This article was first published on Concurrences and is reproduced on this blog with the authorization of its authors.
Allegedly, tax competition has gone too far. For the past ten years, the international community has tried to find ways to reduce its intensity, both at the international level, with the various OECD initiatives (BEPS base erosion and profit shifting) and then the two pillars’ solutions), and at the European level through the vote of directives, such as the Anti-Tax Avoidance Directive (ATAD), and the use of the State aid control mechanism.
This article focuses on the latter phenomenon, specifically the challenge of a specific type of tax ruling: the advance pricing agreement (APA) as illegal State aid. In a nutshell, it will argue that APAs do not have any significant effect on the efficiency of the European economy but only raise equity issues. As such, they should be left out of the scope of Article 107 of the Treaty on the Functioning of the European Union (TFEU).
We will start by briefly describing the tax background, which may be somewhat unfamiliar to competition law specialists: What is tax competition exactly (I.), and what is an APA (II.)? A brief analysis of the economic consequences of an APA will be presented, along with a discussion on the extent to which it can be considered a State aid (III.). We will finish with some elements of conjecture for the future when Pillar 2 is fully enacted in Europe (IV.).
International tax competition: A long story short
The historical view
The story has been told elsewhere in more detail1: the second wave of globalization of the economy has seen the fragmentation of the production process, the development of the international value chain and the rise of gigantic multinational enterprises (MNEs).
As a side effect of this structural change in the economy, the balance of power between states and taxpayers has been turned upside down. Back when the economies were mostly closed, the state was in a sort of monopolistic situation, where taxpayers had almost no other way than to pay whatever taxes were enacted. Fast forward to the 2000s, large MNEs are so big that they can have states compete to attract them in their jurisdiction. In a framework where all countries do not have the same locational advantage (low wages, good education, cheap energy, etc.), tax policy proved an efficient tool to attract both productive capital and taxable profit. A “tax war” therefore started and got more intense in the 1990s, triggering a trend change in the tax mix of many rich countries, with a reduction in capital taxation and a corresponding increase in taxation on labor and consumption2. This tax war proved hugely successful for the winners, like Ireland or Luxembourg, which significantly reduced their corporate income tax (CIT) rate and saw impressive growth in their GDP per capita over the period—not so much for other European states.
In the aftermath of the 2008 crisis, the low tax rate paid by certain large multinationals started to make the headlines. Rich states that were struggling to balance their public budget started to realize that they had lost a lot with the tax war and that it might be time for them to reconsider the international tax rules. They then turned to the OECD to design a structural tax reform: the BEPS project (2015). BEPS closed certain loopholes but did not solve the issue that was at the core of the matter: how to stop competition to attract investment and profit. Shortly after, the G20 asked the OECD for something more ambitious, which led to the two-pillar approach that is in the process of implementation3.
European history runs parallel to this one4. Since the early 1960s, the Commission has tried to reduce the internal frictions caused by inconsistent tax systems. Such problems were identified very early; however, since direct taxation is mostly the responsibility of Member States, the Commission had very few possibilities to correct them, except by issuing a non-binding “code of conduct.”5 After the 1992 Ruding report6 that recommended harmonizing the corporate income tax rates (which never took place), the 1998 code of conduct7 focused on certain national tax regimes that were considered abusive and incompatible with the common market. In order to go beyond the simple recommendation, the Commission mentioned that abusive regimes could be considered illegal State aid, which gave it some bargaining power to have the states remove them from their Tax Code (e.g., the Belgian headquarters regime). Schematically, before 1998, State aids were mostly concerned by protectionism; after 1998, they started to focus on general tax regimes that were intended to divert capital flows. The last piece of the puzzle took place in the early 2010s, after the topic of tax optimization of MNEs attracted media (and political) attention, when the Commission started to investigate individual rulings, particularly APAs.8
The analytical view
The expression “tax competition” is widely used in the debate on European tax policy, but it is sometimes misleading as it can refer to several different usages of tax by Member States to serve a strategic purpose9. Three of them need to be clearly distinguished:
- A state can give a tax break to a specific firm (or a group of firms in an industry) as a subsidy in disguise, so that they perform better against competition. The Italian textile case10, where Italy allowed textile companies to pay less social security contributions, would be an example.
- A state can implement tax policies to attract productive capital from MNEs (R&D center, plant, etc.) in their jurisdiction. The Indian states “bidding war” to attract Tata Motors’ plant would be a canonical example11.
- Lastly, a state can implement policies to attract the taxable base (i.e., profit) of MNEs. The most obvious way of doing that is by reducing the local tax rate.
