Back in 2016, the U.S. Department of Justice announced it was hitting Deutsche Bank with a $14 billion fine relating to its dealings in mortgage-backed securities. The irregularities occurred years earlier, during the buildup to the 2008 financial crisis.
Around the same time, the European Commission announced it was hitting Apple (indirectly) with a tax assessment of €13 billion, delivered under the guise of a state aid proceeding directed against Ireland.
Given the prevailing exchange rates between dollars and euros, there wasn’t much difference in the amount of these colossal sanctions — as if to offset each other when combined on some imaginary ledger of transatlantic grievances.
I’m not given to conspiracy theories, but I’ve often wondered if the events concerning Deutsche Bank and Apple were connected, despite outward denials. The timing and the amounts are simply too convenient.
The diplomatic subtext makes sense: If Washington dares to mess with one of Europe’s national champions, Brussels will find a way to return the favor in a way that’s proportional. Tit-for-tat gamesmanship is not limited to bilateral trade.
If you were an EU official seeking some way to retaliate against an icon of U.S. industry, Apple’s international tax planning must have seemed like an easy target.
The world already knew from the OECD’s base erosion and profit-shifting project that multinational enterprises were enjoying huge profits and low effective tax rates. BEPS action 1 implicated the digital economy as an unsolved problem for which no practical solution was readily available. This was a few years before BEPS 2.0.
All the commission needed was a legal basis for portraying the enablers of tax avoidance — those pesky advance rulings — as a form of prohibited state aid.
Never mind that direct taxation is acknowledged to be an area of national competence. Article 107 of the Treaty on the Functioning of the European Union makes competition policy everyone’s problem.
This article projects how that saga will end. The Court of Justice of the European Union’s decision in Fiat Chrysler Finance Europe v. Commission, joined cases C-885/19 P and C-898/19 P (CJEU 2022), issued in early November, sets things right over the interplay between state aid doctrine and transfer pricing.
Fiat itself isn’t so important, but it makes the CJEU’s forthcoming decision in Apple a foregone conclusion. It has taken several years, but the grand state aid misadventure seems to have run its course.
The Reference Base
There was always something odd about how Denmark’s Margrethe Vestager, then serving as EU competition commissioner, emerged as Europe’s leading tax cop. As a technical discipline, taxation was never part of her regulatory portfolio.
The European Union had a separate tax commissioner — yet he was out of the picture, powerless to lead the charge against member states’ indulgent habits.
Another odd thing was that so many experienced tax professionals couldn’t understand how Ireland could rightfully tax Apple’s profits to the degree the commission was calling for.
While discussing Apple’s state aid dispute with reporters in 2016, the OECD’s Pascal Saint-Amans gave voice to what many others had been thinking: In transfer pricing terms, “the bulk of the profit clearly belongs to the United States.”
In effect, the commission was trying to unilaterally soak up profits that the United States had declined to tax itself. That would parallel the OECD’s undertaxed profit rule under pillar 2, except that here the relevant member state (Ireland) wants nothing to do with soaking it up. Seen through the lens of our use-it-or-lose-it metaphor, two nations displayed forbearance.
The overreach seemed clear to non-Europeans and was well documented in these pages. The commission was unable to impose its preferred version of the arm’s-length standard on taxpayers by direct EU legislation.
Rather than concede that such matters were beyond its reach, it chased a backdoor approach to tax harmonization using the powers afforded under TFEU article 107. In lieu of a formalized UTPR (formerly known as the undertaxed profits rule), it relied on the arm’s-length principle.
The scheme almost worked. The General Court of the European Union (GCEU) went along with the commission’s novel reasoning on several occasions, including the decisions in Fiat (Luxembourg v. Commission, joined cases T-755/15 and T-759/15 (GCEU 2019)) and Starbucks (Netherlands v. Commission, joined cases T-760/15 and T-636/16 (GCEU 2019)), on the appropriate reference base for determining selective advantage.
The court’s September 2019 decisions in those cases drew heavily on an earlier CJEU decision (Belgium and Forum 187 v. Commission, joined cases C-182/03 and C-217/03 (CJEU 2006)), which authorized the commission’s use of the arm’s-length standard in some circumstances.
