cash flow volatility and net assets and that group treasury will follow this “vision, strategy and policies”. (However, this seems to be rather at odds with the comment in paragraph 40 that group treasury may make key decisions with regard to risk management and investments).

Further Consideration of the Arm’s Length Interest Rate

For some reason, the guidance on interest rates is split between two sections of the draft chapter, ending in Box B.4 and resuming in paragraph 47 after the discussion of treasury functions, which is not helpful.

Paragraph 52 suggests that the granting of security over its assets is not necessary for a subsidiary which is borrowing from its parent because the parent already has control and ownership of those assets. It follows that if those assets are not already pledged as security to another lender, any parental loan could be priced as if it were secured. This is an important and controversial viewpoint about which courts have reached different conclusions in different jurisdictions.

Paragraph 54 makes another controversial suggestion, which is that when a borrower’s assets are not pledged to a third party, its best option is usually to seek a secured loan from a related party (which would be at a lower interest rate—the implication would be that a tax deduction would only be allowed for the lower interest rate on almost every related party loan). The exception would be where a borrower needed to keep its collateral available in case it might need a further loan.

Paragraphs 55 and 56 allude to the choice which an independent party would have made between a fixed interest rate or a floating rate, the term of its loan and the option to prepay without a penalty, but only indirectly, and it would be helpful to be more explicit on these questions.

Box C.2 asks for comments on a rather strange proposal, which is that the credit rating of any subsidiary can be assumed to be the same as that of the parent company of the group as a whole. Clearly, no ratings agency or bank would take such an approach; rather, ratings agencies calculate the standalone rating for the subsidiary and then the adjustment (if any) for implicit group financial support, while banks start with the parent’s rating and reduce it for a possible lack of group financial support. The possibility of implicit support/passive association benefits is considered in paragraphs 67 to 74 without reaching any conclusions, but simply listing the factors which ratings agencies consider in their implicit group support models (although without acknowledging this).

In a similar vein to paragraph 52 in respect of security over assets, paragraph 78 makes the controversial suggestion that it can be assumed that there is a key financial ratio maintenance agreement by a borrower to its related-party lender by reason of the visibility of the borrower’s financial information to the lender, even though such an agreement does not exist. (It is not stated, but this would have the effect of reducing the interest rate).

Paragraph 80 notes, correctly, that interest rates charged between independent parties are usually in addition to loan arrangement or commitment fees, and that this should be taken into account when deciding on related party interest rates when no such additional fees are being charged. (Here it should be added that the absence of loan exit fees should also be adjusted for, where relevant).

Paragraphs 83 and 84 note that there is plentiful information on loans between unrelated parties with which to use the comparable uncontrolled price method.

Another strange suggestion is made in Box C.7, which is that the average interest rate paid by a group on its external debt can be used as an internal CUP.

Cash Pooling

Paragraphs 97 to 99 distinguish between two broad types of cash pooling, namely “physical” and “notional” pooling. In physical pooling, the related party cash pool leader has to borrow from the bank to make good any net deficit in the group’s daily balance (although it can deposit any net surplus with the bank). In notional pooling, the bank charges or pays interest on the notional net balance looking across each company’s daily balance and this net is allocated across the participants according to a formula. The cash pool leader makes little contribution to the arrangement and the main transfer pricing issue is how to allocate the interest savings in such a way that it reflects the net contribution or burden imposed by each participant.

The two hypothetical examples in paragraphs 114 to 118 and in 119 to 123 illustrate how the reward for a cash pool leader should be determined. In the first example, the leader sets up the arrangement with a bank and has the net balance with the bank. These are not thought to be “bank-like” functions and as such, it is not thought to be appropriate for the leader to earn an interest spread like a bank. In the second example, the cash pool leader is also the group treasury company and as such it also sets the financial management strategy, manages group liquidity, raises funds for the group and makes related party term loans. It bears credit risk, liquidity risk and currency risk and manages those risks. As a result, it is concluded that it should earn part or all of the spread between the borrowing and lending positions which it adopts.

In paragraph 126, it is noted that ideally the method of sharing the benefits of pooling between the participants would reflect how much each had contributed to the total interest saving, which could vary according to the source or sources of that interest rate saving (for example, smaller net balances with the bank and/or better interest rates). Paragraphs 127 to 129 then set out three possible methods for sharing these benefits, which are: to offer a better interest rate to participants with larger balances in the pool, whether they are credit or debit balances: where the participants have similar credit profiles, to use the same interest rate for all participants regardless of whether they are depositors or borrowers: and, where there is a genuine credit risk to the depositors, to only share the benefits of pooling among the depositors. Clearly, these are very different methods and more guidance will be needed in the next iteration of the discussion draft.

