The clock is ticking. The European Commission has a few more weeks to decide whether it will appeal the July 15 judgment of the General Court of the European Union (GCEU) in Apple Sales International and Apple Operations Europe v. European Commission, T-778/16 and T-892/16 (GCEU 2020).1 The decision is a no-brainer. The commission must appeal the judgment to the Court of Justice of the European Union.2
Apart from the considerations presented below, the most compelling reason for the appeal is that it goes to the heart of the commission’s unique status as overlord of the EU project, to say nothing of its decades-long quest for greater coordination in the field of direct taxation. This is a battle that must be fought for the sake of continued relevance. Essentially, this is the same justification offered a century ago by Himalayan adventurer George Mallory when asked why he was trying to summit Mount Everest. The commission simply must appeal Apple because it is there.3
Who’s Driving This Train?
The commission’s case before the GCEU was far from perfect. Many observers agree that the court was on to something when it observed that the commission failed to sustain its burden of proof on the issue of profit attribution to the Irish branches of Apple Sales International (ASI) and Apple Operations Europe (AOE), which benefitted from the two advance rulings at the center of the dispute. It remains a little baffling that a party with the collective resources and expertise of the commission didn’t do more to develop that portion of the case.
What would it have cost the commission to hire an independent consultant to perform a rudimentary transfer pricing analysis showing that more of Apple’s profits were properly attributable to the Irish branches? Not that much. If there are two things we know about transfer pricing litigation, it’s that outcomes are largely driven by expert testimony of this sort, and that you can find an expert witness to attest to just about anything. The commission ignored that wisdom, to their detriment. Had the usual considerations of litigation risk been at play, you’d expect a rational litigant to spend a bit of coin (say, €200,000 or so) to bring them closer to a €13 billion recovery. Normally that’s money well spent, but it didn’t happen here.
That the burden of proof was botched this badly can’t be explained away as some type of near miss. It’s not as though the judgment had to be a binary outcome, in which the Irish government was going to be forced to recoup every cent of the alleged tax benefit or nothing at all. The recovery of impermissible state aid could have been for any amount less than €13.1 billion and more than €0. Before learning of the outcome, I fully expected the court to split the baby in this manner. A mixed result would have awarded the commission far less than they were seeking, reflecting the deficiencies in their case, while acknowledging that Apple was indulging in shameless profit shifting. That prediction wasn’t even close. The zero recovery tells us that the commission didn’t just miss the target as to the burden of proof, it ignored it entirely.
What explains this foolhardiness? The best guess is that the commission omitted that portion of the case because they felt it wasn’t necessary. They never intended for the outcome to hinge on an evidence-based approach to profit attribution. You can see their point. If this kind of dispute were to boil down to competing bodies of evidence, the commission would probably lose every time. Apple can afford some pretty good tax planners; nobody is going to beat them on evidence. The commission’s best path forward is to make the case about policy — in particular, applying the arm’s-length standard in a manner that precludes stateless income. For purposes of the appeal, they should double down on the argument that interpreting the standard under EU law is entirely their prerogative.
In reading over the judgment, one could conclude that the commission is already halfway there. The court handed them some significant favors. Notable among these is identification of the proper reference system for finding a selective advantage under article 107 of the Treaty on the Functioning of the European Union. The court agreed that the arm’s-length standard is the appropriate criteria for determining whether companies that comprise part of a multinational enterprise are tax-advantaged relative to a stand-alone firm dealing with unrelated parties. That was despite the fact that the arm’s-length standard technically wasn’t a part of Ireland’s domestic law until 2010. Both Ireland and Apple must be frustrated that the case turned on how the standard is to be applied.
As the court saw it, the commission erred when it came to applying the standard because it deviated from the authorized OECD approach (AOA). Adherence to the AOA would have required analysis of commercial functions, assets, and risks — the usual suspects. But that type of conventional analysis wouldn’t cause more of Apple’s profits to be attributed to the Irish branches, which weren’t doing much beyond routine activities. Despite a cost-sharing arrangement that transferred Apple’s intellectual property rights beyond the reach of the U.S. tax base, the court’s recitation of facts made it clear that control over the IP was retained in the United States.
It does the commission no good to point out that the cost-sharing arrangement is problematic as to the roles of participants after the fact. One expects participants will do something beyond co-funding the development, but that’s not the EU’s concern. It wasn’t their homegrown IP that migrated offshore. For its part, the U.S. cost-sharing regulations are more concerned with adequacy of buy-in amounts and cost contributions, not how participants spend their time.4
Instead of relying on the AOA, the commission invented its own novel theory of profit attribution. Its exclusionary model lacks precedent in the broader transfer pricing world. It should be the job of the CJEU to determine whether that matters. The exclusionary model would tag the Irish branches with relevant profits because ASI and AOE didn’t have any foreign activities that were worth mentioning. As if by the process of elimination, Apple’s large bundle of profits had to be attributed to the Irish branches because there was nobody else around.
Think of it as the transfer pricing equivalent of how bartenders make sure that the bill for drinks always gets paid — they stick it on the last person standing after everyone else in the group has exited the bar. Under the exclusionary model, somebody in the multinational group is going to be on the hook. The approach is crude and unconventional. Time will tell whether it proves effective.
