BEPS and Interest: A Bad Formula

By James J. Tobin, Esq.*

Ernst & Young LLP, New York, NY

I have to admit I am getting a bit tired of BEPS. I wasn't intending to write this commentary on BEPS. And I'm sure many readers of my commentaries (I do hope there are some readers) are tired of BEPS as well. Plus, there are lots of other interesting developments to write about, like Notice 2014-52's expansion of the definition of "United States property" for §956 purposes to include non-U.S. property or the recently proposed and wildly overreaching U.K. Diverted Profits Tax. But after reading the OECD's "Discussion Draft on BEPS Action 4: Interest Deductions and Other Financial Payments" (the Draft), I couldn't resist the urge to comment/rant about the proposals in this latest BEPS draft.

To my mind, the Draft starts from a flawed premise and combines that with a biased view of multinational corporations (MNCs) as unfettered manipulators of the current global tax system.  Add to that an ambition by the drafters to propose a drastic solution to the perceived problem, coupled with the constraints of the very tight OECD BEPS timetable for producing recommendations for all 15 Actions. The result is a radical, misdirected, unworkable proposal for allocating external group interest costs within a global group.

The faulty premise is that compensating a lender for the use of borrowed funds is somehow a base erosion or profit shifting event. To my mind, fairly compensating a lender or investor for a return on funds advanced for a borrower's use in the business is a fundamental cost of doing business for the borrower with appropriate income recognition to the lender. A fundamental principle of corporate taxation in virtually every country in the world is that there is a difference in the tax treatment of debt versus equity, with the result that interest payments are tax deductible (within limits) and interest income is taxable. So this result is neither base erosion nor profit shifting, but merely the basic tax result of the worldwide tax system.

Of course, there are opportunities for tax planning in connection with the use of debt. For example, the use of hybrid debt to obtain a one-sided deduction for interest expense in certain cases — hence, the inclusion in the BEPS project of Action 2 on hybrid mismatch arrangements. Perhaps also the use of low-tax group finance companies — which no doubt will be a major focus of BEPS Action 3 on CFC rules. The Action 4 draft acknowledges this but still continues on the premise that simple intragroup debt, or indeed the choice of location within a group for external debt, is itself an abuse.

The drafters are clearly bothered by the existence of tax rate differences across countries. In one example they conclude that a decision by a group to incur debt in a high tax rate (35%) country versus a low tax rate (15%) country produced a tax savings which they equate with a negative effective rate of taxation. Hard to understand their rationale. Certainly tax rates are one of the factors that companies will always consider in making investment decisions, and governments consider this from a tax and investment policy standpoint in setting their tax rates. This is the reason for the significant reduction in corporate tax rates worldwide over the last several decades.  The seeming premise of the drafters that finance cost should be incurred in the lowest tax location is radical indeed.

The preferred approach in the Draft is to require a proportional allocation of a group's external finance cost among members of the group based on their proportionate amount of income.  Such allocated amount would serve as an annual cap, i.e., the maximum amount of interest which could be deducted by that local entity or local country group. So no aggregate interest deduction would be allowed to an MNC group in excess of its external interest/finance costs and, because the allocated amount would serve as a cap, inevitably the aggregate tax deductible interest expense for any group would be significantly less than its external finance costs.

As far as I am aware, no country currently uses such a formulary approach. Several countries, including Germany and Australia, have adopted an "escape" clause which permits a local company to deduct finance costs above local thin capitalization or earnings-stripping type limits where it can demonstrate that its level of local debt is in line with the overall leverage burden of the worldwide group. But these are taxpayer benefit provisions and in my experience are rarely resorted to in large part due to the complexity and administrative burden of producing and proving the required proportionate group calculations.