It is important to note that, maybe surprisingly, items 2 and 3 are not necessarily the same thing. Under the current international tax system (based on the arm’s length principle), it is possible, and sometimes even mandatory, to largely dissociate the location where a firm has physical capital and the location where its profit is allocated. Profit generally goes where key decision power and key intellectual property (IP) are located,12 rather than with factories and warehouses. An MNE can therefore decide to build a large factory in France because this is what makes most economic sense and have its IP and headquarters with key decision-makers in Ireland, so that most of the profit derived from the factory’s activity would be allocated in Ireland, which would be perfectly in line with the arm’s length principle.
Taking this into account, states would not necessarily use the same tools to attract physical capital and to attract profit. Decreasing the CIT rate would attract mobile intangible assets, people, and profit13 but not necessarily physical capital. Attracting a large plant generally rather entails local tax exemption, energy subsidies, etc.
What is an APA?
One of the main objectives of the International Tax Regime is to allocate the profit of firms (circa $8 to 10 trillion per year14) between the states where they operate. This is done through the regulation of the prices at which transactions between subsidiaries take place—the famous “transfer prices.” If subsidiary A sells an input to subsidiary B at a high price, this will shift profit to A, and vice versa if the price of the input is low. In OECD countries, the transfer price regulation is based on the arm’s length standard according to which the “right” transfer price for tax purpose is the theoretical “market price” of the good or service transferred.
This regulation is based on outdated economics and has many flaws15. In practice, goods that are exchanged within an MNE are almost never sold on a market (which is why MNEs exist in the first place). Assessing the theoretical price at which a theoretical buyer would purchase these goods therefore involves a certain level of subjectivity. The uncertainty on the right price entails significant financial risks for MNEs because if states A and B disagree on how to calculate the transfer prices of the transaction between their subsidiaries, the same profit can be taxed multiple times16.
To avoid the risk of multiple taxation on certain large and complicated transactions, MNEs can ask a pre-approval to tax administration on their transfer price. This type of ruling is called “advance pricing agreement” or APA. APAs are generally a very efficient and useful tool to give tax certainty to MNEs. Because of the complexity and duration of an APA negotiation, it is, however, a rather uncommon practice in Europe17.
Of course, APA can sometimes also be used to obtain security for a scheme which, although legal, can lead to an allocation of profit that hurts an intuitive notion of tax justice. Let’s take the example of a tech company that has a valuable IP that it uses to manufacture goods in its Irish plant and suppose that the correct allocation of profit, according to the arm’s length principle, would be 90 with the IP owner, and 10 in the Irish plant (most of the value is in the IP). If the IP owner were in the US, the license transaction between the IP owner and the plant would create no tax risk, as it is reasonably standard. If instead of the US, the IP owner is located in a tax haven, in an entity with limited economic substance, even if the profit allocation is the same (90 with the IP owner and 10 in the plant), the transaction becomes more questionable, and the Irish tax administration could consider that the whole scheme is an abuse of law. To prevent that risk, an APA can be concluded with Ireland.
This example illustrates the kind of APA that has been scrutinized by the European Commission in the past few years. They are characterized by the following factors:
- They benefit the MNE by allocating part of its profit to a low-tax jurisdiction, with a high level of security;
- The state that gives the APA does not necessarily lose anything with the APA—in the above-mentioned example, the “losing state” would be the US, not Ireland, since Ireland is getting exactly its fair share of the group profit.
Should APA be in the scope of State aid regulations
What is the economic effect of an APA?
We will focus here on APAs securing the allocation of profit in an entity with limited substance and a low tax rate18.
We can try to give some qualitative elements to assess the economic impact of these APAs following the three categories identified above: Do they improve the competitive position of the beneficiary? Do they create distortions in the geographical allocation of capital? And lastly, do they create distortions in the geographical allocation of profit?
- With regard to inter-company competition, it is important to note that a lower CIT rate is not at all the same as a subsidy. Let’s take the example of a firm with a production cost of 10 operating in a market where the price is 8. Absent state intervention, that firm would get bankrupt. If the state grants that firm the right not to pay part of its social security contributions, reducing costs by 3, then the firm’s operating cost becomes 7, and it can beat its competitors on price. If instead, the state gives that company a more favorable CIT rate, it would not make its cost structure more competitive, since you only pay CIT when you generate profit, meaning when you can set a price higher than your cost. For the same reason, benefiting from a better CIT rate would not allow a firm to engage in predatory pricing behavior, as in the Deutsche Post case19. Generally speaking, firms are expected to make pricing decisions that maximize their profit, yet since CIT is a percentage of profit, maximizing profit before and after tax is the same thing, and there is no reason that the CIT rate would change the equilibrium price on a given market20. Everything else being equal, a lower CIT rate would allow the firm to obtain more cash flows with a given level of profit, serve better dividends or have more funds to invest and buy other assets. The effects of that on competition are indirect and unclear, particularly in a context where financial markets are reasonably efficient21.