Similarly, in Amazon (Luxembourg v. European Commission and Amazon EU Sàrl v. European Commission, joined cases T-816/17 and T-318/18 (GCEU 2021)) the GCEU allowed the commission to rely on the OECD’s arm’s-length standard to benchmark a member state’s use of the same principle. The court confirmed the basic approach of Fiat and Starbucks, though it ultimately found that the commission failed to meet the burden of proof, resulting in a victory for the taxpayer.
In hindsight, the court’s reliance on Forum 187 was misplaced. The CJEU never specified whether the commission’s use of the arm’s-length standard was justified because the member state in question (Belgium) had incorporated the concept as part of its domestic tax code, or because the standard was generally applicable any time the commission sought to apply it.
It all began in 2012 when Luxembourg granted an advance ruling to Fiat Finance and Trade Ltd. (FFT). FFT was later reorganized as Fiat Chrysler Finance Europe, which is the name referenced in the commission decision and ensuing court proceedings.
FFT handled intragroup financing for operations in North America and portions of Europe, excluding Fiat’s home country of Italy. The ruling covered the tax years 2012 through 2016.
The ruling allocated profits to FFT based on a combination of risk renumeration and functions renumeration. The risk factor was based on a hypothetical measure of regulatory capital (€28.5 million) and a pretax return of 6.05%.
The estimate of risk capital relied on the Basel II framework. Significantly, the estimate excluded equity relating to FFT’s stakes in two of the Fiat group’s other finance affiliates: Fiat Finance North America Inc. and Fiat Finance Canada Ltd.
The explanation for the exclusion isn’t entirely clear, though it followed from a transfer pricing report prepared by KPMG that supported FFT’s ruling request. The functions renumeration was based on a separate estimate of the capital required to perform FFT’s treasury services (€93.7 million) and a presumed market interest rate of 0.87%.
The trick, so to speak, was that the agreed-upon measure of regulatory capital for risk evaluation understated the return on equity, producing a reduced allocation of profits.
Strictly speaking, the commission’s complaint is not about whether Fiat was getting away with clever tax planning — which was certainly the case — but the appropriate benchmark for determining a selective advantage. However favorable the ruling was, it would not be considered a selective advantage if such terms were available to all comers.
The advance ruling was issued in accordance with article 164(3) of Luxembourg’s income tax code and a circular that authorized the national revenue authorities to issue advance tax rulings. The article reflects Luxembourg’s codification of the arm’s-length standard, and the circular provides an explanation of how arm’s-length prices are to be determined. The ruling contained a recital that its terms respected arm’s-length principles.
At no point did the commission try to determine whether the favorable tax ruling ran afoul of Luxembourg’s domestic transfer pricing regime. It could have done that, but purposefully chose not to go there. Doing so would have signaled that domestic law could dictate the reference base for identifying any derogation from normal outcomes.
The commission argued that scrutiny of local practices was an unnecessary distraction; the critical elements of selectivity must be gauged only by applying its own version of the arm’s-length standard. Ireland intervened in the litigation, given what it had at stake in the Apple appeal.
The justification for all of this was that Luxembourg’s corporate income tax had the ostensible purpose of taxing all profits. The longer you dwell on it, the more you realize that reasoning is flimsy. Don’t all corporate tax systems have the stated purpose of taxing profits — apart from the odd gross receipts tax that can be ignored in this case?
By extension, there’s no tax regime across the whole of Europe that wasn’t vulnerable to being superseded — because each of those systems seeks to tax an expansive definition of gross income that nominally applies equally to group members and stand-alone companies. The implication is that the commission can freely disregard a member state’s transfer pricing outcomes that are not to its liking — which is to assign competition policy a remarkable remit. That can’t be the correct result given what we know about direct taxation being the exclusive competence of national governments.
For me, the disingenuous part of the commission’s case is how it fixates on the idea that Luxembourg’s objective was to tax all profits. The commission takes that to mean an intention to tax all resident entities on equivalent terms.
Formally, Luxembourg never contested the point that its tax system seeks to tax all profits, including those of group members and stand-alone companies. However, the professed linkage between the state’s objective and the commission’s entitlement to override national law has never been substantiated.
