O.E.C.D. Discussion Draft On Financial Transactions – A Listing Of Sins, Little Practical Guidance

INTRODUCTION

In early July, the O.E.C.D. Centre for Tax Policy and Administration (“C.T.P.A.”) released Public Discussion Draft on B.E.P.S. Actions 8-10: Financial Transactions (the “Discussion Draft”). The Discussion Draft addresses financial transactions (e.g., loans, guarantees, cash pools, captive insurance, and hedging). Like many of the other initial B.E.P.S. Project drafts, the Discussion Draft does not represent a consensus among the O.E.C.D. Member States and requires commentary, input, and further work before becoming a chapter in the O.E.C.D. Transfer Pricing Guidelines.

THE O. E.C. D. IS LATE TO THE TABLE

A reader might wonder how or why the Discussion Draft has emerged at this particular point in time, as tax authorities and legislators in various countries have already provided guidance on a unilateral basis. In a short recap of developments in financial transaction transfer pricing, we arrived at this point in approximately the following chronology:

• 1972 – Decision in Mixon v. Commr., 464 F. 2d 394 (5th Cir 1972), establishes thirteen factors that can be determinative of bona fide debt
• 1998 – Decision in Laidlaw Transportation Inc., et al v. Commr., T.C. Memo 1998-232, further establishes that certain of the thirteen factors can be determinative of bona fide debt
• 2009 – DSG decision from the U.K. Tax Tribunal on captive insurance pricing
• 2009 – “Implicit support” of a subsidiary by a parent (discussed below) emerges from the GE CapitalCanada case
• 2010 – GE Capital Canada “guarantee fee” case decided for the taxpayer on appeal
• 2013 – The B.E.P.S. Project decides not to abandon the arm’s length principle in favor of formulary apportionment
• 2014 – Draft reports are issued on B.E.P.S. Actions 4 (interest deductibility), 8, 9, and 10
• 2015 – O.E.C.D. C.T.P.A. signals the start of a project on transfer pricing for financial transactions
• 2016 – The U.S. issues proposed regulations under Code §365 issued on treatment of related-party indebtedness
• 2017 – McDonald’s loan case decided in Spain in favor of the tax authority
• 2017 – Chevron loan case decided in Australia in favor of the tax authority
• 2017 – Hesse Norge A.S. loan case decided in Norway in favor of the tax authority
• 2017 – Adverse S BV loan case decided in Sweden in favor of the tax authority
• 2017 – The O.E.C.D. issues revised O.E.C.D. Transfer Pricing Guidelines (the “2017 Guidelines”)
• 2017 – The U.S. Tax Cuts and Jobs Act amends the Code §163(j) business interest deduction limitation
• 2018 – Exxonmobil Production Norway Inc. loan case decided in favor of the tax authority
• 2018 – The Discussion Draft is released for comment

As the reader can see, the Discussion Draft is somewhat late in providing guidance. This is unusual for the O.E.C.D., which typically provides transfer pricing guidance prior to tax law, administrative guidance, and jurisprudence. Stated differently, O.E.C.D. guidance generally has been issued where little authority existed, and in a multilateral context, it has provided direction to minimize double taxation.

The Discussion Draft arrives at a time when several of the key questions in financial transaction pricing have been settled in a substantive way, though not necessarily by all O.E.C.D. Member State tax authorities. Any resulting O.E.C.D. guidance will be applied in conjunction with already existing tax law, administrative guidance, and jurisprudence when determining the appropriate treatment of a controlled financial transaction. The potential for double taxation may arise where treaty partners give varying deference to particular O.E.C.D. guidance. The lack of consensus amongst O.E.C.D. Member States on the Discussion Draft may foreshadow difficult double-tax cases between competent authorities.

CONDITIONS FOR RECHAR ACTERIZING DEBT

The Discussion Draft deals at length with the conditions that must exist before a treaty partner may recharacterize a debt instrument or a guarantee as equity and the means by which recharacterization could be achieved. It suggests that descriptions of recharacterized outcomes will be a focus of future work. The Discussion Draft strongly signals that the tax authority’s view on financial transactions appears to skew toward transaction recharacterization, and away from providing guidance that will help companies characterize financial transactions appropriately at the issue date or help tax authorities adjust a transaction price in a reasoned way. This likely will be an area that attracts significant industry comment and demand for examples and guidance on how to be compliant given different fact patterns.