- It is also unlikely that an APA would have a significant impact on the allocation of physical capital within Europe. Indeed, it is not necessary to allocate significant physical capital to the entity that signs the APA. As a matter of fact, looking at APAs that have been scrutinized by the Commission, they generally did not involve entities with a lot of assets. Additionally, a lot of these APAs were signed with entities that were already in the country before the agreement was signed; no further investment was needed. For a state, communicating on the fact that it is giving “sweetheart deals” to certain firms does not seem to be a good strategy to attract capital22.
- Lastly, it is true that APAs have an impact on geographical profit allocation—it is indeed the whole point of their existence. It should, however, be noted that the allocation of profit supported by the APAs under scrutiny by the Commission is generally not to the detriment of Europe. Take again the example of a US MNE that has operations all over Europe, a headquarters in Ireland and an IP owner in a low-tax jurisdiction, where a lot of the profit is allocated as per the APA. The equity (not legal) issue here is related to the amount of profit allocated in a small Pacific Island. But that profit did not belong to Europe in the first place; rather, probably, to the US, where the IP had been developed. As a matter of fact, if any European country (except Ireland) considered they did not get their fair share of profit, they would be free to perform a tax audit in their jurisdiction, since they are not bound by the APA.
To summarize: APAs are about allocating corporate profit. The vast majority of them are fine, and—subject to further analysis—those that are debatable from an equity standpoint do not seem to have consequences in terms of trade, competition and even allocation of physical capital.
What problem do APAs raise, and can it be solved by State aid control?
APAs can contribute to securing the allocation of MNEs’ profit in low-tax jurisdictions, resulting in a lower CIT rate for them, which, even when legal, can be debated from an international equity standpoint. It is precisely that equity argument that has been used by Commission officials to justify the wave of tax ruling scrutiny that started in the mid-2010s. Margrethe Vestager: “We are in a situation today where some companies are engaging in aggressive tax planning, made possible by lack of fiscal harmonization in the EU and loopholes in national taxation systems (. . .) Especially in challenging economic times, with many EU citizens having had to tighten their belts, it is even more important that large companies pay their fair share of tax23.” Pierre Moscovici: “It is time to guarantee equity and transparency in the taxation of firms in Europe24.” The list goes on.
The question is: Has State aid control been designed to serve that kind of purpose? In an enlightening article, Mason25, after a thorough analysis, argues that the main objective of Article 107 TFEU is fair trading, preventing distortion in inter-firm competition. Even if it has also been used to reduce the intensity of tax competition between states, such as in the famous Gibraltar case26, it should not be the main purpose of State aid control.
Article 107 TFEU prohibits State aid which distorts or threatens to distort competition in so far as it affects trade between Member States. This wording clearly refers to an efficiency objective, but what exactly is a distortion of trade? As noted by the OECD: “[T]he very meaning of these notions of competition distortion and affectation of trade has evolved over time as well. In Philip Morris, the [Court of First Instance] held that competition distortion relates to a change in the position of an undertaking compared with other undertakings in intra-Community trade. But in the Framework for Research, Development and Innovation (FRDI) the European Commission also mentions changes in the location of economic activity as a possible distortion27.” Looking at this from an economic standpoint, it is clear that state measures that affect private competition should be in the scope of Article 107 TFEU, since they create clear distortion in the functioning of competitive markets. Including measures that only have an impact on the allocation of investments is already a stretch, since the effects of such measures on welfare are debatable28. It seems, however, hard to go beyond that point, and it is hard to find a justification for using Article 107 TFEU to solve an international equity issue that has no clear efficiency consequences.
Unfortunately, in the various APA cases that have been judged, the impact on trade has been mostly presumed, as is often the case. Considering the sensitivity of the tax topic within Europe, and the amount of State aid recovery at stake (around €14 billion for Apple), it could, at least, be argued that State aid cases concerning tax rulings should require a higher level of analysis of the effect of the measure, so that there is a general and clear understanding of why State aid is an appropriate tool to use in these circumstances.
Even though State aid control might not be the right tool to deal with the ongoing tax war, something should nonetheless be done in that regard to avoid having our international tax regime fall apart.
Fortunately, after a long and painful process, Pillar 2 has been included in a directive voted in December 2022, and it will be turned into positive law in Europe in the next few years. In effect, Pillar 2 will implement a “minimal taxation” level of 15% worldwide. For instance, if a French-based MNE allocates 100 of its profits to a country with no CIT, France will levy an additional tax of 15, so that the foreign profit is taxed at a minimal level.
