10 October 2019

EU General Court Rules On Starbucks And Fiat State Aid Cases

On September 24, 2019, the EU General Court (General Court) issued its long-awaited judgments in relation to the appeals brought against two European Commission (EC) decisions of ...
European Union Tax
To print this article, all you need is to be registered or login on


On September 24, 2019, the EU General Court (General Court) issued its long-awaited judgments in relation to the appeals brought against two European Commission (EC) decisions of 2015 concluding that tax rulings granted by The Netherlands and Luxembourg conferred illegal state aid on Starbucks and Fiat, respectively.1 This follows the General Court’s decision earlier this year to annul the EC’s decision that the Belgian “excess profit” ruling regime amounted to a state aid scheme. That judgment was based on the purely procedural ground that the EC should have analyzed each individual “excess profit” ruling and could not rely on a holistic analysis at the level of the scheme.2 This week’s judgments, in contrast, for the first time clarify the court’s thinking on the substantive state aid assessment of tax rulings, which has given rise to controversy in recent years.

Key Points of the Starbucks and Fiat Judgments

  • The General Court confirmed the EC’s powers to examine whether tax rulings by member states confer state aid. While member states have autonomy in direct taxation matters, national tax laws should still comply with EU law, including state aid law.
  • The EC is entitled to use the arm’s length principle as a “tool” or “benchmark” to investigate whether a tax ruling gives rise to a selective advantage under state aid rules.
  • The EC has the burden of proof that a tax ruling gives rise to an advantage, i.e., that the tax ruling resulted in a reduction of the tax burden compared to the situation absent the ruling. The EC cannot merely point at methodological deficiencies regarding the grant of the ruling, but must also demonstrate that the alleged error by the member state led to an outcome outside an arm’s length range. The EC had failed this test in its Starbucks decision, which was annulled, while the Fiat decision was upheld.
  • These judgments significantly raise the bar for the EC to prove the existence of an advantage granted to a tax ruling beneficiary. They also give further arguments to companies currently involved in appeal or investigation proceedings.
  • The judgments may still be overturned by the Court of Justice (CJEU). It may therefore take a few more years for additional legal certainty to emerge on these issues.

The 2015 EC Decisions

In its decisions, both issued in October 2015, the EC concluded that Luxembourg and The Netherlands granted a selective advantage in favor, respectively, of Starbucks and Fiat, by issuing tax rulings which artificially lowered the corporate tax that the two companies paid as compared with the liability calculated under the ordinary rules. The EC viewed the tax rulings as state aid because they allegedly endorsed an artificial methodology for the calculation of taxable profits which “did not reflect economic reality.” The EC maintained that in application of the state aid prohibition in Article 107(1) of the Treaty on the Functioning of the EU (TFEU), intragroup transactions should be remunerated as if they were agreed by independent companies operating under market conditions, in compliance with the arm’s length principle (ALP). Tax rulings inconsistent with the ALP, in the EC’s view, confer a selective advantage on integrated companies over stand-alone companies (who transact under market conditions) and may result in illegal state aid under EU law.

In Starbucks, the EC pursued a Dutch advance pricing arrangement (APA) with Starbucks Manufacturing EMEA BV (SMBV), part of the Starbucks group. The objective of that arrangement was to determine SMBV’s remuneration for its production and distribution activities within the group. Thereafter, SMBV’s remuneration served to determine annually its taxable profits on the basis of Netherlands corporate income tax. In addition, the APA endorsed the amount of the royalty paid by SMBV to Alki, another entity of the group, for the use of Starbucks’ coffee roasting intellectual property. Specifically, the APA provided that the amount of the royalty to be paid to Alki would correspond to SMBV’s residual profit. Residual profit was determined by deducting SMBV’s remuneration, under the APA, from SMBV’s total operating profit.

In Fiat, the EC took issue with a Luxembourg tax ruling in favor of an undertaking in the Fiat group that provided treasury and financing services to the group companies established in Europe (FFT). The tax ruling at issue endorsed a particular method for determining FFT’s remuneration for these services.

The EC argued that both rulings “did not reflect economic reality” and, thus, amounted to state aid. It ordered recovery in the range of €20-30 million for each company.

EC May Pursue Tax Rulings Under State Aid Rules Using Arm’s Length Principle

On appeal, the applicants heavily criticized the EC’s novel use of the state aid rules to tackle individual tax rulings. It was alleged that the EC’s enforcement actions amounted to a breach of the member states’ exclusive competence in the field of direct taxation. Also, even if the EC would be competent to pursue tax rulings under the state aid rules, the applicants strongly opposed the EC’s allegedly novel approach to “import” the ALP into the traditional analytical framework for the assessment of potential state aid.

In last week’s judgments, the General Court appears to have validated some of the core legal principles underlying the EC’s approach for the state aid assessment of tax rulings.

First, the court made clear that the EC did not encroach upon member states’ fiscal sovereignty. The court recalled that despite member states’ autonomy with respect to direct taxation, national tax laws should still comply with EU law, including state aid law. Since the EC has the power to monitor compliance with the state aid rules, it cannot be accused of exceeding its powers by examining if certain tax rulings entailed state aid.

