Our final NYU Tax Policy Colloquium of 2014 concerned the above-named paper by my colleague Mitchell Kane. The paper addresses current OECD efforts to address the challenges for transfer pricing that are posed by intangible assets, including the synergy value of using common control and ownership for operations in multiple countries. The following is a nice example in the paper showing one might visually depict common control synergies:
l l l l
a $75 b $50 c $75 d
Country A Country B
In this example, operations in Country A would earn $75 if separately owned and operated, as would those in Country B, but the multinational that owns both of them earns an extra $50, or $200 total, due to common control synergies.
Under transfer pricing done right, Country A will get to claim at least $75 of the company’s worldwide income (i.e., ab). Country B will likewise to claim at least $75. But what about the other $50? There is no obvious way to answer this question, whether through the analysis of “comparable sales” or otherwise.
Some argue that this example exposes a horrifying logical contradiction in the underlying “arm’s length” concept that underlies transfer pricing. How can you say what the arm’s length price would be for something that the parties can only realize by acting not at arm’s length? Others note that this can be viewed as a standard bilateral monopoly bargaining problem. That is, if one imagined separate owners of the operations in A and B negotiating over the terms of the merger, the problem would simply amount to one of how to divide the $50 surplus. Alas, this doesn’t help much, unless you think the “logical impossibility” matters for its own sake, since bilateral monopoly negotiating outcomes are context-dependent and notoriously hard to predict.
The paper is mainly engaged in legal analysis, in the sense of interpreting existing OECD guidelines rather than evaluating the policy issues against a blank slate. The chief conclusions it reaches are:
(1) The OECD shouldn’t respond to the problem by creating a novel category of intangible assets (e.g., “common control synergies” with their own set of distinctive rules),
(2) In dealings with each other, OECD members and bilateral tax treaty partners should allow pure taxpayer electivity with respect to which of the two countries gets the $50 of common control synergy income included in its tax base.
I have some sympathy with conclusion # 1, but have issues with conclusion # 2. The following is a fuller discussion of some of the main issues in the paper, including numerous points on which I agree with it, and thus am just putting my own gloss on a conclusion that it reaches as well:
1) Does the example show that multinationals with common control synergies are earning rents?
Sometimes, in international tax policy debate, common control synergies are treated as evidence that multinationals (MNEs) are earning rents. To illustrate the apparent intuition, suppose that, in Figure 1 up top, earning $75 in Countries A and B constituted a normal rate of return, given the amount of capital that is being employed. Then merging them and earning an extra $50 (for $200 total) would afford the MNE a rent.
I would argue, however, that, whether or not MNEs are earning rents, the existence of common control synergies has no obvious bearing on the question. If such synergies are significant in particular areas, that merely indicates that one would expect them to be part of the production process. Thus, in Figure 1, it’s just as natural (and perhaps more so) to assume that $200 is a normal rate of return, and that the separately operated entities would not be earning high enough separate rates of return to stay in business.
By analogy, suppose we did a chart in which we compare the amounts that multistate firms in the U.S. would earn (a) if they still used horses to ship goods, as compared to (b) given their use of trucks. We could portray them as having special “truck use profits,” but that wouldn’t tell us that they are earning rents.
Or to put it another way, consider economic sectors in which cross-border common control by MNEs would reduce productive efficiency, and we therefore don’t observe it. One wouldn’t say that in these cases the separately owned firms enjoy rents from a “separate ownership windfall.”
2) Two-country vs. three-country synergy
The paper argues that, in interpreting current OECD doctrine, it is legally correct, but also desirable as a policy matter, to let MNEs assign common control synergy income to whichever country they like, among the production sites that enjoy the synergy, so long as it’s reported somewhere. In other words, full electivity is the suggested approach.
How big an advantage is this? In Figure 1, it is plausible that Countries A and B both will have significant tax rates, since both are actual production countries. Pure tax havens tend not to have enough productive resources to support a lot of local economic activity. Thus, the benefit to MNEs of full electivity seemingly would be limited, albeit far from negligible.
