Back to Taxes Committee publications

Kit Fowler
Katten Muchin Rosenman UK, London
kit.fowler@katten.co.uk

 

Report on a conference session at the ninth annual IBA Finance & Capital Markets Tax Conference

Tuesday 21 January 2020

 

Session Chair

Guglielmo Maisto  Maisto e Associati, Milan

 

Panellists

Nikolaj Bjørnholm  Bjørnholm Law, Copenhagen

Emilie Lecomte  August Debouzy, Paris

Jonathan Schwarz  Temple Tax Chambers, London

Carol Tello  Eversheds Sutherland, Washington, DC

 

Guglielmo Maisto opened the session by outlining that the panel intended to consider some recent case law in the international tax sphere and that it would address a few cases that might be of interest to all present, regardless of the jurisdiction within which the case originated. Maisto summarised the topics:

• statistical trends in tax litigation;

• the United States courts’ deference to tax regulations;

• the Canadian case of Alta Energy Luxembourg SARL v R and how it may shed light on the use of the principal purpose test; and

• the impact of Court of Justice of the European Union (CJEU) Danish cases on tax treaties.

Statistical trends in tax litigation

Emilie Lecomte explained that, on 23 October 2018, there had been an increase in the penalties associated with French taxation, including the creation of the French tax police and that the French Ministry of Finance has come to see tax fraud as a criminal issue.

However, due to a range of factors there are inconsistencies between the intended strict approach to tax fraud and actual actions. Ministers understand the need to distinguish between taxpayers who have made honest mistakes in their tax returns and those who have been deliberately fraudulent. The tax attractiveness of France as a jurisdiction is a constant consideration for policymakers. French tax authorities keep losing tax cases and tax audits in France are decreasing (in 2018 the number of tax audits was two times lower than in UK and two-and-a-half times lower than in Germany).

Tax issues concerning transfer pricing remain; audits on transfer pricing policies are a key component of the Ministry of Finance’s plan. Article 57 of the French Tax Code was used over 400 times by French authorities in 2018, resulting in assessments to tax of €3bn. In 2018, the average delay to get a prior transfer pricing agreement filed was 29 months.

Lecomte outlined the salient points arising from the French Google case (CAA Paris 25-04-2019 no 17PA03065, 9eme ch, Société Google Ireland Limited) and its impact on questions of permanent establishment. The French tax authorities considered that Google had a permanent establishment in France. It was, however, Google Ireland, not Google France, which was providing services to French customers. The French judge took a legalistic approach to that fact and found that Google did not have a permanent establishment in France. In response, the French tax authorities lodged a complaint for tax fraud against Google France and a complicity complaint against Google Ireland, firing the starting gun on criminal proceedings. Google subsequently accepted a public interest judicial agreement on a transfer pricing issue and avoided criminal conviction. A fine of €500m was agreed, to be split between the two Google entities.

Maisto asked how the case shifted from a civil issue to a criminal charge and Lecomte explained that since 2011 there has been an automatic transfer to the financial prosecutor in France if a certain financial threshold is met. The French tax authorities can keep using this method of transfer from the civil courts to the financial prosecutor to work round the ‘Google’ ruling on permanent establishment.

The panel then discussed statistical trends in other countries. Maisto commented that, in cases between US clients and the Italian tax authorities, where the Italian tax authorities have appealed to the Supreme Court, the Supreme Court finds around 70 per cent of the time in favour of the authorities. In general, the tax authorities are more likely than not to be successful in the lower courts.

Jonathan Schwarz queried the efficacy of HM Revenue and Customs (HMRC)’s success rate data but pointed out that, in terms of litigating tax avoidance cases, there is a very high success rate for HMRC. Taxpayers do, however, have better odds when contesting technical points of law.

On the US front, Carol Tello was unsure whether the Internal Revenue Service (IRS) provided similar statistics but at the highest level, the US Supreme Court has only taken on two cases: PPL (PPL Corporation v Commissioner, 133 S.Ct. 1897 (2013)) and Home Concrete (United States v Home Concrete & Supply, LLC, 132 S. Ct. 1836 (2012)).

Nikolaj Bjørnholm explained that the statistics provided in Denmark were not very good but in the upper courts, the taxpayer wins around 15 per cent of tax litigation cases. That figure rises to about 30 per cent in the courts of first instance.

US court deference to tax regulations

Maisto introduced the next topic by suggesting that a discussion of US deference may give an indication of taxpayers’ likelihood of success in litigious cases.

Tello began by sketching an outline of what deference is for US purposes. She explained that deference is a judicial doctrine, rather than black letter law: it is a presumption that an administrative agency should be deferred to when interpreting its own regulations, as it should be in the best position to interpret due to technical expertise. Firstly, Tello explained the Mayo case (Mayo Foundation v United States, 562 US 44 (2011)), where the issue was whether grad school medics were students or employees for tax purposes. The court stated that the Treasury should be deferred to and thus the medics were found to be employees (and liable to pay social security).

