EU state aid – no letting up

In our last edition, we reported on how an increasing number of multinational enterprises (MNEs) were finding themselves on the radar of the EU Commission's State Aid investigations for having received favourable tax rulings from Member States or benefitted from a favourable state sponsored tax regime. 

Invariably State Aid was found to exist with the prospect of the tax "saved" having to be repaid. While the appeals process works its way through the European Courts, a process that will take years to resolve, we report in this edition on one of the latest State Aid investigations and what this decision means for MNEs.

If the initial focus of the EU Commission's State Aid investigations appeared to target US-based MNEs (and we look at Apple's case below), EU multinationals have also been the subject of successful investigations. ENGIE, the French energy group, has seen its Luxembourg finance structure successfully challenged, which allowed it, in the context of a purely Luxembourg financing, a tax deduction for their financing costs without a matching tax receipt.

Running through the investigations and decisions are a number of common themes:

• The Commission does not like the use of tax planning and the obtaining of rulings that create "stateless" income or income which is not taxed;
• Advance tax agreements that are essentially unilateral are more likely to be challenged than those that are bilateral;
• The State Aid requirement that there has to be a "selective advantage" granted to a taxpayer or group of taxpayers is a single not double test and so is simpler to prove; and
• The Commission is not a fan of the Transactional Net Margin Method (TNMM) for determining an arm's length price and will scrutinise such transfer pricing methodologies extremely closely.

We look below at these themes in the context of the Apple decision and seek to identify whether there are any pointers to the future direction of the State Aid cases.

Apple

The facts of the Apple case are relatively straightforward.  Two Irish incorporated but non-Irish resident corporations - ASI and AOE - obtained rulings from the Irish tax authorities as to the profits to be allocated to the trading branches each had in Ireland.  The companies were considered non-Irish resident because they were ultimately controlled by a person resident in a Treaty Country (Apple Inc. in the US) and they had a trading activity in Ireland.

ASI, the more significant of the two companies and therefore the focus of this article, owned the iphone and ipad intellectual property outside the Americas under a cost-sharing agreement but it did not have any Irish based employees involved in developing or managing the IP. It also, through its Irish branch, procured products from equipment manufacturers, and sold and distributed Apple products to European customers.  The Irish branch had a sizeable workforce supporting these activities.

The rulings granted by the Irish tax authorities to ASI in 1991 and 2007 allocated only a small amount of profit to the ASI Irish branch based on a margin on operating costs, a methodology proposed by Apple's advisers. Those profits were subject to Irish corporation tax at the 12.5 per cent rate. The remaining profits were allocated (although not expressly so) to the intellectual property which belonged to the head office of ASI that was in the US where the majority of ASI's directors were based and not the Irish branch.

Those profits would not be subject to US tax unless and until repatriated to the US since ASI did not have a taxable presence in the US.

The Commission, in its final decision, approached the issue of State Aid in the same way as it had in a number of previous decisions – Fiat, Starbucks, Amazon.

The key issue is whether there is a "Selective Advantage", which for State Aid purposes means any economic benefit which an undertaking would not have obtained under normal market conditions, i.e. in the absence of the state intervention through its ruling.

Pared down to its simplest, this requires the improvement in the financial situation of an undertaking as a result of state intervention by comparison with the position of that undertaking had the intervention not been granted. This is the "Advantage", but it also needs to be selective. Here the Commission confirmed its previous jurisprudence in Case C-270/15 Belgium-v.-Commission, that in the case of an individual aid measure (as opposed to a scheme, such as the Belgian Excess Profit scheme) identification of an economic advantage, is, in principle, sufficient to support the presumption that it is selective.

The standard three-step process by which the Commission determines the existence of a selective advantage, namely:

1. Identify the Reference System or benchmark
2. Determine whether the State Intervention is a derogation from the reference system
3. If so, is the intervention justified

can be reduced to simply the second of these two steps.  That is because in Apple's case

• comparing the economic position of ASI as a result of the ruling with its position had it not had the ruling, is the same issue as

• identifying a derogation from the reference system.

The EC jurisprudence has stated that this latter question is answered by comparing the economic position of operators who are in a comparable factual and legal situation to the beneficiary of the State Intervention, and who do not benefit from State Intervention.

In arriving at the answer as to whether ASI was in a better position than an Irish resident company carrying out the same activity and functions, the Commission applied the OECD arm's length principle to an "intra-company" profit allocation, notwithstanding that the arm's length principle was not an express part of Irish law.  That allowed the Commission to apply its favoured phrase – "Is the branch's taxable profit on which corporation tax is levied determined in a manner that reliably approximates a market-based outcome in line with the arm's length principle?"

In essence, the answer to the question of whether there is a Selective Advantage becomes a debate about transfer pricing and whether the right methodology had been chosen, the functions and analysis had been correctly identified and the right comparables adopted.

Given then that:

• no transfer pricing report had been produced at the time the rulings were requested; and

• there was an unsubstantiated assumption that the IP should not be allocated to any extent to the Irish branch of ASI

it was inevitable that the Commission would determine that the rulings conferred a selective advantage on ASI.

In arriving at that conclusion, the Commission were particularly critical of the Irish tax authorities' failure to rigorously determine the allocation of assets used, functions performed and risk assumed by the Irish branch. That factual investigation was conducted by the Commission in its lengthy investigative phase which came to the conclusion that the ASI Irish branch played a significant role in generation of ASI's profits, and that the IP should be allocated to the branch and not the head office (which had no staff and no actual physical location).