So why the radical proposal? I mentioned at the outset of this commentary a feeling that there was a biased view and the tone of the Draft very much has that feel. The drafters repeatedly refer to the perceived ability of a group to "manipulate" its debt and equity on a local level. Statements illustrating this point of view include:

  • The potential use of "orphan entities – to disguise control of an entity";
  • "The use of foreign exchange instruments to manipulate the outcome of rules";
  • "…equity levels can be easily subject to manipulation, for example by a controlled entity issuing new share capital to its parent which does not correspond with any increase in economic activity".

I must have missed some tax planning classes as I am not sure how any of these manipulations work in practice.  (I don't know any "orphan entities", forex movements are unpredictably volatile and not a tool for tax "manipulation," and increasing an entity's equity increases its economic value, doesn't it?) But these statements and others like them seem to reveal the mindset of the drafters.

The two main areas of concern with respect to what MNCs can achieve by these "manipulations" are the allocation of external debt to higher tax countries and the use of internal leverage such that the aggregate interest deduction of a group exceeds its external finance costs. It never occurred to me that either of these was an abuse. For local subsidiaries, the decision regarding the extent of debt-versus-equity funding is a foundational issue from a tax, treasury, and corporate law standpoint, which is subject to myriad local country rules from a tax, exchange control, and regulatory point of view around the world. And where internal debt is used locally, for every unit of interest expense there is a unit of interest income on which tax must be paid, with an increasing number of countries imposing either CFC rules or interest deduction limitation rules if the tax rate paid on the interest income is too low.

Some examples are provided in the Draft to illustrate the effect of the proposed rules. I would start with a few even more simplified examples to illustrate my point:

  • Assume Group 1 has no net interest expense – a cash-rich group.  Assume parent lends 1,000 to sub 1 in country X in conjunction with an acquisition and receives interest of 100 per year. Result – no deduction in country X because the cap is zero but parent would still be taxed on 100 of interest income.
  • Assume Group 2 has 100 of net interest expense all incurred by parent in country A (a common structure from a treasury management standpoint).  Assume parent has subsidiaries in say India and China and has not been able to push any debt down to those subsidiaries due to local tax or regulatory constraints. And further assume that the EBITDA in all these countries is equal. Result – the interest cap for each country would be 33. Thus, 67 of interest would be disallowed in the parent country and the "cap" of 33 each in India and China would be unused and likely unusable.
  • Assume Group 3 incurs external interest cost of 100, all in parent country X. And assume parent pushes down debt to country A, B, and C subsidiaries and realizes 75 of interest income from each. Thus, parent X has net interest income of 125 (225 less 100). Further, assume the same level of EBITDA in each of the four countries X, A, B, and C. Result — subsidiaries A, B, and C each has an interest cap of 25 and thus each has a 50 disallowance, but parent still pays tax on net interest income of 125.

In all these cases, significant double taxation or double non-deduction results. The drafters acknowledge this potential result, but conclude it would be too complex to include mechanisms to rectify it. (A linking rule requiring corresponding tax-free treatment of the disallowed interest element would not seem to be that complex a mechanism – certainly not in comparison to the mechanism the OECD developed for hybrid mismatch arrangements under BEPS Action 2.) Perhaps they just don't have sympathy for the tax planning manipulators that have brought them to the point of requiring such dramatic action.  What jumps to my mind is one of my favorite scenes from Monty Python and the Holy Grail. "She's a witch, can we burn her?" ask the villagers of the lord after dressing up a town woman to look like a witch. Seems like the drafters have in their minds dressed up all MNCs as witch-like manipulators and are proposing "appropriate" punishment.