It is not uncommon for foreign tax authorities and transfer pricing practitioners to give the 2017 Guidelines the deference of enacted law. However, in the Discussion Draft, one could argue that no deference is appropriate. There are too many comments concerning the behavior of independent parties that are not confirmed with empirical evidence. Rather than making law, the document is better construed as making recommendations on best practices.

An example of an unsubstantiated statement is found in Box B.4 of the Discussion Draft. It states that if a lender lacks the functions to manage lending, it should receive a risk-free return while the managing entity receives the residual return. This is contrary to what is seen in the market. Companies and individuals pay fund managers a fee to invest capital because, ultimately, the risk to capital resides with the individual or entity investing the funds. An investment advisor may be unhappy that its client lost 50% of their investment in a risky venture, but it is the client that loses that capital, not the advisor. If this approach is adopted, we may see asset and risk considerations take a backseat to functions, contrary to market evidence.

Similarly, cross-guarantees are stated in paragraph 131 of the Discussion Draft to have no value despite the well-established practice of banks requiring cross-guarantees on material loans. Collateral is also stated to be valueless in the related-party context in paragraph 52, as ownership of shares is assumed to imply ownership and control of assets.

CREDIT RATINGS

Credit ratings and their determination are discussed at length in the Discussion Draft. Several controversial ideas on how to calculate and apply credit ratings in an intercompany context are advanced.

Credit ratings are issued either for a company or for a specific issue of debt, not a corporate group in the aggregate. The rating tells the market what the odds are that a borrower will meet its debt obligations. While ratings are not issued for a corporate group, certain market participants may make the simplifying assumption that group members share the same credit rating.

In Box C.2, the Discussion Draft asks commentators to entertain a controversial presumption that an independently derived group credit rating may be taken as the credit rating for each member. The question is whether this would be useful for tax administrations and tax compliance. The answer is no, unless multiple nations agreed to create a safe harbor of this presumption. Why? Because the assumption fails to hold in important ways. Except in rare circumstances, a subsidiary can never have a higher credit rating than its parent company. A company can have less liquidity than a parent or sister company, greater relative debt service burdens, or sovereign factors that make it more prone to default. Therefore, the assumption of a group credit rating fails when tested, unless one makes strong assumptions about implicit support.

IMPLICIT SUPPORT, OR THE CREDIT ” HALO EFFECT”

The Discussion Draft takes the presumption of implicit support as a given. Implicit support, which assumes a Multinational Enterprise (“M.N.E.”) group member is too big to fail and therefore any default would be backstopped by the M.N.E. group or parent, is not supportable in an arm’s length analysis. The Discussion Draft starts with passive association but makes a logical leap to the assumption that (i) due to the importance of an entity to a group, it would be bailed out, and as such, its creditworthiness should be somehow elevated and (ii) this credit enhancement is not compensable. This ignores the O.E.C.D.’s consensus in the 2017 Guidelines at 1.159 on synergistic benefits of group membership:

This assumption, which was stated in a footnote, may have been overlooked in the rush to issue the Discussion Draft. However, the prior guidance is still logically sound. The 2017 Guidelines suggest that deliberate support in a financial transaction context is compensable. However, the Discussion Draft indicates that a guarantee exists by default, for which no compensation is warranted.

Many descriptions of implicit support are akin to hand-waving exercises for which hard data does not exist. The Discussion Draft disparages the use of bank opinions at paragraphs 92-93 as being a departure from an arm’s length approach, yet a bank opinion is likely the most credible evidence of a quantification of implicit support. In fact, in an example that references the 2017 Guidelines at paragraph 1.164, implicit support is predicated on what seems to be a bank opinion. It also fails to clarify how a potential financial bailout does not represent a (compensable) commitment of assets as described in paragraph 1.159 of the 2017 Guidelines.