Meanwhile, the Chevron case (Chevron USA, Inc v Natural Resources Defense Council, Inc, 467 US 837) ('Chevron') provides the highest level of deference available. Chevron created a two-step test for deference: firstly, has Congress addressed the issue? If not, then is the Treasury interpretation reasonable? If so, then deference will apply.

Tello pointed out that the second step is not a difficult step to apply from the tax authorities’ point of view. Tello went on to explain that the US also has the Auer (Auer v Robbins, 519 US. 452 (1997)) deference, where an agency is interpreting its own regulations, as opposed to Chevron, which looked at an agency’s interpretation of statute. The Kisor (Kisor v Wilkie, No. 18-15, 588 US (2019)) case limited application of deference by requiring a five-part test to be satisfied. Finally, Skidmore (Skidmore v Swift & Co, 323 US 134 (1944)) deference applies a ‘power to persuade’ standard to sub-regulatory guidance.

Having sketched out the relevant case law, Tello asked what the corpus of cases actually meant. She raised the argument that there is a failure in separation of powers when deferring to unelected bodies and the counterargument that the agencies are ultimately accountable to the President. Two important west coast tax cases are Altera (Altera Corp v Commissioner Nos 16-70496, 16-70497 (9th Cir, 7 June  2019)) and Amazon (Amazon.com, Inc & Subsidiaries v Commissioner No 17-72922 (9th Cir, 16 August 2019)). In the latter case, Amazon lost following the court’s application of Kisor. Altera, meanwhile, applying the Chevron standard, saw the court rule originally for the taxpayer, a decision that was reversed by the ninth circuit.

A brief discussion of deference in other jurisdictions followed. In the United Kingdom, Schwarz reminded those present that foreign case law is persuasive, but that UK courts have taken a firm line against the persuasiveness of US cases decided on basis of deference and against reliance on the executive. In Italy, Maisto was clear that a tax liability cannot be created by the interpretation of a tax authority’s regulations, following a European Court of Human Rights decision. The situation is similar in France, as Lecomte explained, with judges being empowered to appraise the statute and regulations regardless of the Government’s position.

A separate question was raised by Schwarz, who asked whether, given that the Amazon case dealt with the definition of intangibles, a court should be giving such a degree of authenticity to the definition of intangible found in the transfer pricing guidelines, given the known issues with the definition. This was dismissed by Tello, who stated that the question was settled from a US perspective.

Canadian Court, Alta Energy SARL v the Queen

Discussion of the next topic began with Schwarz outlining the facts of Alta Energy Luxembourg SARL v The Queen (2018 TCC 152 (CanLII)), a recent Canadian tax case. The matter involved an investment fund in the US which owned shares in a Canadian mining company. The US/Canadian tax treaty would permit Canada to tax the gains arising on a sale of the shares. Following a restructuring, a Luxembourg company was interposed between the US fund and the Canadian company. This frustrated Canada’s attempts to tax the gains arising from the share sale, as under the US/Canadian tax treaty the Luxembourg company was outside the scope of Canadian taxation.

Unsurprisingly, the Canadian taxation authority challenged this state of affairs under the general anti-abuse rule (GAAR). The Canadian GAAR is very similar to the principal purpose test in structure. Questions arose as to the role that the preamble to the US/Canadian tax treaty plays in interpreting both the Canadian GAAR and the treaty generally. The first two elements of both the Canadian GAAR and the principal purpose test were accepted by the taxpayer. However, the third element, as to whether the restructuring frustrated the purposes of a treaty provision (ie, whether the restructuring was abusive) was contested. The judge found that the preamble was indicative of the general purpose of the treaty (to address double taxation and fiscal evasion). However, the preamble was vague regarding the application of specific articles: the generality of the preamble affected said application. In which case, Schwarz suggested that the question becomes whether the new additional wording to the preamble alters the generality of the position and subsequent analysis. Schwarz was of the opinion that the additional wording does not alter the position: in other words, the preamble was too vague. In practice, then, application of the GAAR is narrower than was first supposed.

From a US perspective, Tello pointed out that this case illustrates why the US government has not signed onto the principal purpose test. Instead, the US would prefer a limitation on the relevant benefits provision, which would address the situation.

France takes a literal approach which, according to Lecomte, is very business friendly. She explained that it was agreed in the Energy case that the restructuring had happened to take advantage of a specific tax provision. In France that would be enough to deny treaty benefits and apply penalties, as treaty shopping is in and of itself an abuse from a French point of view. From 1958, the French domestic GAAR has been applicable to tax treaties. In 2017 it was agreed that, even if appraising a double tax treaty is difficult, it cannot be possible that the intent of the treaty was to grant treaty benefits to obvious treaty shopping.

Maisto interjected, asking why it should matter if the preamble is too vague in circumstances where Articles 6 and 7 intercede, given that they are specific provisions.

