Dr Stephen Daly

In the past few months, I have been busy pulling together my notes on the Commission’s tax ruling campaign – its litigation against Ireland, Luxembourg and the Netherlands in respect of rulings granted to Apple, Engie, Fiat and Amazon, and Starbucks, as well as the ongoing investigations into the rulings granted to Huhtamaki (by Luxembourg), and Ikea and Nike (by the Netherlands).

Three factors have motivated my thinking. The first is the prediction that Apple and Ireland will succeed in their case before the ECJ. This will be the final nail in the coffin of the Commission’s failed approach to regulating interactions between tax authorities and taxpayers. The second is that I think there should be some regulation of these interactions. The third is that the Commission will need to change tack if it is to fulfil its role as Guardian of the Treaties.

This has led to a string of outputs:

• The first is a lengthy case note in the British Tax Review on the Engie case (“Commission v Luxembourg and Engie – (another?) mortal wound in the Commission’s campaign”). This, to my mind, produces a more significant defeat for the Commission than the Amazon and Fiat decisions.
• The second is a debate article (with Werner Haslehner) in Intertax (“The Advocate General’s Opinion in Commission v. Ireland and Apple“) whereby I critique Advocate General Pitruzzella’s opinion in the Apple case and explain why I think his opinion should not be followed by the ECJ. This will be published in Issue 6/7, but the pre-publication is available already.
• The third is a 10,000word article in European Taxation (“Plus Ça Change: The Fate of the Commission’s Open Investigations”) whereby I canvass the Commission’s options, post Amazon, Engie and Fiat, for its ongoing investigations into Huhtamaki, Ikea and Nike. I suggest that the Commission must change tack but that it could be successful if it were to base its cases around internal inconsistency (Ruth Mason’s idea), administrative impropriety (my idea), or traditional selectivity analysis. This will be published in Issue 6, but the pre-publication is available already.
• The fourth is a speech that I gave at the University of Vienna in April 2024 (“The €13bn Question: Is The Fiscal State Aid Era Over?”). This gives a broad overview of the developments in the last 10 years and tries to predict whether the Apple judgment, expected some time in the next few months, will signal the end of the fiscal State aid era?

Whenever the Apple judgment is handed down, I will also produce a lengthy case note detailing its consequences for fiscal State aid regulation.

Since time immemorial it has been accepted that taxpayers with a close connection with a jurisdiction should pay tax on their income in that jurisdiction. This norm informs the taxing rights of countries – literally, their right to impose tax. This means that countries will generally have the right to tax people on their income in the place where they are based – their country of residence. But this would obviously not satisfy countries – referred to as source countries – where operations and customers might be based and hence where income is generated. Both countries will have a legitimate claim to tax the income, but allowing double taxation of this sort would hinder cross border trade. As a result, taxing rights in international tax law are allocated on the basis that source countries have the “right” to tax active income – such as trading and employment income – whilst residence countries have the right to tax passive income – such as dividends, interest and royalties. For these “rights” to have any practical effect for taxpayers, both source and residence countries must have signed a double tax treaty to reflect the allocation of taxing rights.

The injustice of the non-dom rules is that they unilaterally grant relief from tax on income (and gains) without asking whether the income (or gain) has already been subject to tax and irrespective of what any double tax treaty provisions might say. How does this work?

Let’s just work with the rules on income to illustrate the point. The non-dom rules allow somebody who is resident in the UK but is domiciled abroad to only pay tax on income sourced in the UK. Meanwhile anything generated outside the UK, so long as the money is not brought in to the UK, is not taxed in the UK. A person’s “Domicile” is generally the country where a person (or their father) is born or where a person plans on returning.

Here the distinction is made not between source and residence, but rather between source and domicile. That might be justifiable if every other country adopted the same approach – meaning that the income would be taxed in either the domicile or source country, and crucially somewhere – but of course that is not the case. Instead, countries adopt the source/residence distinction and that distinction is enshrined in double tax treaties.

Let’s take an example: if we look at the India/UK double tax treaty, we see that passive income in the form of dividends, interest and royalties will be taxed in the country of residence (save for withholding taxes at either 10% or 15%; see Articles 11, 12 and 13). This means that if an Indian domiciled person were to take up residence in the UK, then the double tax treaty dictates that India cannot tax that person on their passive income derived from Indian sources (except in respect of withholding tax). Instead, the UK has the “right” to tax that person. But the UK does not do so (so long as the money is not brought into the UK). This means the UK non-dom resident does not pay tax in the UK – owing to the non-dom rules – and does not pay tax in India (save for the withholding taxes at low rates).