Further, the Discussion Draft provides no guidance on how to measure the credit rating impact of implicit support. In the examples given at 1.164-1.166 of the 2017 Guidelines and at paragraphs 157-159 of the Discussion Draft, the credit rating effect of implicit support is simply assumed, with no guidance on quantitative estimation. Taxpayers are asked to quantify implicit support in the absence of concrete guidance, akin to asking a company to describe an unknown counterfactual state. Regrettably, the world of tax compliance places a low priority on forgiveness of flawed prior assumptions.

This lack of detail will make competent authority proceedings difficult for both tax authorities and taxpayers, where achieving relief from double taxation will be inhibited by a lack of common approach founded on reliable principles.

PRACTICAL GUIDANCE AND EXAMPLES ARE LACKING

The Discussion Draft lacks practical guidance on how to evaluate transactions. Though there are some useful points, the Discussion Draft does not offer guidance on how to analyze transactions to determine if adjustments are necessary and how to adjust terms to achieve comparability. Worse, in some cases, the proposed approach is inconsistent with arm’s length practices.

In paragraphs 62 through 66, the Discussion Draft does contain a reasonable discussion of the factors taken into account to arrive at a credit rating. The comment at paragraph 62 correctly observes that it is challenging to estimate a credit rating for certain entities (e.g., start-ups, special purpose vehicles, etc.). While this is true, as is the statement that independent lenders would conduct a due diligence process, there is no guidance or request for comments on what that process looks like practically. In effect, the draft says it is easy to sin here but doesn’t give guidance on how to stay on compliant.

A useful observation is made in paragraph 65. It suggests that related-party transactions can influence any quantitative ratios and should be adjusted. Practitioners will be looking for some qualification to these observations to say that, to the extent controlled transactions influence the credit rating, those transactions must be shown to be at arm’s length for the credit rating exercise to be reliable. Making such a qualifying statement and emphasizing the order in which transactions are examined would be important and would assist companies computing synthetic credit ratings and tax administrations evaluating the computations.

The insurance industry is as close as possible to the ideal of a transparent pricing model. In Box E.2, commentators are asked whether an actuarial analysis is an appropriate method for determining non-arm’s length premiums. There is a widely-promulgated set of general methodologies available to actuaries, with available data, clear assumptions, and guidance on their application. To the extent, related-party transactions are truly insurance transactions and actuarial models exist that fit the transactions (as proposed in paragraph 166), the O.E.C.D. clearly has the opportunity to advance a robust principles-based arm’s length pricing approach. We expect some debate from commentators about when such an approach is warranted. Such debate would indicate good progress toward multilateral guidance.

The Discussion Draft makes several statements about the consequences of recharacterization that are somewhat impractical. For example, in paragraph 140, the proposition indicates when a related party receives a guarantee that enhances not just its credit rating but also raises its debt capacity, a portion of the borrowed funds should be deemed to have been borrowed by the guarantor and considered a capital contribution to the borrower. There is no basis, of which we are aware, in the arm’s length market. Further, how tax administrations and M.N.E.’s would go about executing a recharacterization is unexplained, likely for good reasons. A guarantee can be considered analogous to insurance for a lender. The proposed approach implies that, in some circumstances, the insurer should gain an equity interest in the borrower or policyholder in exchange for its pledge to the third-party lender. Surely, a more practical approach would be to price the impact of the credit enhancement and then price the value of the debt capacity enhancement, both of which are feasible exercises that would acknowledge the fact that the guarantor is not actually borrowing from the lender.

One might question whether some of the “solutions” to common problems proposed in the Discussion Draft should have made it into a document for public commentary. For example, the question to commentators in Box C.7 is a request to identify situations in which an M.N.E. group’s average interest rate paid on external debt could be considered an internal “C.U.P.” (comparable uncontrolled price). The answer seems to us to be clearly never, if we were to recognize that an average is derived from more than one number. Borrowing is highly dependent on the term of the loan, the date of the transaction, the creditworthiness of individual borrowers, etc. And, stepping back to definitions, a C.U.P. is an average of prices that, absent comparability considerations, does not constitute a price. At best this would be an alternative method, without logical underpinnings. The lack of justification for this proposition undermines the creditability of the Discussion Draft.