The apparent rationale is that by possessing the objective to “tax all profit,” the state is deemed to embrace a conceptual market orientation, and strict adherence to the OECD’s arm’s-length principle is offered as the only means by which that condition can be satisfied.
How such a conclusion naturally follows from the initial assumption of legislative intent is something of a mystery. In context, it’s a power grab that gives the commission oversight over all transfer pricing arrangements to which any EU member state is a party. How is that any different, in substance, from a general ability to regulate taxation of an integrated company?
Not only does that clash with the TFEU on the exclusive competence concept, but it separately violates the principle of legality because private actors aren’t in a position to reasonably foresee tax liabilities as they come due.
The Commission’s Last Stand
It took a while, but reason prevailed in the form of Advocate General Priit Pikamäe’s opinion, issued in December 2021. The opinion invited the CJEU to annul the GCEU’s decision, finding that the commission overstepped its bounds and infringed on member state sovereignty.
Pikamäe framed the issue in straightforward terms. The commission’s selectivity analysis for state aid purposes must be confined to national law:
“In order to avoid any encroachment on the exclusive competence of the member states in the area of direct taxation, the existence of an advantage within the meaning of article 107 [TFEU] can be verified only by reference to the normative framework outlined by the national legislature in the actual exercise of that competence.”
In short, the commission was not entitled to grab at a hypothetical interpretation of the arm’s-length principle. Pikamäe observed there was precedent for the rejection of a “fictitious” reference base, citing the CJEU decision in Commission v. Poland, C-562/19 P (CJEU 2021).
The CJEU decision, released November 8, mirrors the advocate general’s opinion in every important respect. It held that the GCEU erred as a matter of law in allowing the commission to determine the reference base through external elements.
Rather than find that Luxembourg had violated an EU state aid law, the lower court had infringed on Luxembourg’s exclusive competence over corporate taxation, as follows:
“By disregarding national tax rules, the General Court infringed the provisions of . . . the Treaty relating to the adoption by the European Union of measures for the approximation of Member State legislation relating to direct taxation, in particular Article 114(2) TFEU and Article 115 TFEU. The autonomy of a Member State in the field of direct taxation, as recognized by the settled case-law . . . cannot be fully ensured if, in the absence of any such approximation measure, the examination carried out under Article 107(1) TFEU is not based exclusively on the normal tax rules laid down by the legislature of the Member State concerned.”
Is Forum 187 still good law? That’s not entirely clear. You can’t read the CJEU’s decision in Fiat as anything but a rejection of the idea that the commission may use an external version of the arm’s-length principle to check the legitimacy of a member state’s transfer pricing regime. At paragraph 102 the Court writes:
“Contrary to what the General Court held . . . the judgment of 22 June 2006, Belgium and Forum 187 v. Commissioner . . . does not support the position that the arm’s-length principle is applicable where national tax law is intended to tax integrated companies and stand-alone companies in the same way, irrespective of whether, and in what way, that principle has been incorporated into that law.”
It’s now settled that Forum 187 doesn’t mean what the commission thought it meant. The CJEU stopped short of declaring the earlier decision null and void, though at this point it’s not clear what the case stands for.
The implications are probably fatal for the commission’s appeal in Apple. There, the GCEU agreed with the commission that an external version of the arm’s-length standard could be imported into the state aid analysis.
Ireland prevailed because the commission was unable to establish that a selective advantage was conveyed. Before the ruling in Fiat, many observers had been expecting the CJEU to merely endorse the GCEU’s holding and end the matter in the taxpayer’s favor.
Given this latest development, however, I would expect the CJEU to hand the commission a more resounding defeat that finds an error of law in the lower court’s reasoning.
Going forward, the commission remains free to bring other state aid charges against member states, but it will be foreclosed from using outside elements beyond national law to determine the reference base — which is everything in these cases.
With the commission having definitively lost on that issue, you must wonder if there’s still any purpose in its bringing tax-motivated infringement cases under article 107 — unless the member state’s conduct is wildly divergent from its own transfer pricing regime.
The commission has lost the battle over what constitutes normal taxation for an integrated company. With that, it’s presumptively lost the war.