Happily, the Draft is just a discussion draft and it does ask lots of questions and seeks input. But I fear the questions are premised on the presumption of "manipulation" and they seek input mainly on detailed technical issues, such as the use of earnings or assets for the allocation formula, how to treat foreign exchange, and how to deal with GAAP/tax differences and timing issues, rather than on the threshold question of the merit – or not – of the formula approach. Seems kind of like decreeing that all tax advisors should be sent to prison for life but asking for their input on the color of their cells. One is tempted to answer the question – as my favorite color is blue, it would be a shame to spend the rest of my life in a green cell. But, in my view, the proposals are based on a completely flawed premise, the debate certainly should not begin with the details, and indeed the debate should never get to those details at all, ever. But you won't be surprised that I can't help myself, so here are some observations on problems with the formula:

  • An allocation of consolidated external interest expense based on relative earnings/EBITDA would likely result in huge volatility year to year and therefore necessarily a mismatch with the actual level of an entity's or country group's debt levels, resulting in lots of overs and unders relative to the permitted cap – with the result of lots of external interest with no deduction.
  • An allocation based on an earnings metric would ignore the fact that some types of income, such as services, are less dependent on capital levels than others, such as manufacturing or natural resources – again, with the result of lots of overs and unders relative to the permitted cap, particularly for MNC groups with diverse business lines.
  • The use of consolidated GAAP financial information as the denominator for the allocation fraction would require all tax authorities to understand and potentially audit an MNC's consolidated data – a scary and impractical idea given the multiple flavors of GAAP in the world and the complexity of those rules.
  • Consolidated GAAP data contains many potential distortions when compared to local country taxable or statutory account concepts, including, for example, consolidation elimination, purchase accounting step-up in assets, impairment charges/write-down in assets, etc. So, in the case of a formula that would use local tax/statutory account data for the numerator in the formula and consolidated GAAP for the denominator, there would be significant distortion in any allocation which could result in an aggregate allocation significantly above or below 100% of the external interest finance costs.
  • External interest and finance costs may not be fully apparent from the GAAP financials. Some interest could be capitalized for GAAP purposes. Some could be part of cost of goods sold. On certain instruments the amount of interest expense for GAAP purposes could differ from interest expense for tax purposes. Forex gain or loss treatment for tax and GAAP purposes can be very different, etc.
  • It is not at all clear how this would be applied to for privately owned groups that may not file audited consolidated accounts.
  • One-sided taxability of intergroup interest income with potential (virtually certain) denial of deductions for a portion of the interest expense would be bad policy and particularly impactful for groups that maintain internal finance/treasury companies.

I'll stop there due to space limitations, but you get the point. A group allocation rule is a bad, unnecessary idea. Not worth further study, but if it is nonetheless pursued, lots more evaluation would be needed with respect to how such a rule could be implemented.

In addition to the recommended external interest allocation cap approach, the Draft discusses the fixed ratio limitation approach adopted by many countries and evaluates whether, as an alternative to the allocation approach that they favor, a best practice recommendation approach with respect to the use of fixed ratios should be considered.  As is obvious from my comments on the allocation approach, I would favor this best practice alternative. Over recent years, more and more countries have adopted earnings-stripping-like interest deduction limitations based on some maximum percentage of earnings and also have been adopting other targeted deduction limitation rules. At the same time, these countries have been aggressively auditing intercompany debt arrangements, and in many cases have been showing a good deal of creativity in attempting to disallow or limit intercompany interest deductions. Silly me had been hoping that the OECD would focus on recommending a best practice approach for the fixed ratio method in order to simplify the burden on MNCs having to comply with a myriad of different percentages and formula definitions and also would admonish countries to be restrained in their zeal for new special measures – which in my mind are all aimed at merely increasing local tax revenues, often with little, if any, policy underpinning.

Unfortunately, in addition to their inclination toward the complex and impractical allocation approach, the drafters' discussion of the fixed ratio approach and other measures does nothing to encourage more consistency among countries and seems to go out of its way in a number of places to give deference to inconsistent country-by-country decisions on the specifics of limitation approaches.  Further, in reviewing existing fixed ratio approaches, the drafters observe that most countries that have adopted a fixed ratio of earnings as a limitation have set the ratio at 30% (as opposed to the §163(j) limitation of 50%). But the Draft then concludes, based on "anecdotal evidence from a number of sources…" plus a staff study of net interest-to-EBITDA data for a specific group of MNCs, that 30% is too high on the grounds that the specific group's average appears to be less than 10%.

