Margrethe Vestager, infamously labelled “the tax lady” by former US president Donald Trump, is returning to her role as EC Competition Commissioner after a brief hiatus campaigning for the top job at the European Investment Bank. One of the many policy issues in her in-tray is what to do with the EU-wide campaign against “sweetheart" tax deals, which she launched ten years ago using the EC’s tools as supreme State aid enforcer in Europe.

Despite all the time and efforts the EC has put into its campaign, the prospects for the authority do not look bright. In the span of ten days at the beginning of December last year, the EC saw its cases against Engie and Amazon fall apart before the CJEU.

The EC’s interest in so-called “tax rulings” dates back to 2013 when details about Apple’s arrangements with the Irish tax authorities were unveiled during US Senate hearings. Such rulings are in principle lawful but risk coming into conflict with EU State aid rules if they provide for a more advantageous treatment than the normal corporate tax regime in the Member State concerned. These cases are all about how you determine the benchmark ―the normal tax burden for companies― against which to measure whether a tax deal offers an advantage.

In the Fiat Chrysler case (discussed here), the CJEU laid down the simple yet fundamentally important principle that it is the Member States themselves that fix the benchmark. This is inherent in the fiscal sovereignty Member States enjoy in the field of corporate taxation.

Engie and Amazon: further defeats for the EC before the top EU court

In its 5 December judgment in the Engie case, the CJEU (Grand Chamber) cited its findings in Fiat Chrysler about who determines the applicable benchmark. By contrast to most tax ruling cases, the Engie case was not about the use, or perceived misuse, of OECD transfer pricing guidelines. Instead, the crux of the matter was how to interpret certain provisions in Luxembourg tax law. The tax ruling in the case at hand concerned complex intra-group financing structures that were treated as equity for the parent company and as debt for the subsidiary, with the result that the bulk of the profits remained untaxed. The EC argued that this departed from the general rule in Luxembourg domestic tax law. The CJEU however was less than impressed with how the EC had interpreted the relevant legal provisions. Not only did the EC depart from a literal interpretation of the rules but also from the interpretation made by the Member State authorities. The EC had superimposed its own logic on the national tax provisions to an extent the CJEU was not willing to accept, which led to the EC’s decision being quashed. As a result, Engie escaped having to fork out €130 million in back taxes.

Nine days later, the EC suffered another blow when the CJEU upheld the GC’s annulment of the EC’s decision in the Amazon case. This outcome hardly came as a surprise. In its 2017 decision, the EC had relied heavily on the arm’s length principle in the OECD’s transfer pricing guidelines to hold that the tax ruling the company had obtained (regarding royalty payments for a license to use intangible assets) in Luxembourg triggered aid. The CJEU had already criticised using this principle as a benchmark for State aid law purposes in its 2022 ruling in the Fiat Chrysler case, which also concerned a tax deal in Luxembourg. The CJEU restated its disapproval when handing down its ruling in Amazon. The OECD guidelines are not binding (unless incorporated into national law) and the Member States are free to depart from them when applying national tax law. Amazon ducked a €250 million recovery order.

Again, the key takeaway is that the EC must refrain from imposing benchmarks to identify State aid that are foreign to the national tax system. To do so would encroach on the sovereign tax powers of the Member States.

Not all is doom and gloom for the EC: important court cases still pending

The Belgian Excess Profits case concerned a series of tax rulings in which the Belgian tax authorities accepted certain deductions to the tax base for multinational groups. The EC argued that these rulings amounted to a specific tax scheme for multinational groups which was more favourable than the normal rules in the Belgian corporate tax regime. Belgium countered that the scheme was aligned with general tax principles such as the avoidance of double taxation.

In its September judgments, the General Court (GC) sided with the EC and ruled that the scheme amounted to a misapplication of Belgian domestic tax law, which went beyond what could reasonably be justified based on the aim of avoiding double taxation. Once again, the decision of the EC and the GC relies on an interpretation of Belgian law which goes against the interpretation given by the Belgian tax authorities and legislator. So, the EC should not start counting its chickens just yet. Most of the multinationals that paid tax in Belgium relying on this scheme have appealed the GC’s ruling to the CJEU, relying on the latter’s decisions in the aforementioned cases.

Another glimmer of hope for the EC in an otherwise bleak landscape was Advocate General Pitruzzella’s 9 November opinion regarding the Apple decision; the tax ruling saga that towers above all other cases due to the staggering €13 billion recovery order imposed by the EC in 2014. This is another investigation in which the OECD transfer pricing guidelines played a key role. The GC however ruled that the EC’s reasoning was flawed and set aside the EC’s decision.

The Fiat Chrysler ruling in 2022 hinted at a victory for Apple before the CJEU. However, in a surprising turn of events, the Advocate General suggests that the case shall be referred back to the GC for a new judgment. In his view, the GC interpreted the EC’s decision incorrectly and the court’s reasoning is vitiated by a series of inconsistencies and errors. Nevertheless, it is far from a foregone conclusion that the CJEU will follow Pitruzzella’s proposal. Whatever flaws the GC’s ruling may contain, the CJEU may still take issue with the EC’s use of the OECD transfer pricing guidelines and conclude that the GC was correct to annul the EC’s decision; in other words, that the GC was right but for the wrong reasons. In fact, this is what happened in the Amazon case.

We will soon know the answer. The CJEU’s judgment in Apple is expected in early 2024.

What’s on the horizon?

Whatever the outcome of the Apple and Belgian Excess Profits cases, resorting to State aid law to rein in Member States’ use of tax rulings will be much more challenging after the Fiat Chrysler, Engie and Amazon rulings from the EU’s highest court.

However, the EC is making progress, albeit slowly, in its parallel efforts to reform EU and international tax law and policy to put an end to practices identified as “tax base erosion and profit-shifting”. This will not yield results overnight and will perhaps grab fewer headlines than multi-billion euro recovery orders against allegedly “tax-dodging” multinationals. It also requires painstaking consensus-building even beyond the EU’s borders. However, championing tax reform allows the EC to steer clear of accusations of tax harmonisation “through the backdoor” and the risk of seeing complex years-long State aid investigations collapse when appealed to the Union courts. Vestager and her fellow commissioners in Brussels need to carefully think through what route to prioritise in the future.