But suppose the MNE can put a tiny amount of locally generated income in Country C, a tax haven that has tax treaties with Countries A and B, and is an OECD member. (The paper treats this as a precondition to accessing the benefits of full electivity.) This may open up further tax planning opportunities, as shown here:
l l l l
a $70 b $50 c $70 d
Country A Country B
Using this structure lowers pretax profits by $5. But the business as a whole still has common control synergies that include Country C, where $5 is actually being earned. But now, with full electivity, the MNE can report the synergy income in Country C and pay little or no tax on it.
3) How important is it to tax everything exactly once?
Suppose all countries are homogeneous, including in the tax rates they apply. Then taxing everything exactly once promotes global efficiency. It also means that given countries may be unalarmed by full electivity in the above scenarios, since there is no reason to think that it would systematically disfavor them.
With heterogeneity and tax rate differences, this changes. For example, there is no reason for Country A or B to think that the scenario in Figure 2 is just fine by reason of, say, the synergy income’s paying, say, a 10 percent rate in Country C. Even if high-tax countries may benefit from the use of tax havens to lower the effective tax rates that MNEs may earn on inbound investment – an issue that I can’t address here – it would have nothing to do with how many times a given dollar of income is being taxed at different rates.
Does the fact that the paper’s analysis is limited to OECD members and treaty partners mean that we are actually in the homogeneity scenario after all? Consider that OECD members which have tax treaties with the U.S. include Ireland (12.5% corporate rate), Slovenia (15%), and Luxembourg (almost 30% as a statutory matter, but generally with a much lower effective rate).
4) Transfer pricing versus formulary apportionment
In Figure 1, suppose that Countries A and B tax rates differ. This will affect decisions about where to locate economic production. But suppose instead that we treat production locations as fixed, and ask only whether taxes will distort the decision whether particular A and B operations should be owned and managed separately, or by an MNE that exercises common control. In this scenario, transfer pricing’s potential virtue is that, so long as it results in the assignment of at least $75 of the MNE’s income to each country, it will not distort the common control decision, even with full electivity, whereas formulary apportionment (FA) potentially will.
Does this provide a significant reason for preferring transfer pricing to FA, even under full electivity, so long as it is hitting the bottom of the target (i.e., assigning at least $75 to each country)? Not necessarily. For example, consider Figure 2, where it distorts location decisions because we have added to the choice set moving some operations to Country C. In addition, suppose that the U.S. stays with transfer pricing to achieve the 2-country efficiency gain, and accepts the paper’s proposal that the entire $50 of synergy income be shifted to other countries. Depending on the relevant ratio of revenue raised to deadweight loss incurred by U.S. individuals who own stock in MNEs, the U.S might be better off distorting the choice in exchange for getting some revenue.
5) A better solution than full electivity?
I am far more skeptical than the paper appears to be about the desirability of basing tax liability on formal entity lines, such as between corporate parents and wholly owned foreign subsidiaries. Why give significant tax consequences to something that is likely to have little if any economic significance?
The paper sees valuable “information” here that we shouldn’t discard. But since the entity lines are meaningless economically, the only informational content I see is that it’s predictive of how other countries will tax a given MNE, given that in practice they rely on entity lines. This in turn reflects assigning more normative weight to avoiding “double taxation” than to limiting tax planning and the creation of potentially huge tax advantages for MNEs.
If avoiding double taxation and formal double non-taxation were all that mattered, why not advocate the “ROSE” system? That’s my new acronym for “Registered Official Sourcing Election.” An MNE announces to all countries where it is reporting its taxable income, and all agree to heed this.
I might lean more towards sharing the paper’s predominant concern about over-taxation if I thought that countries were strongly prone to over-reach. For example, suppose in Figure 1 that both Country and Country B want to grab at least $125 (i.e., the local $75 plus the full $50 of synergy income), so MNEs with normal returns overall face very high worldwide tax rates. However, I am not persuaded that this is how even high-tax countries generally act. MNEs not only can wield interest group power but may be treated favorably due to well-grounded concerns about their mobility.
My preferred approach, therefore, even assuming we use transfer pricing in Figure 1 with the aim of “correctly” placing at least $75 in each of the two countries, is to use something like FA for the residual $50.