Second, the court confirmed that the ALP may be used as a “tool” or “benchmark” to assess whether a tax ruling entails state aid. It did not consider the ALP as an integral part of EU state aid law, but rather found the principle to be inherent in the national tax laws of the member states that had granted these rulings.3 The court stressed that national corporate income tax systems which do not make a distinction between integrated undertakings and stand-alone undertakings intend to tax profits of integrated undertakings, which are involved in intragroup transactions, as though they had arisen from transactions carried out at market prices. Accordingly, the court held, Article 107(1) TFEU allows the EC to verify whether the price level of intragroup transactions, accepted by the national authorities for determining the tax base of an integrated company, corresponds to a price level of a transaction negotiated in market conditions. The court further confirmed that the transfer pricing guidelines of the Organisation for Economic Co-operation and Development (OECD), although not formally binding, constitute significant guidance, in light of the international consensus achieved in relation to transfer pricing. 

EC Must Prove Advantage to the Requisite Standard — Starbucks Wins, Fiat Loses

The court was highly critical of the EC, at least in the Starbucks case, regarding the analysis performed by the EC to establish the required “advantage,” i.e., that the ruling resulted in a reduction of the tax burden compared to the situation absent the ruling. The court pointed out that the mere noncompliance with methodological requirements is insufficient to conclude that the beneficiary of the ruling actually received an advantage. According to the court, the EC must demonstrate that the alleged error in the ruling led to an outcome that is outside an arm’s length range. We infer that this therefore allows member states a degree of latitude, given that arm’s length pricing by reference to comparables frequently produces a range, to ensure that pricing within the relevant range does not fall foul of state aid rules.

In Starbucks, the court was not convinced by any of the EC’s main or alternative lines of reasoning with respect to the alleged advantage enjoyed by Starbucks as a result of the tax ruling. It disagreed that the use of the “transactional net margin” method (TNMM) for determining SMBV’s remuneration instead of the comparable uncontrolled price method (CUP) for the Alki royalty constituted an advantage. The EC failed to prove that that choice had necessarily led to a result that was non-arm’s length, because it provided no analysis of what result would have been obtained using the CUP method. The court concluded that the EC’s finding that the royalty had not been properly analyzed in the APA was not sufficient to prove that the royalty was not in line with the ALP. According to the court, the EC failed to prove to the requisite standard that the alleged errors in the APA did not result in a reliable approximation of a market-based outcome.

In Fiat, the court did back the EC’s finding that the application of the TNMM endorsed in the tax ruling was incorrect and that it was inappropriate for Fiat to have segregated its capital by reference to its different activities for the purposes of its TNMM analysis. The EC correctly found that the whole of FFT’s capital should have been taken into account and a single rate should have been applied. The EC had also correctly considered that the method consisting, on the one hand, in using FFT’s hypothetical regulatory capital and, on the other, in excluding FFT’s shareholdings in Fiat Finance North America (FFNA) and Fiat Finance Canada (FFC) from the amount of the capital to be remunerated could not result in an arm’s length outcome. This was despite the fact that (as the court recognized) the OECD Guidelines neither expressly authorize nor prohibit such an approach. The court confirmed the EC’s view that the methodology approved by the ruling minimized FFT’s remuneration, resulting in an advantage.


The Starbucks and Fiat judgments by the General Court are expected to result in continued enforcement by the EC of the state aid rules in the area of tax rulings. The General Court has made clear that the EC is competent to investigate the state aid character of tax rulings, and that the arm’s length principle as clarified in the OECD Guidelines may be used as a “benchmark” or “tool” in that analysis, provided that the principle can be found in the national tax laws of the member state under investigation.

Competition Commissioner Margrethe Vestager announced following the judgments that the EC will continue to look at tax planning measures as part of the commission’s wider goal that all companies pay their fair share of tax through a combination of efforts to make legislative changes, enforce state aid rules and change corporate philosophies. Vestager has been nominated by commission President-Elect Ursula von der Leyen to stay on as competition commissioner in the new commission that is due to take office later this year, alongside her new role as commission first vice president for digital.

While further investigations may be expected, these judgments significantly raise the bar for the EC to prove the existence of an advantage granted to a tax ruling beneficiary and provide further arguments to companies involved in appeal or investigation proceedings. The EC cannot simply presume the existence of an advantage because the procedure for granting the ruling was allegedly arbitrary or flawed, but should prove to the requisite standard that the ruling results in an outcome that is outside an arm’s length range. As a result, the EC may now become more selective when identifying further cases.

The General Court’s judgments in Starbucks and Fiat may still be appealed to the EU’s second-tier Court of Justice, which would limit its review to points of law. It will likely take several more years until the EU’s top-tier court will have an opportunity to rule on these questions and provide further legal certainty for tax ruling beneficiaries.


1 Cases T-760/15 Netherlands v Commission and T-636/16 Starbucks and Starbucks Manufacturing Emea v Commission; T-755/15 Luxembourg v Commission and T-759/15 Fiat Chrysler Finance Europe v Commission.

2 Cases T-131/16 Belgium v Commission and T-263/16 Magnetrol International v Commission. The EC has, as a result, started investigations into a number of these companies individually.

3 The court is yet to rule on whether the ALP can be used as a “tool” or “benchmark” in relation to the application of local tax rules by a member state which does not enshrine the ALP into its domestic laws.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.

See More Popular Content From

Mondaq uses cookies on this website. By using our website you agree to our use of cookies as set out in our Privacy Policy.

Learn More