That is the injustice of the non-dom rules – they unilaterally exempt from tax where there is no risk of full double taxation. This is how the rules generate a tax break for rich people with a foreign connection who have accumulated wealth in another country. Not only should the less well-off be upset by having to pay marginal rates of tax much higher than the non-doms, but even rich people who have had the misfortune of being born in the UK should feel aggrieved!

Is that tax break justified? It seems that few still think so. Many that do are harbouring a misconception that the non-dom rules somehow encourage investment. But that cannot be right, given that the rules only work where the non-dom individual does not bring their foreign sourced income or gains into the UK. So the non-dom cannot invest in the UK without having to pay lots of tax. As Arun Advani has argued, the rules actually discourage investment.

The point of this post is not to suggest that there should no rules to prevent double taxation, even on the rich. That’s what double tax treaties are for. The non-dom rules however go far further than is necessary to achieve that aim. All that can be said for the rules is that they encourage non-doms to come to the UK and spend some money consuming goods and services. This generates some taxation receipts for the State and some employment no doubt. But how much, and at what cost…

Dickson Poon School of Law, Somerset House East Wing, Strand, London WC2R 2LS

Friday 8 September 2023

The ECJ handed down a highly significant judgment today in the case of Fiat and Luxembourg v Commission concerning the application of the EU State aid rules to tax rulings. The case concerned a 2012 tax ruling provided to Fiat Finance and Trade (FFT), a Luxembourg company which provided treasury services and financing to Fiat Chrysler Group.

In short, the Court of Justice found that the Commission’s arguments were misconceived as they were not grounded in Luxembourg law. Whilst all countries have some objective basis for determining profit allocation, meaning their rules in the abstract can often be expressed as seeking “arm’s length” treatment, this does not override the need to consider the specifics of how that principle finds concrete expression in domestic law.

This blogpost details how we got to today’s result, through the ruling 10years ago to the judgment today.

The ruling

The 2012 tax ruling concerned the method for determining the profit allocation to FFT. The profit allocation was determined using a transfer pricing method known as the Transactional Net Margin Method (TNMM). This determined the profit of FFT to be EUR 2,542 million (with a range of +/- 10%), consisting only of the remuneration due to the company, as calculated by reference to the capital needed by FFT to bear its risks (FFT faced market risk, credit risk, counterparty risk and operational risk) and to perform its functions (FFT was involved for instance in market funding and liquidity investments; relations with financial market actors; and financial coordination and consultancy services to the group companies). However, no remuneration was due under the ruling in respect of FFT’s holdings in the non-European entities (FFC in Canada and FFNA in America).

The Commission decision

On 21 October 2015, the European Commission communicated its decision that the ruling amounted to State aid, as it improperly reduced FFT’s tax burden since 2012 by €20 – €30 million. Whilst the Commission found that the use of TNMM was appropriate (recital 247), it found that there were a number of aspects of the methodology used in the analysis which were flawed (recital 248):

• The Commission for instance found that the use of hypothetical regulatory capital as a profit level indicator was inappropriate as regards determining remuneration due to FFT (recital 249), with the consequent effect of significantly reducing the remuneration due to FFT (recital 256). Instead, FFT’s total capital ought to have been used (recital 249).
• Notwithstanding this conclusion, the Commission also found that the calculation of the hypothetical regulatory capital was too low, both in terms of calculating the regulatory capital that would be required by Basel II (minimum capital requirements on financial institutions) (recital 268) and the deductions that would be appropriate from that capital (recital 277).
• Further, the Commission concluded that the level of return applied to the capital to be remunerated was too low (recital 292).

The Commission concluded by finding that the entirety of FFT’s capital ought to have been taken into account for the purposes of calculating the appropriate remuneration due to FFT, to which a single rate should then be applied (recital 311).

GCEU

In a judgment handed down on 24 September 2019, the General Court rejected the appeal and upheld the Commission’s decision:

• The Court agreed with the Commission’s finding that the entirety of FFT’s capital ought to have been taken into account for the purposes of calculating the appropriate remuneration to which a single rate should then be applied (paragraph 279).
• The Court agreed with the Commission that the use of hypothetical regulatory capital was inappropriate (paragraph 279).
• The Court agreed with the Commission that the Luxembourg tax authority was wrong to conclude that no remuneration was due in respect of FFT’s holdings in the non-European entities (FFC in Canada and FFNA in America) (paragraph 279).