It therefore followed that the choice (and acceptance by the Irish tax authorities) of the TNMM to determine profit allocation to ASI's Irish branch, would not find favour with the Commission as approximating to a market-based outcome. That methodology is appropriate, in the context of a unilateral or one-sided ruling request, for testing the least complex party, not the party which carries out more significant functions and risks. Similarly the use of a margin on operating expenses as a profit level indicator was also an inappropriate choice for an operation that was more than just minimal risk and with significant activity. Furthermore, the Commission rejected the inappropriately low levels of return on operating expenses set out in the comparability analysis provided by Apple.

So, a finding of unlawful State Aid, was the outcome.  Although unspecified in the decision, the EC's estimate of the Aid is in the region of €13 billion.  This could, however, be adjusted as a result of restating the accounts for ASI (and AOE) correctly (to reflect better cost contributions to the IP development by the branches) and allowing for corresponding payments to other Apple group companies (where they may have undertaken greater local activity and assumed more risk).

While this is not good news for Apple or indeed the Irish Revenue, who are appealing the decision to the General Court, what does this mean for other multinationals?

The first point to note is the very thorough investigation undertaken by the Commission to establish the functional analysis and risk assumption by ASI and AOE and their branches. Clearly, where multinationals, in a one-sided ruling request disclose limited information as to the activities being carried on by both the company seeking the ruling and other companies within the group that contribute to the profitability of the group, the risk will be that any such ruling will carry little weight with the Commission if investigated. Many unilateral rulings disclose only the activities of the requesting company so that the tax authority does not see or cannot appreciate the extent of other activity within the group and, more relevantly, whether or not that activity is being taxed. Groups may also make a number of unilateral ruling requests to different tax authorities which describe only the activities of each company rather than the totality of the group activity. Often, as in the case with Apple – the other activity being the holding of valuable intellectual property – the other activity resides in a tax-free environment. While the treatment of another company in another jurisdiction that is party to transactions with the requesting company should not in principle impact on how the requesting authorities tax the transaction, the existence of "stateless income" causes the Commission a concern. On a wider basis, we have seen in the context of anti-hybrid legislation promulgated by the OECD, that the tax position of the counterparty to a transaction is a relevant factor. We expect it to be inevitable that tax authorities in giving rulings will seek to request more information about the tax impact of the group as a whole.  

Multinationals with existing rulings should be reviewing those to see how full and complete the requests were. The same is equally true of the transfer pricing analysis and methodology that supports these ruling requests. The majority of the Commission's State Aid decisions so far have involved rulings based on acceptance by the tax authorities of the TNMM.  Adoption of this particular methodology provides the Commission with three avenues of attack and they have been successfully used in a number of the decisions including the Starbucks, Fiat and Apple decisions. The Commission acknowledge that the use of the TNMM as an accepted transfer pricing methodology but only where the tested party is not making unique and valuable contributions. In other words, it is the least complex of the enterprises involved in the group transactions and does not own valuable IP or unique assets. The Commission's first challenge will always be that the actual factual situation of the tested party will be more complex than that set out in the ruling request. In the Starbucks case, for example, the Commission had doubts that Starbucks' Dutch company, which sought an advance pricing agreement in the Netherlands, was a Toll Manufacturer as claimed. Challenging the factual basis of the rulings is step one and which is why a company that does find itself being investigated by the Commission will see itself sucked into a long and rigorous process of request and disclosure of factual circumstances.  Where that investigation unearths significant facts and circumstances that were not apparent or produced at the time of the ruling, an unfavourable outcome can be expected.

The second weapon in the Commission's armoury is for the Commission to argue that the ruling adopted the wrong methodology to use. Obviously that argument is unlikely to be successful in the context of TNMM if the factual investigation conducted by the Commission concludes that the requesting company was indeed the least complex party. So far, few companies investigated have proved otherwise.

The third step is to argue that even if the methodology is correct, the application of the right profit level indicator and/or the third party comparables used to compare profit, were incorrect. This was the argument successfully run by the Commission in the Fiat case, having accepted that the TNMM was the right methodology and the return on capital employed was the right profit level indicator. They concluded that the determination of capital was incorrect and the comparable financial institutions were not sufficiently comparable.

Conclusion

The Apple decision is another clear sign that the Commission is not letting up on its attack on multinationals through the State Aid process. Both the fact that the decision has caused opprobrium in the US Treasury and the sheer magnitude of the State Aid (c.€13 billion) have not deterred the Commission. They are understood to have asked each Member State for all ruling requests based on a TNMM approach, and have increased the size of their investigating team.

Inevitably, there will be new and time-consuming investigations initiated on an individual company basis where there have been unilateral rulings and possibly also in relation to state "schemes", or state legislation, as was the case with the Belgian Excess Profit Rulings scheme and the more recent GDF Suez case in Luxembourg.  Multinationals would be well-advised to review any such ruling requests.  What might slow the Commission's fervour is a decision by the General Court in one of the appealed cases, particularly one which considers the question of "selective advantage". Unfortunately, in one of the few cases to reach both the General Court and the European Court of Justice (Autogrill Case C-20/15), the ECJ took a view of selectivity which appears to accord with the Commission's view, namely that selectivity analysis should be limited to determining whether a tax measure:

• derogates from the ordinary tax system; and
• discriminates against taxpayers which are in a comparable situation but cannot benefit from the relevant tax advantage.

The next instalment is keenly awaited!