In response, Schwarz suggested that Article 6(1) of the multilateral instrument purports to include a preamble in modified tax treaties – but, in Schwarz’ view, this does not insert a preamble. In general, a preamble informs the interpretation of the principal purpose test (they travel together). The real issue in the Energy case is that the principal purpose test directs one to look to the object and purpose of the article, not the treaty as a whole. Schwarz concluded that in applying treaties you must go to first principles of treaty interpretation. In response to Lecomte, Schwarz posited that a literal interpretation is insufficient. Instead, one must look to the object and purpose of the treaty.

Lecomte reiterated that, in France, if treaty shopping is identified, there is no need to look for other abuses of tax treaty. The holding period of the Canadian company would be a clue to the artificiality of the arrangement.

In summary, Maisto stated that when interpreting the principal purposes test, lawyers must look at domestic anti-abuse rules with the understanding that the principal purpose test reflects what domestic law says.

Impact of ECJ Danish cases on tax treaties

The discussion then segued into looking at Danish cases, with Bjørnholm explaining that the Energy decision is difficult to reconcile with recent European decisions.

Bjørnholm outlined that he had been examining an Italian Supreme Court case, six Danish cases which went before the CJEU (two concerning dividends and four concerning interest), a Spanish case and a Dutch case. Bjørnholm further explained that there was a common theme running through these cases: they were all lost by the taxpayer. This, Bjørnholm believed, is the reality of base erosion profiting shifting (BEPS) and lawyers should be concerned.

The Danish cases consisted of T Denmark and Y Denmark v the Danish Ministry of Taxation (joined cases C-116/16 and C/117/16), being the cases concerning dividends, and N Luxembourg 1, X Denmark A/S, C Denmark I and Z Denmark ApS v the Danish Minister of Taxation (joined cases C-115/16, C-118/16, C-119/16 and C-299/16) – the cases concerning interest. In the interest cases, the issue was one of beneficial ownership. The CJEU also discussed treaty abuse and it not clear how to separate the CJEU’s discussion of abuse from its discussion of beneficial ownership. The CJEU stated that it is for Member States to assess whether treaty abuse has taken place.

Next, Bjørnholm outlined what may actually constitute treaty abuse. There are elaborate guidelines that enable domestic courts to constitute abuse easily. Factors include looking at artificiality, purpose and timing. One interesting – and in Bjørnholm’s view worrying – point is that many structures have a US parent company acquiring a Danish company through an intermediary. There would be no withholding taking place without intermediaries, so the presence of intermediaries should not alter that position. However, the CJEU has stated that, if the first entity in the chain is abusive, then the fact that the ultimate parent is ‘good’ does not affect their finding of abuse. In terms of substance from a Danish perspective, the Danish courts do not care about premises or cost but look to the persons who decide issues. Bjørnholm fears that the CJEU endorses that approach and therefore more formal, physical substance will become irrelevant.

Maisto commented that the CJEU looks at artificial arrangement and treaty purposes in order to establish abuse, which means that if the artificiality of an arrangement cannot be established then it seems unlikely that the CJEU would rule in favour of the relevant tax authority.

The question of artificiality for Schwarz goes to circumvention – that is, whether a circumvention of the treaty purposes has taken place. On that basis, the Danish cases do not present a real departure of principle. The question remains regarding what the provision in question is actually supposed to do.

Bjørnholm demurred, stating that he would be surprised if a Danish judge reached that conclusion, instead looking at an overall assessment of the facts and circumstances. Lecomte opined that granting substance to a Luxembourg company would not be enough to obviate all risk: there is a need to balance economic interests and tax interests.

Bjørnholm touched on the aforementioned Italian case (ITW, 19 December 2018). Here, the Italian authorities prevailed before the Italian Supreme Court. The case involved simple structuring: the German company interposed between the US parent company and the Italian company had substance. It is worrying that inserting an intermediary to obtain a treaty benefit was set aside by the Italian Supreme Court.

Finally, Maisto posed a question for the panel, asking how domestic treaties would apply with regards to ultimate beneficial owners.

In France, Lecomte commented, the tax authorities and the judge only look at the interposed company. This can be frustrating: for example, if a German company is interposed between another German company and a French company, a risk of double taxation arises. However, there would be no treaty shopping issue as there would be a real business activity and the parent would be already located in Germany.

In the US, Tello highlighted that intermediaries were acceptable in some cases, but not always (eg, the Akin case, where back-to-back loans were present and the insertion of a payment company was not found to be acceptable). Anti-conduit regulations do exist in the US, and these must be examined to ensure that the structure is not caught within these regulations.

From a Danish perspective, Bjørnholm stated that he would be concerned should disregarding an intermediate entity for treaty purposes be considered abusive.

The final word went to Schwarz, who reminded the audience that any interposed company cannot benefit from the treaty between the countries in which the companies at the top and bottom of the structure reside. According to the CJEU, there is no need to identify the beneficial owner for the relief to be denied.

 

Back to Taxes Committee publications