The limited anecdotal evidence is not really provided for review and comment. The specific group of MNCs "tested" consisted of a list of Global Top 100 companies by market capitalization. A problematic approach and again to my mind a results-driven analysis for a number of reasons:

  • Local country EBITDA limitations take into account the local subsidiaries as independent entities, which will have capital structures that on a standalone basis would not necessarily mirror a consolidated capital structure.
  • The local EBITDA limitations would be based on taxable income principles, not consolidated GAAP. I question whether sufficient information could really be discerned from public financials to ascertain net interest expense or GAAP EBITDA in a meaningful manner.
  • The sample group selected represents the most valuable companies in the world. Their level of finance costs would not be remotely reflective of a broad universe of MNCs.
  • We are currently in an all-time low interest rate environment so benchmarking today's ratios to use as a fixed limitation for the future would be completely inappropriate.

The Draft does not conclude on what a "best practice" fixed ratio would be, thankfully. However, in a series of examples at the end of the Draft, a 15% level is used for illustration – seeming indication that the drafters were heavily informed by the flawed study and the anecdotal evidence shared among themselves.

So what conclusions do I draw about the work on this Action 4 (and to some extent on all the BEPS output to date)?

So what conclusions do I draw about the work on this Action 4 (and to some extent on all the BEPS output to date)?

• The OECD has been given too tall a task to accomplish effectively in too short a time. The 15 actions cover virtually every aspect of international tax and have multiple interdependencies. The drafters of each action's discussion draft to date are showing too much ambition in “fixing” the system with their own action proposals and are seemingly inclined toward more revolutionary than evolutionary change. As evidenced by the Action 4 draft, much more study and analysis is needed, as well as much more consideration of the interaction with the other actions, in order to realistically assess what “solutions” are needed.

• Proposing radical change that requires concerted country action is OK if it is just a proposal for further review and where there is some constraint with respect to country action that isn't concerted. For example, the Common Consolidated Corporate Tax Base proposals in the European Union are radical. However, as long as there is a requirement for common EU approval and adoption, there hopefully are appropriate safeguards in the system. I fear that, although the OECD encourages wide implementation of its recommendations when finalized, the reality will be that countries will pick and choose among the OECD BEPS recommendations in amending their local legislation, which will further increase the burdens and risk of double taxation on cross-border business and trade flows.

• I see no justification for, and much risk in, a group allocation approach. It is antithetical to the arm's-length standard and is a big step toward a formulary approach in the transfer pricing area. Local subsidiaries should be able to adopt a capital structure based on their own assets/activities just as a local stand-alone entity would do. The drafters presume that this would allow serious manipulation. I disagree. Benchmarking the appropriate debt levels for a separate entity isn't a difficult transfer pricing exercise—there is much data available on debt levels/credit ratings. Likewise on determining appropriate interest rates to be charged on a debt instrument. Maybe the drafters should focus instead on best practices in this area.

• A best practice approach to fixed ratios that is done realistically to put a reasonable upper limit on annual deductions and that includes some suggested guidelines for carryovers and carrybacks to limit double taxation outcomes would be helpful—if the OECD would also “encourage” consistency among countries in adopting any such rule. 

I take comfort in the fact that this Action 4 document is only a discussion draft. I'm hopeful that the final version of the recommendations under this and other Actions trend in the directions outlined above. However, I fear that my naturally optimistic nature is being tested.


This commentary also will appear in the March 2015 issue of the  Tax Management International Journal.  For more information, in the Tax Management Portfolios, see Daher, 536 T.M., Interest Expense Deductions, Maruca and Warner, 886 T.M., Transfer Pricing: The Code, the Regulations and Selected Case Law,and in Tax Practice Series, see ¶2330, Interest Expense, and ¶3600, Section 482 -- Allocations of Income and Deductions Between Related Taxpayers.


  The views expressed herein are those of the author and do not necessarily reflect those of Ernst & Young LLP.