HINDSIGHT AND TIMING IN FINANCIAL TRANSACTIONS

Financial transactions are entered into at a specific point in time. They are based on the best quantitative and qualitative data available at this time. In hindsight, transactions can look unnecessary or excessive, such as insurance premiums paid for a fire that never occurs. The Discussion Draft acknowledges the importance of timing when it considers economic circumstances but fails to give this factor its due consideration.

The ability to manipulate the timing of a financial transaction – even within a given year – can lead to significant changes in the effective interest rate and should be of concern to tax administrations. To counteract this impact, certain requirements could be imposed, such as a requirement to demonstrate that credit analysis occurred before the issue date or that fund movements inform pricing dates and the requisite analyses.

Pricing is incredibly sensitive to timing and transactional terms can have a significant impact on interest rates. The discussion regarding the ready availability of loan data in paragraphs 83-84 does not address the fact that loans are not liquid or traded instruments and that loan data may not always be available or relevant. This is the reason why bonds, with clear terms that generally are consistent with the terms of loans and trade at volumes resulting in the reporting of pricing and other issue data, provide a practical alternative to loans when considering a source of pricing data.

Use of bond data in concert with credit ratings also addresses the fact that credit ratings are issued for issuers and securities that are actively traded. Public companies and bond issues are the primary sources of credit ratings. The Discussion Draft does not raise the comparability risks that may arise in using a credit rating to determine an arm’s length interest rate on an illiquid intercompany loan. The draft skirts around the edges of this difference in paragraph 63, discussing how banks and alternative lenders utilize their own models for determining credit worthiness in special circumstances. However true, these models are specialized and proprietary and, therefore, not helpful when companies and tax authorities attempt to verify the pricing approach the other has taken.

Somewhat surprising is the Discussion Draft’s lack of caution against the use of hindsight. Within the 2017 Guidelines, there are eight references to hindsight and the care required when this approach is used. Hindsight and restructuring or recharacterization go hand in hand. Paragraph 1.123 of the 2017 Guidelines cautions that restructuring of legitimate business transactions would be wholly arbitrary and lead to inequity. Therefore, great care should be taken in “delineating” between debt and equity where clear forward-looking guidance is not provided by tax administrations. This guidance is quite common and takes the form of debt-equity ratio rules, E.B.I.T.D.A.-denominated thresholds, or other mechanisms.

SUMMARY

Perhaps a plain-spoken Discussion Draft for the penitent multinational could have stated that manipulation of credit ratings can significantly impact interest rates and then recommended certain approaches to prevent companies using pricing techniques that are recognized as either non-arm’s length or gross simplifications. This could have been followed by some illustrative numerical examples. A “safe list” of commercially common terms to a loan transaction, such as prepayment terms, lack of security, and liquidity requirements, would have been helpful. A “less safe list” could have been established based on how exotic a particular term is in the market and how much judgment is required to price the marginal effect of the term in practice.

The O.E.C.D. released a final report on how to value “H.T.V.I.” (hard-to-value intangible assets) at the same time as it released the Discussion Draft. The H.T.V.I. report was intended for use by tax authorities. The preface to the Discussion Draft does not exclude multinational corporations as a user group. Nonetheless, the O.E.C.D. has clearly signaled that it intends to depart significantly from both well-established financial transaction pricing practices and standards that are market-based or can be understood with reference to market data.

For U.S. subsidiaries of parents resident in an O.E.C.D. Member State, the Discussion Draft suggests that, eventually, loans may be accorded different treatment depending on the jurisdiction of the borrower. This may require significant modification to a multinational company’s generalized global transfer pricing policy for financial transactions at the country level.

One might question whether the Discussion Draft trades clarity, for tax authorities and companies, for adherence to the arm’s length principle. Commentary is invited on a number of potential approaches that are not arm’s length as we have pointed out. We expect companies will tell the O.E.C.D. that this compromise in principles is not needed. We hope the O.E.C.D.’s aggressive opening bid will be met by a request for a return to principles and practical guidance that can coexist with country tax law and the behavior of financial market participants.