INTERVIEW: Ednaldo Silva

Ednaldo Silva is an economist and transfer pricing expert, with a PhD from the University of California at Berkeley. He was a senior economic adviser at the Internal Revenue Service, working with the Advanced Pricing Agreement program. He founded RoyaltyStat LLC, a royalty rates database of transfer pricing comparables. He is now the managing director of EdgarStat LLC, a transfer pricing analytics firm. We spoke on April 30. (Note: RoyaltyStat, now part of Exactera, is a sponsor of this newsletter.)

The first thing I wanted to talk about, how the arm's length standard is surviving in this new digital economy where comparables are much different than they than they used to be. At least, that's what you often hear, so I'm curious for your thoughts on that.

The challenge of the digital economy is primarily the location of who is recognizing the taxable income. Is that the source or the destination, is that the location of the service, or the location of the consumer? There are some elemental issues that need to be resolved before we discuss comparables. My sense is that the digital economy must invoke the combined profits split, as opposed to comparable transactions. To me, the digital economy invokes a profit split regime. I do not mean the residual profit split based on comparables. I mean the profits split, based on value added factors like payroll.

Interesting, we talk about the split between the location of the transaction and location of the consumer, because that's not something that's new. They debated that 100 years ago, how does the new economy affect that?

Well, it is true that it's an old debate, but my understanding is that the taxable income is recognized in the jurisdiction in which the service provider, or the manufacturer, or the distributor, or the retailer is located. And so, this debate is reopened with the digital economy, because you can have one central location with users or consumers in many jurisdictions. And the service provider has many options for where to recognize income, including recognizing income in a jurisdiction in which the servers are not located and in which there are no consumers. You could have a triangle. So, it is true that the debate is old, but with the digital economy it becomes interesting.

You said that you think the way things are going, there will be a move away from transfer pricing based on comparables and towards more profit splits in this new economy. Can you elaborate on why you think that is?

I think that it creates no drama. Because the concept of arm's-length is an axiom. And we can define it any way we want. The colloquial expression "arm's-length" is mistranslated in French. It's mistranslated in Spanish as a principle of competition. But the arm's-length standard defined in terms of comparables, with respect to function performed, and geographic market has little to do with competition.

So, we can define it in terms of limits to deductions. Which is, I would say, the prevalent regime in most developing countries, in which you control related-party transactions by limiting the deduction to royalties, limiting the deduction for service fees, limiting the deduction for interest payments. It can go to its limit, which is to disallow any intercompany deduction. You also have a system of safe harbors, which can be conceived as arm's-length. You just need your trading partners to help you with the revised concept. Safe harbors are already acknowledged in the U.S. regulations, and also the OECD's guidelines.

But when you have the mixing of accounts, such as expenses, and receipts, or a situation in which the location of the value-add is fluid, those situations call for the combination of the profit, and then the splitting of the combined profits. Now, if you apply the unrealistic U.S. insistence on splitting the profit as a residual--I say unrealistic because that presumes that you have at least two sets of comparables, if you have a party and a counterparty. To use the residual profits that you need comparables for the party and for the counterparty so that you can determine the routine profits. So my suggestion is, comparables are very difficult to find in an oligopoly economy. We should combine the profit and split by the fact that determines value-added and that is payroll.

Because it's ultimately the people who are creating the value, is what you're saying.

Since classical times, the first economists, who were either medical doctors or lawyers, conceived of value-added as payroll and a residual, which was profit. But the profit was the price of the product, minus the wages or the payroll, the residual is what profit is. The determinant of value-added is payroll. Now, you may say, how are you going to deal with the economy that's robotic? That's a complexity we can deal with at a later stage. Because this is also contemplated, as strange as it may seem, by the classical economists, by developing the concept of direct and indirect labor. It's a very old concept.

You were talking about safe harbors and limits on deductions, I wanted to ask you about Brazil and their the way they've changed their system to align with the OECD. That is what their system was based on, if I understand it. Do you have thoughts on that on what's going on there right now?

You're correct that Brazil had a mixed regime, it had the regime of limits to deductions of royalty payments, which is always complex to enforce because tax planners embed the payment. For example, you have a 5% limit on the deduction of the royalty payment. You embed the payment. You want to sell some service, the cost of goods, you embed it, so it becomes undecipherable, difficult to audit. So that is a challenge. It was mixed with limits to deductions and a system of safe harbors for the transfer of tangible goods. And they decided to switch to the OECD regime based on comparables and they are going to have the same problem that United States has, that all the OCD countries have. Which is that comparables have become the source of controversy.

Why is it that comparables are so controversial?

It's because the regulations contemplate an economy that does not exist. The regulations contemplate an economy that's competitive. And in this regard, competition is characterized by two attributes. One is a large number of entities producing a homogenous product or service--by homogenous I mean, with similar substitutes. Competition for an economist, or in economics is equivalent to the statistical law of large numbers. So that is one attribute, the other attribute is that you have free entry and exit. These two attributes do not reflect the reality of any country, whether it is rich OECD countries, or BRICS [Brazil, Russia, India, China and South Africa] countries or even poorer countries. All countries are dominated by oligopolies. Except for non-traded goods, for example, the baking of bread, local construction, residential construction. They are not subject to international trade. But goods that form the basket of the Consumer Price Index, or the Producer Price Index primarily reflect the output of oligopolies.