The salient difference between the commission’s approach and the AOA lies in their default positions. The exclusionary model — to the extent that one can understand it — takes stateless income off the table by ensuring that profits are attributed somewhere. The details of how that occurs seem to be of secondary importance. The attribution result has priority over the sanctity of the attribution method. To do that, the exclusionary model refuses to get hung up on analysis of functions, assets, and risks. It wants us to regard the omission not as a failure but as a virtue. Whereas the AOA enables stateless income under the right circumstances, the exclusionary model prevents it altogether.
In fairness, it must be acknowledged that the exclusionary approach is not what most people have in mind when they speak of the arm’s-length standard. The court rejected the exclusionary model outright but did so without adequately explaining why the AOA should hold sway in the first place. It’s on this critical point that one detects a fuzziness in the court’s reasoning. You can read over the court’s opinion several times and still not understand who put the OECD in charge of determining how the arm’s-length standard must be interpreted for purposes of EU law. Sure, the AOA can fairly be described as an accepted international standard. So what? Its usage elsewhere doesn’t necessarily require that it be accepted here, for purposes of applying article 107. That’s an acute vulnerability that could be unsettling to the CJEU. The issue must be tested on appeal.
The court also found error in how the exclusionary model is out of sync with Irish domestic law on the attribution of profits to a permanent establishment. Here, the court turned to Irish case history, holding that foreign profits shouldn’t be attributed to a local branch unless it exercises control over the assets in question. We’ve already determined that the Irish branches of ASI and AOE weren’t controlling the IP responsible for the profits in question. It’s here that the branch’s shallowness works in its favor.
However, it’s fair to ask whether the court’s reading of old Irish tax cases is up to snuff. The court relies on previously obscure case law, the Dataproducts LTD decision from the 1980s, for the principle that Ireland wouldn’t normally attribute these profits to the branches. I suspect the commission’s legal team is licking their chops at the opportunity to distinguish Dataproducts LTD from the present facts. That’s a second vulnerability that should be properly vetted on appeal. Setting aside the whole spat between AOA and the exclusionary model of profit attribution, the commission could prevail on appeal by showing that the GCEU misapplied Irish domestic law. Pinning €13 billion on the relevance of Dataproducts LTD seems a bit shaky. Again, the CJEU should have the final say.
To Quit Now Is Premature
There’s something to be said for coordinating national tax systems within the EU bloc. It’s supposed to be a single economic market, after all. Tariff and trade policies are harmonized, which gives member states a collective clout that would be unattainable if they acted individually — notwithstanding U.K. fantasies about post-Brexit trade policy. Taxation is different, of course. Member states retain primary competency over the taxing function. They’re reluctant to give up the slightest pinch of sovereignty, but EU-level synergies are there to be had. Think how disjointed the VAT environment would be if indirect taxation weren’t coordinated to a large degree. People will always complain about the EU VAT directives, but the headaches would be worse without them.
Achieving that level of coordination for direct taxation has proven next to impossible. That’s not for lack of trying. The common consolidated corporate tax base is a fine idea that will never see the light of day as long as qualified majority voting doesn’t extend to direct taxation. That’s why the state aid doctrine is necessary. Its presence has a chilling effect on the darker impulses of member states, which individually don’t care about tax-induced market distortions as long as they can be twisted to their myopic advantage.
If the CCCTB had been up and running from the early 1990s, would some member states have been able to develop their prolific advance ruling programs? Probably not. We’re not talking about rulings that merely provide useful certainty to taxpayers (think of private letter rulings as used in the U.S. tax system) but rulings whose sole purpose is to barter away pieces of the corporate tax base in a nontransparent manner for other perceived benefits. One likes to think the CCCTB might have prevented the rise of EU tax havens, and thereby deterred Ireland from granting these curious tax rulings that enabled Apple to accrue stateless income.5 Brussels doesn’t have the CCCTB in its quiver, but it does have the state aid doctrine. It needs to be periodically used, boldly if necessary, to keep multinationals honest.
The commission has been invested in the state aid doctrine for many years, including as a regulatory tool over tax policy. The depth of this investment has been amplified in recent years, such that it makes no sense to back down at this point. The dispute over Apple’s tax rulings has garnered sufficient attention that it has become emblematic of the commission’s regulatory ambitions and all that article 107 is capable of achieving.
When confronted with this kind of existential battle, the two choices are to emerge victorious or die trying. Quitting shouldn’t be an option. The GCEU isn’t even Europe’s highest court. Here, anything other than an appeal feels like an abdication of responsibility.
Decision time is fast approaching. Will the commission choose to be meek or muscular?
FOOTNOTES
1 For prior coverage, see Ryan Finley, “EU Court Rules Against European Commission in Apple State Aid Case,” Tax Notes Int’l, July 20, 2020, p. 301. See also Lee A. Sheppard, “What About Cupertino?” Tax Notes Int’l, July 27, 2020, p. 429.
2 For prior analysis, see Finley, “Will the EU Commission Cut Its Losses After the Apple Decision?” Tax Notes Int’l, Aug. 3, 2020, p. 602.
3 I note that Mallory suffered a miserable death in the process.
4 The final report on base erosion and profit-shifting actions 8-10 addressed the substantive role of cost-sharing arrangement participants. The changes have been adopted into the OECD transfer pricing guidelines, but the U.S. cost-sharing regulations haven’t changed. Nor are they likely to given U.S. views on risk reallocation and recharacterization.
5 Arguably you could say the same thing about BEPS action 5, addressing the transparency of advance tax rulings.
END FOOTNOTES