Let’s make one thing clear at the outset – it was right for the Commission to take a look at some of the tax ruling practices carried on by some Member State tax authorities. Professor Omri Marian has written that ‘we have almost no knowledge of the legal and institutional environment that facilitated an almost completely unchecked administrative process. We do not know who instituted the process, who directed it, and why.’ In his 2017 Harvard Business Law Review article The State Administration of International Tax Avoidance, Marian took a forensic look at the Luxleaks documents and highlighted that on a single day where 11 tax ruling applications were made, 8 were approved that same day (at p. 17).

The Commission’s investigation of Amazon subsequently highlighted similar anomalies. By way of background, Amazon had received a ruling in 2003 (which was prolonged in 2011) from the Luxembourg tax authority in respect of the amount of tax due on activities in Luxembourg. At the core of the ruling was an agreement as to how to calculate a royalty payment from one Luxembourg entity to another. But this ruling was granted just 11 days after application, there was no contemporaneous transfer pricing report and there was a cosmetic presentation of the calculation for the royalty. In short, the agreement on the part of the Luxembourg tax authority does not appear to have been backed up by robust contemporaneous evidence. When the Commission handed down its final decision in 2015, finding that the ruling amounted to unlawful State aid because more tax was due than had been paid (according to its own analysis using the arm’s length principle), it must have thought: “how could we possibly lose this case?”

And yet on Wednesday 12 May the General Court found against the Commission and upheld the appeal of Luxembourg and Amazon. How could the Commission have lost this case?

The scientific answer is that the Commission failed to satisfy the evidentiary and legal burden, according to the General Court, required to demonstrate that there had been an underpayment of tax. But this is to look at the trees and not the wood.

The arguably improper administrative practices on the part of the Luxembourg tax authority and other Member State tax authorities have not been placed at the centre of the tax ruling cases. Instead, the litigation has proceeded on the claim that the tax authorities undercharged the multinationals. And the Commission has used the arm’s length standard along with the OECD guidance to support its arguments.

The outcome in the case is part of a trend whereby the General Court has responded, even in respect of those cases that the Commission has won (namely FIAT and Engie), by making it incredibly difficult for the Commission to prove that there has been an error in calculating the taxes due. The General Court is doing this deliberately because it wants to make incredibly narrow the window of opportunity for the Commission to challenge old tax arrangements. It does not want pandora’s box opened!

That this is the state of play is unfortunate, but it is an inevitable outcome of the way that the cases have been argued. There is certainly a legitimate place for State aid rules in regulating potentially dodgy relationships between large taxpayers and tax authorities – and indeed, if one reads through the Commission’s opening decisions in these cases, that appears precisely to have been its initial concern. And that initial concern was merited. To follow through on that logic then, improper and poor tax authority practices ought to be at the forefront of the State aid analysis, an argument I have made in the Law Quarterly Review in a paper entitled ‘The Power to Get it Wrong’. Once you can prove that a tax authority has acted improperly in its relationship with a large taxpayer, then a prima facie demonstration that tax has been underpaid ought to suffice. 

So where does the Commission go from here? On account of the fact that the Commission has appealed against the General Court’s judgment in Apple, one might be minded to think that the Commission will appeal this case also. But recall that the Commission also lost the Starbucks case and decided against appealing there. And a decision to litigate is not inconsequential – it costs money, money that will ultimately come from the monies generated from EU citizens, businesses and employees of EU institutions.

Most fundamentally, what exactly does the Commission have to gain from even winning this case? Luxembourg has subsequently reformed its rulings’ processes, with a rulings body established to approve rulings (rather than a single individual) and annual reports summarising, in an anonymized form, advance rulings that have been provided. Moreover, there are two interrelated reasons why the Amazon tax structure significantly reduced taxes payable – a hybrid mismatch and the previous US deferral system. The European profits sat with a Luxembourg limited partnership which was not taxable in Luxembourg (because the partners were US companies) and not taxable in the US until remitted into the country. Since the investigation began in 2014, the US has amended its rules with the TCJA and the OECD successfully spearheaded the Multilateral Instrument which among other measures made reforms to the treatment of hybrid mismatches. Sure, winning this case might result in some back taxes being paid to Luxembourg, but it seems clear that the Commission has already won in terms of these bigger points!

There is wood out there – not just trees.