When you say oligopolies, what do you mean by that?

An industry that is represented by a small number of entities in which the products are differentiated. They may be substitutes but they are differentiated. Because if the product is differentiated, the prices are not the same.

One thing I'd wanted to ask about, we've heard that there might be more use of the economic substance doctrine, at least in U.S. transfer pricing and maybe around the around the world. I know that's not technically a transfer pricing concept, but I'm wondering if that's going to influence transfer pricing, when you have the government looking beyond structures and evaluating what the substance is beneath them.

Economic substance is part of the US transfer pricing regulations. And also the OECD guidelines. When we do a transfer pricing audit, we presume that there is economic substance in the controlled transaction. But there are at least two conditions in which economic substance trumps the transfer pricing analysis, meaning you don't even go to transfer price, because the transaction lacks economic substance. One scenario, which I see quite often, is when the counterparty does not have any payroll. For example, you have a distributor or manufacturer or retailer, that has a factory entity. It's a special purpose entity to factor the receivables, but it has no payroll. And when you look at the profit margin, it's close to 100%, because it has no cost. So that kind of entity is a fiction, it's legal fiction. It has no economic substance. And that kind of transaction should be disallowed. We are not even at the level of transfer price. Tax planners are extremely aggressive, and unrealistic, because they have presumed that the tax authority cannot detect this flaw.

So that's one scenario. The other scenario is when you have an integrated entity, for the example a manufacturer, and the distributor. And you split those two functions, you create two separate entities. But when you create the entity, you forget the test, the present value test, that the entity that is now dismembered must have a present value of streams of profit that cannot be lower than what it was prior to the reorg. Let's say that you and I belong to the same group. You're the manufacturer and the distributor, or even the manufacturer and the R&D service provider. And tax planners come to us and say, you should create two separate entities because we have, in terms of tax savings, an opportunity. And then you or me, as managers of the group, say, well, to survive audit the plan must be designed in such a way in which the profit of the manufacturers, the major element of the group is not decreased. Because the act involves more than one year, you have to take the present value of the stream of profits. You take the present value of the stream of profits of the integrated entity, and you compare to the present value of the stream of profits of the members of the bifurcated entity. What happens is that this principle is usually violated in corporate reorganization for transfer pricing purpose. I think if the tax authority were careful in its enforcement, they would impeach many corporate re-orgs, because they violate the economic principle that a corporate reorg cannot decrease the stream of profits of the entity being dismembered.

That makes me think of the whole concept of DEMPE [Development, Enhancement, Management, Protection and Exploitation] functions that the OECD introduced, I guess it's about 10 years ago, as part of the first BEPS project. It has similar concepts of looking through to what the people are doing, and if they are involved in the value driver. I'm curious if you think that's has been affecting transfer pricing.

Well, let us go back to how this problem was conceived, in the '92/'93/'94 U.S. regulations. The temporary and then the final U.S. regulations. We have a problem regarding the transfer of intangibles, in which there was a mismatch of benefits and cost or risks. For example, you have a foreign intangible owner that did not pay for the promotion, development, and protection of the intangible in the US. There was a famous example called Fromage, that in subsequent revisions, it was removed from the US transfer pricing regulations -4, regarding the transfer of intangibles. So, there are three scenarios that were contemplated. GlaxoSmithKline, that was settled in '06, was fundamental on this issue, on this controversy.

There were three options that were contemplated. One, we found a mismatch, of the owner of the intangible being foreign, but the intangible is being developed in the U.S., and the expense to develop the intangible is booked in the U.S. And the US entity benefits from that deduction, and so does the group. One scenario was to disallow the, for example, the R&D with the advertising expense of the U.S. entity. In the example, you have a drug that is well known in Europe, the R&D is recognized in Europe, and when it's transferred to the U.S., it's not known, it's not authorized. The U.S. entity books all those expenses and to promote the drug in the detailing. So then, you disallow the local expense, meaning you relocate to the owner of the intangibles. So then you have a match of benefits and expenses such that the expenses are used to offset the royalty income. That's one option which the U.S. was pursuing and Glaxo, but subsequent to that it has recanted by discarding the Fromage example. The other option is to say that the U.S. entity is a co-owner of the intangible. So, the intangible is owned by the foreign entity, but to the extent that the U.S. contributed to the recognition development of the DEMPE functions, the U.S. is the co-founder. So the payment, the royalty payment has to be reduced, has to be adjusted downward, to offset the U.S. expenses. And the IRS also pursued this theory in Glaxo. And then the third alternative, which is what is prevalent today, both in the U.S. and the OECD, is to say that the entity that is not the owner is providing a service to the owner of the intangible. And the service is classified as DEMPE, it can be any of the DEMPE functions. And the services must be reimbursed, plus an arm's-length markup. This is the position that today is dominating.

Contact the author at amparkerdc@gmail.com.