Tax Notes legal reporters Ryan Finley and Kiarra M. Strocko discuss the EU General Court’s judgments in Amazon and Engie and their effect on state aid rules.
David Stewart: Welcome to the podcast. I'm David Stewart, editor in chief of Tax Notes Today International. This week: the state of state aid. On May 12 the EU General Court released its judgments in two tax cases: Amazon and Engie. In the first, the court annulled the European Commission's 2017 ruling that advanced pricing agreements approved by Luxembourg resulted in €250 million in state aid to Amazon. It held that the commission's transfer pricing analysis in the Amazon case was fundamentally flawed.
In the Engie case, the EU general court upheld the European Commission's finding that Luxembourg's tax rulings allowing the French fuel supplier to categorize transactions as both debt and equity violated state aid rules.
What do these judgments in both cases mean for the commission's efforts to reign in tax avoidance through state aid rules? Here to talk about it are Tax Notes legal reporters Ryan Finley and Kiarra Strocko. Ryan, Kiarra, welcome to the podcast.
Ryan Finley: Thanks for having me.
Kiarra Strocko: Thanks, Dave. I'm excited to be here discussing these two very important state aid cases.
David Stewart: Ryan, let's start off with you. Can you tell us a bit about the Amazon case and its journey to the EU General Court?
Ryan Finley: The Amazon case is about an APA that was originally issued in 2003. The APA dealt with the license of Amazon's technology and other intellectual property from a Luxembourg partnership, which was not subject to any entity level tax, to a Luxembourg operating company, which was a taxable corporation. That corporation was the one that actually actively oversaw Amazon's EU business.
Under the APA, the royalty payable under this license was calculated indirectly using a method called the transactional net margin method (TNMM). In this case, it gave the operating company a markup on its operating expenses and allocated all the residual profits from Amazon's EU operations to the partnership as a royalty.
The commission thought this left the taxable operating company with too little profit and the nontaxable partnership, which in this case didn't even have any employees, with too much profit. The commission made a bunch of arguments as to why this represented illegal state aid, but its primary argument had to do with how the TNMM was applied.
The way the method works, it gives one of the parties, called the tested party, some kind of fixed return based on a comparison with the profitability of functionally comparable companies. And as a one-sided method, any residual profit or loss remaining after that return goes to the other party by default.
The commission's argument was that the partnership and not the operating company should have been the tested party. It would have been the partnership that only received a kind of fixed routine return while all the residual would remain with the operating company.
Under the OECD transfer pricing guidelines, which the EU courts have accepted as guidance in these state aid cases involving transfer pricing, the tested parties usually supposed to be the less complex entity. That's because it's typically easier to find comparable companies for a less complex entity.
Owning unique IP or bearing IP development risk like the partnership did in this case generally means that that party can't be the tested party. The commission still argued that the partnerships should be the tested party because it should not be treated as the real owner of the IP. It argued that this was the case because it was the operating company, not the partnership, that performed and controlled all the key functions and risks associated with IP.
Now, importantly, the commission had to cite concepts from OECD guidance that wasn't issued until years after the 2003 APA was issued to make this argument. The court didn't go for that.
The decision says you can't judge a 2003 APA based on standards that came out years later, and that kind of killed the commission's main argument because there really isn't much in the 1995 version of the OECD transfer pricing guidelines that would support disregarding IP ownership in this way.
The commission also added some alternative arguments, including that the profit-split method, and not the TNMM should have been applied, but the court rejected these on kind of burden of proof grounds. Basically the court said that the commission had to justify its own methodological choices and actually prove that the changes of method led to a tax advantage. Since the court said they didn't do that, the commission lost.
David Stewart: Well, Amazon and its Luxembourg subsidiary having transfer pricing issues rings fairly familiar. Are these the same entities that were involved in the dispute that was recently appealed in the U.S. courts?
Ryan Finley: Yes, it does actually. The partnership was a participant in a cost-sharing arrangement with its ultimate parent Amazon Inc. That cost sharing arrangement was the exact issue in dispute in the U.S. case. That case was about whether the IP that the U.S. developed and partially transferred to that Luxembourg partnership was valued appropriately. This case is about whether the royalty that the partnership charged for sublicensing that IP was valid.
David Stewart: OK. Well how did the parties in this case — Luxembourg, Amazon, and the European Commission — how did they react to this decision?
Ryan Finley: Well, as you'd expect Luxembourg and Amazon welcomed the decision. They both released statements to the effect that the decision proves that Amazon never received any selective tax benefit that would run afoul of EU state aid rules.
The commission released a cautious statement saying that it would study the judgment and consider its options, but it also added the commission remains committed to its overall goal of combating tax avoidance with state aid law.
David Stewart: Kiarra, let's turn to you now. Can you tell us a bit about the Engie case?
Kiarra Strocko: Yes. To simplify the seemingly complex facts of the case, the transactions involved a transfer of assets and liabilities in connection with Engie's intergroup financing structure. As you touched on before, the transactions between the group's subsidiaries were categorized as both debt and equity.
In the end, the parent entities were essentially not taxed because they benefited from the participation exemption regime. Between 2008 and 2012 Luxembourg issued tax rulings that allowed these financing structures. But the commission ended up opening a formal investigation into the rulings in 2016 after the authorities requested information from the commission.