The next draft of the document could also benefit from a discussion of loans and deposits with an in-house bank of up to one year, which can be repaid or withdrawn immediately and for which a liquidity premium needs to be added to the margin on the loans and deducted from the margin on the deposits.

Paragraphs 130 and 131 discuss the issue of fees for the cross-guarantees which banks usually require from cash-pooling participants. It is suggested, rather unconvincingly, that because each party does not know the amount which it will have to guarantee for each of the other participants, no fees should be charged, and rather more controversially, that any financial support arising because of the default of another participant should be treated as a capital contribution. This is an area in which international case law could usefully inform the guidance.


Paragraphs 132 to 134 discuss hedging, concluding that arranging a hedging contract for a related party is a service which should be paid for, but that if the group treasury company enters into the hedging contract itself and does not match the position with the related party in question or with any other group company, then “more difficult transfer pricing issues arise”, but no suggestions are made as to how these should be addressed.

Financial Guarantees

Section D of the discussion draft, comprising paragraphs 137 to 161, is concerned with financial guarantees. Paragraph 142 concludes that only legally binding commitments to pay merit a guarantee fee, and not, for example, letters of comfort. Paragraph 143 notes that not all bank requests for related party guarantees mean that the bank believes that its risk is reduced by these guarantees, sometimes because it may assume financial support from related parties. Paragraph 144 notes that guarantees can have a value even if they are by parties with the same credit rating as the borrower, because they increase the likelihood that a debt can be repaid. It also notes that in situations of multiple cross-guarantees, it is difficult to calculate the net benefit which one party is providing, and suggests (controversially) that as a result no fees should be charged.

With regard to the pricing of guarantees, paragraph 148 rules out the CUP method because independent parties do not guarantee each other’s loans. Four alternative methods are discussed in paragraphs 149 to 155. The first of these is the “yield approach”, which is the difference in interest rates with and without the guarantee (after adjusting for the effect of implicit financial support on the interest rate). As noted in the paragraph, this is the maximum fee which would be paid.

The second method to be discussed is the “cost approach”, which seeks to cover the guarantor’s loss in the event of default “or” the capital required to support the risks. There seems to be some confusion here as these are both elements of the same method, which is the “actuarial” or “insurance” method of calculating the minimum fee which the guarantor would require, along with a third element which is the unexpected loss for which capital also has to be set aside. This seems to be closer to the “valuation of expected loss method” proposed in paragraph 154, although there appears to be some confusion. The fourth method described in paragraph 155 is the “capital support method”, which is the expected return on the amount of capital which the borrower would have to add to its balance sheet to give it the same credit rating as the guarantor. (Clearly, this method, and the yield approach, do not work if the guarantor’s rating is not higher than that of the borrower).

Quite rightly, paragraph 153 states that the arm’s length guarantee fee will be somewhere between the outcome of the yield approach and the cost approach.

Paragraphs 156 to 161 contain two hypothetical examples of guarantees. The first example is a straightforward situation in which the impact of implicit support is to make an actual guarantee fee too high if the yield approach is used (i.e., the GE Capital Canada situation). The second assumes that prices for comparable arm’s length guarantees can be identified, which seems to be at odds with paragraph 148.

Captive Insurance

The remainder of the discussion draft is concerned with captive insurance. Reference is made to very useful insurance industry functional analysis information in Part IV of the OECD’s 2010 Report on the Attribution of Profits to Permanent Establishments. Paragraph 166 lists some tests to determine whether a captive insurance service really is being provided.

Paragraph 181 suggests that an actuarial approach may be needed in order to calculate arm’s length premiums, which would be sufficient to cover the expected loss, the administrative costs, and a return on capital (taking into account the investment income earned on the premiums).

Paragraph 184 notes that, like cash pooling (and centralised goods procurement), captive insurance can create group synergies, in this case by pooling group risk so as to obtain cheaper premium arrangements with third party reinsurers. The captive should receive a reward for its “basic” services, and the rest of the group synergy benefit should be shared out among the insured participants in the form of reduced premiums.

Paragraph 187 discusses what is similar in essence to the U.K. DSG situation, with a high street retailer being paid commissions to sell policies which are insured with a captive. Although the rate of commissions is found to be consistent with those charged between independent parties, the captive insurer is found to be too profitable compared to independent insurers. Paragraph 188 suggests that this excess profit arises from the agent’s exploitation of its “point of sale advantage” to sell the policies, and that the agent would have negotiated this excess profit for itself in an arm’s length negotiation.

We can expect the next iteration of this discussion document to provide more detail on the pricing methods taking into account the comments received in the consultation exercise, probably in early 2019.

Author Information

Danny Beeton is Of Counsel at Arendt. He can be contacted at:

Request Transfer Pricing Report