As with the other judgments of the General Court concerning State aid and tax rulings (such as Amazon, Apple and Starbucks), it was found that the reference framework should be the corporate tax system. It was also found that the Commission was entitled to use the arm’s length principle to determine whether there had been a deviation from the reference framework and could be assisted in doing so by OECD guidance. 

Advocate General opinion

Arguments in the case were heard in May 2021. In a punchy opinion handed down in December 2021, Advocate General Pikamäe advised the European Court of Justice to overturn the General Court’s decision and to allow the appeal. The AG concluded that only the rules and principles of the Luxembourg legal system should be analysed in assessing the existence of an advantage under Article 107 TFEU. As a result, the Commission should not, as it had done in the case, be permitted to replace Luxembourg rules with rules extraneous to the domestic system. To this end, the “arm’s length principle” was wrongly used as the benchmark for “normal taxation” in assessing whether the Luxembourg tax ruling granted to Fiat Chrysler constituted a selective advantage. The AG concluded that the Court’s erroneous identification of the reference system vitiated the entirety of the Commission’s analysis.

European Court of Justice

Today (8 November 2022), the European Court of Justice in its judgment agreed with the assessment of AG Pikamäe and found that the decisions of the General Court and the Commission should be annulled. The Court stressed that the determination of whether State aid has arisen will depend upon a comparison between the scheme for “normal taxation” and that which applied to the taxpayer in question. To undertake such a comparison, it is the domestic tax rules which should be considered. Rather than using an abstract version of the “arm’s length principle” for this comparison, it is hence the manner in which the arm’s length principle is expressed in the domestic rules which should be considered (which was to be found in Article 164(3) of the Luxembourg Tax Code and specified in the related Circular No 164/2). The Court did not stop there. It also highlighted that resorting to non-domestic rules, as the Commission had done, was contrary to the TFEU rules on the approximation of Member State law – i.e. direct tax rules can only be harmonised in EU law where there is unanimous consent (according to articles 114(2) and 115 TFEU). The approach of the Commission and General Court failed to pay due respect to the Treaties insofar as the “autonomy of a Member State in the field of direct taxation… [is] fully ensured” (paragraph 94) except for specific EU tax rules unanimously adopted. The subsidiary line of reasoning also failed as it was predicated on the principal line of reasoning (despite the superficial references to the relevant provisions of Luxembourg law).

Ramifications

It cannot be overstated how significant this judgment is for the other State aid tax ruling cases, such as Apple, in that the court applied a more rigid interpretation to the scope of the domestic rules than the General Court and the Commission. The ECJ held that the concrete terms of the arm’s length principle are to be derived from national law (see paragraph 95). In the case of Apple, there was no “arm’s length principle” in place when the 1991 tax ruling was granted. This finding then must augur well for Ireland and Apple as the ruling will not be required to align with an autonomous arm’s length principle. Instead, it will merely be the rules in Ireland which existed at the time which sought to ensure that there was an objective basis for determining the profits on a non-resident trading company (see Dataproducts and Belville Holdings cases). A successful finding for the Commission in the Amazon case at the Court of Justice is very unlikely meanwhile given the Commission’s apparent misconception of the scope of Luxembourg law.

Although the judgment is highly critical of the Commission, the Court did however offer a fig-leaf at paragraphs 119-122. State aid can arise from the provision of a tax ruling, such as where the domestic rules are “manifestly inconsistent with the objective of non-discriminatory taxation of all resident companies…pursed by the national tax system, by systematically leading to an undervaluation of the transfer prices applicable to integrated companies or to certain of them, such as finance companies, as compared to market prices for comparable transactions carried out by non-integrated companies” (paragraph 119).

My position (set out more fully in a 2021 LQR article and a 2022 MLR piece) meanwhile has always been that there was another way to tackle these cases – an approach which focused on the key issue at the heart of the original investigations which was a suspicion of administrative impropriety.

Whether the Commission has the stomach for further litigating these issues is another matter…

Where a tax authority fails to collect taxes due, who holds them to account? The answer, at a domestic level, will depend on the jurisdiction but may be the courts, Parliament, the media, specialist committees and so on.

But there is a case for supranational monitoring of tax collection, particularly in the case of the EU, by virtue of the negative spillover effects caused by lax tax administration, as well as the budgetary and macroeconomic consequences.