The commission found that Luxembourg's rulings allowed the company to deduct interest owed to another Luxembourg group entity for payments that were never made and did not create any corresponding interest income. In 2018 the commission formally decided that the tax ruling by Luxembourg violated state aid rules.
David Stewart: What did the General Court decide?
Kiarra Strocko: The General Court ultimately dismissed the appeals brought by both Luxembourg and the company, and ordered the recovery of the €120 million in taxes. But what's interesting about this decision is that it seemed to be a clear-cut win for the commission.
In a handful of other CJEU or General Court decisions, the court will at least side with the other party on one aspect or another. But for the most part the court agreed with the commission's conclusions that the setup was just not justifiable.
To get more into the court's reasoning though, it backed its decision on the matter based on the nonapplication of Luxembourg's abusive law provisions. The court reason that Engie is similarly situated to other Luxembourg taxpayers that cannot benefit from the nonapplication of the antiabuse provision. Luxembourg should've applied its general antiabuse rule to reject the ruling request, but it didn't. So that alone seems to result in a clear selective advantage for Engie. They found that the insanely wacky financing structure resulted in a reduction in the corporate tax rate that would have been payable by other similarly situated taxpayers.
David Stewart: Well, it seems to be a mixed bag for the European Commission's state aid enforcement here. What are you hearing from the tax community on these decisions?
Kiarra Strocko: For Engie specifically there seemed to be an overwhelming amount of commentary about how these aggressive transfer pricing practices cannot be prevented simply by these EU state aid enforcements. However, this begs the question of what corporate tax reform measures need to be implemented to curb the prevalence of these practices.
I spoke with Raymond Luja, a tax professor at Maastricht University, who noted that the avoidance in the Engie case happened within one member state, therefore the transaction was easier to discover and address rather than other cross-border situations that have or may arise in the EU.
Regarding the general antiabuse rule bit in the decision, it seems like the case introduces a new line of attack for the commission. Luja said that the decision will likely result in the commission testing against a member states antiabuse rule standard practice when reviewing these fiscal aids in the future.
David Stewart: Ryan, what sort of things are you hearing about the Amazon case?
Ryan Finley: One theme that I picked up on from multiple observers was that the commission's loss reflects a misplaced focus on transfer pricing when maybe there are other issues at play in these cases. I spoke to J. Clark Armitage of Caplin & Drysdale Chtd., and he specifically said that one of the commission's problems here in this case and in other cases like it is that transfer pricing wasn't the issue. In this case, it was the hybrid tax treatment of that Luxembourg partnership that was licensing the IP and participating in the cost-sharing arrangement.
According to Stephen Daly of King's College in London, the commission shouldn't have focused so much on these transfer pricing technicalities in cases where the real issue and the thing that triggered the commission's suspicion in the first place for these kind of blatantly inappropriate lax ruling practices of countries like Luxembourg that were exposed during LuxLeaks. I think Daly referred to the rulings as back-of-the-cigarette-box type stuff.
David Stewart: Where do these cases go from here?
Kiarra Strocko: I think given the amount at stake in the Engie case, there's a significant chance for an appeal, which I've been hearing amongst practitioners. One hundred and twenty million euros is not a drop in the bucket. However, I will point to the fact that this case has already lasted over half a decade and an option for an appeal would simply drag it on even further.
I think it leaves open a lot of uncertainty for other multinational companies that may have similar tax arrangements. And frankly it poses the question of whether the commission will bring additional state aid cases on these arrangements moving forward.
David Stewart: Ryan, is there something else to expect out of Amazon?
Ryan Finley: No one I spoke to had any kind of clear prediction and there hasn't really been any clear indication yet as to whether the commission will appeal this decision. There are a lot of considerations at play.
At least one practitioner I spoke to, Aisling Donohue of Andersen Tax LLC in Ireland, said it would have been a surprise if the General Court had decided otherwise in this case. The odds look like they're stacked against the commission even on appeal to the Court of Justice. Whether or not that will affect the commission's assessment of whether to appeal or not, we don't know yet. It's certainly something we'll have to consider.
David Stewart: These cases are happening while the European Commission has been attempting to go after tax avoidance arrangements through the state aid rules. Are there other cases out there that have major implications along the same lines that are still pending?
Ryan Finley: There are actually two very prominent, ongoing investigations involving very well-known multinationals. One involves Nike Inc., and the other involves Ikea Systems B.V. Interestingly, they both involve almost the same kind of argument on the part of the commission that this calculation of a royalty as the residual left, after applying the TNMM, basically should have been applied in reverse. The holding company that doesn't really do very much should be the tested party instead of the operating company that does all the important functions.
In those cases as well, the commission had to rely on guidance issued after the APA at issue were granted. It would seem that the recent decision in Amazon might affect the direction they go in with those cases.
David Stewart: Well, those are definitely things we're going to have to keep an eye on, as well as the Amazon and Engie potential for appeal. Kiarra, Ryan, thank you for being here.
Kiarra Strocko: Thank you for having me, Dave.
Ryan Finley: Thanks.