In a forthcoming paper in the Modern Law Review, I suggest that Member States owe an account of the performance of their tax authorities to the Commission. In order for this to be operationalised, I suggest that we should use an accountability framework which harnesses expertise, rather than being antitethical to it –  Onora O’Neill’s model of ‘intelligent accountability’. Using tax rulings as a proxy for determining how tax authorities act in relation to large taxpayers, the paper proposes that tax rulings provided to large taxpayers should be published (in an anonymised format) and assessed by the expert Code of Conduct Group (which monitors Member State compliance with the Code of Conduct for Business Taxation), Member States through peer-review and the public. Appreciating that Member States will have their own mechanisms for holding tax authorities to account, the paper goes on to propose separately that these systems can be harnessed so that an account of the performance of tax authorities can be transmitted to the Commission, which (with the assistance of an expert ‘Tax Administration Group’ – to be established) can then decide whether any changes should be recommended, which would be monitored in the European Semester.

Both the Code of Conduct for Business Taxation and the European Semester are examples of the ‘open method of coordination’ – a framework for cooperation between EU Member States. So the paper then brings together the OMC, a theory of accountability, and makes a case for monitoring, and suggests how to monitor, tax authorities at an EU level.

The article is open access and available here.

The recent Supreme Court case of Tinkler v HMRC [2021] UKSC 39 dealt with the consequences of an administrative error. A notice of enquiry was not sent to the correct address. This is problematic because the statutory provision requires that a taxpayer is given notice of an enquiry.

But nobody seemed to notice this mistake initially – the taxpayer’s agents had been made aware of the enquiry and dealt with HMRC accordingly.

It was only raised by the taxpayer as an issue about a decade after the enquiry was opened.

HMRC sought to rely upon the doctrine of “estoppel by convention” to prevent this mistake from invalidating the enquiry (and there was approximately £635,000 at stake, so this would not have been inconsequential). The Supreme Court ultimately found in favour of HMRC.

Whilst one can have sympathy for both sides to this dispute, there is one lingering problem that I have with the judgment. This is that HMRC actually realised the mistake was made in late 2005. Corrective action could have been taken then (i.e. a new notice of enquiry could have been sent to the correct address of the taxpayer), but it was not. Quite why this didn’t operate to prevent HMRC from raising the estoppel argument I do not know.

A fuller explanation of the case and my thoughts on the Supreme Court judgment have now been published in the British Tax Review.

The abstract for the case note reads as follows:

“Mistakes happen. For all the heat that can be generated by legal arguments, at the heart of the recent Supreme Court case of Tinkler v Revenue and Customs is the question of what approach the law should take where an official within HMRC has made a simple mistake, which went uncorrected – a notice of enquiry was sent to the wrong address.”

It can be downloaded free from SSRN.

In recent years, the UK has been tasked with implementing various tax rules which have been negotiated and agreed at the international level, albeit in very different circumstances, namely the EU Anti-Tax Avoidance Directive (ATAD) and the OECD led Multilateral Instrument (MLI).

Despite the freedom to design tax rules brought about through the UK leaving the European Union, the UK nevertheless has faithfully implemented the rules in ATAD, a point which is actually not surprising when one looks at the role the UK played in the BEPS project and the choices subsequently taken when it came to implementing of MLI into UK domestic law.

My own analysis of the UK’s implementation of these two initiatives can be found in two recently published pieces: an article in Intertax (‘The UK’s implementation of ATAD’ (2021) 49(11) Intertax 938) and a chapter in an IBFD edited collection The Implementation and Lasting Effects of the Multilateral Instrument.

If you would like a copy of either, please email me at stephen.daly@kcl.ac.uk.

One of the issues often raised with plans to introduce is wealth tax is the problem of valuation. Aside from political issues around the tax base, and assuming that value should be determined by reference to market value, how does one go about valuing assets which are not often sold on the open market and thus for which there might be little information from which to estimate value: intellectual property and shares in small private businesses for instance? How many people would, in reality, be confronted with difficult valuation issues and how much would it cost them to hire valuation experts?

In a paper recently published open access in Fiscal Studies, Glen Loutzenhiser (Oxford), Helen Hughson (LSE) and I consider the scale and prevalence of valuation issues under a wealth tax. We examine some of the most problematic asset types from a valuation perspective. We also consider a range of solutions to manage these concerns, drawing on international experience and the approaches already taken for other taxes within the UK system. We conclude that satisfactory options for arriving at a value for wealth tax purposes are available even for the most problematic assets. We also estimate that the absolute number of taxpayers likely to pay substantial valuation fees is small, and that, in aggregate, valuation costs could be contained to around 0.1 per cent or less of total chargeable assets, even if they are substantial for some individual taxpayers.