This measure has been designed to tame tax competition to attract profit (item 3 in the list above), it has already started to produce effects (tax havens are raising their CIT rate), and it is likely that it will significantly reduce the opportunity for MNEs to allocate new IP, people and profit in low-tax jurisdictions. In turn, this will reduce the necessity to secure these tax schemes with APAs, and lastly, we can expect that the EC crusade will finally come to a halt, for lack of new tax rulings to control.
Or maybe not. It would be somewhat naïve to assume that inter-state competition to attract MNEs will stop so suddenly. It will more likely continue but change guise; for instance, certain Swiss counties, which need to raise their tax to comply with the Pillar 2 initiatives, are publicly considering29 a new set of subsidies to offset the tax rate increase. State aid is likely to stay an interesting area of competition law in the near future.
1: For a more complete overview, see R. Avi-Yonah (2000), Globalization, Tax Competition, and the Fiscal Crisis of the Welfare State, Harv. L. Rev., Vol. 113, No. 7, pp. 1573–1676, at 1578; T. Dagan (2017), International Tax Policy, Cambridge University Press; or, for a much shorter and recent piece: J. Pellefigue (2022), A tale of two Pillars, International Tax Review, 12 May 2022.
2: P. Bachas, M. H. Fisher-Post, A. Jensen and G. Zucman (2022), Globalization and Factor Income Taxation, NBER Working Paper 29819.
3: OECD (2021), Statement on a Two-Pillar Solution to Address the Tax Challenges Arising from the Digitalisation of the Economy.
4: On that topic, see for instance C. HJI Panayi (2021), European Union Corporate Tax Law, 2e ed., Cambridge University Press, or for more economics G. Allègre and J. Pellefigue (2018), Quel rôle peut jouer l’Europe dans l’imposition des multinationales ?, Revue de l’OFCE, Vol. 158, No. 4, pp. 331–358.
5: The Commission tried to go much further in the way of harmonization of the European tax system, for instance with the Common Consolidated Corporate Tax Base (CCCTB) Directive, but because of the unanimity requirement, it was never voted. See I. Chelyadina (2019), Harmonization of Corporate Tax Base in the EU: An Idea Whose Time Has Come?, Bruges Political Research Paper 76/2019.
6: Comm. CE, Report of the Committee of Independent Experts on company taxation, Office for Official Publications of the European Communities, Luxembourg, 1992.
7: Resolution of the Council and the Representatives of the Governments of the Member States, Meeting within the Council of 1 December 1997 on a code of conduct for business taxation, OJ C 2, 6.1.1998, pp. 1–6.
8:For an overview of State aid for tax matters, see G. Renard (2004), Les règles communautaires en matière d’aides d’État et la fiscalité, L’Harmattan, Paris. See also R. Mason (2022), Tax Competition and State Aid, Virginia Law and Economics Research Paper No. 2022-10.
9:For a technical economic overview of the topic, see A. Haufler (2001), Taxation in a Global Economy, Cambridge University Press.
10: CJEC, 2 July 1974, Italy v. Commission, case 173/73, EU:C:1974:71.
11: See OECD (2011), Roundtable on Competition State Aid and Subsidies. DAF/COMP/ GF(2010)5, at 29 for a few more examples.
12: See OECD (2022), OECD Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations 2022, OECD Publishing, Paris. For a quantitative assessment of the gap between capital and profit allocation, see T. Tørsløv, L. Wier, and and G. Zucman (2022), The Missing Profits of Nations, Review of Economic Studies, rdac049.
13: A low CIT rate generally attracts so much base that it increases the CIT revenue in absolute terms. Ireland decreased its CIT rate from 50% to 12.5% in the 1990s; CIT increased from 2% to around 5% of their GDP.
14: Author calculation based on US public country-by-country reporting (CbCR) data
15: J. Pellefigue (2014), Prix de transfert : un changement radical s’impose, Bulletin fiscal Francis Lefebvre, 12, pp. 653–658.
16: If the MNE has a profit of 100 and has split it 50/50 between both states, state A can argue its share of profit should have been 80, and the MNE will then pay taxes on a total profit of 130 (80 + 50), 30 of profit will therefore be taxed twice. There are possibilities to eliminate double taxation—if state B accepts on reducing its tax base to 20, but it does not always work—certain double tax treaties do not provide for a way to come up to an agreement that binds both states.
17: According to the EU statistics (Statistics on APAs and MAPs in the EU, https://taxationcustoms. ec.europa.eu/taxation-1/statistics-apas-and-maps-eu_en), France had in 2020 only 36 APAs in force, Ireland has 6 of them, Spain and Italy around 100, and it is the same order of magnitude for other European countries, with the exception of Belgium, which has more than 1,300.
18: In all rigor, US-based MNEs should not be concerned since any profit registered in a tax haven will be ultimately taxed at the US rate, but the taxation is deferred until the profit is repatriated in the US, which can take time.
19: Comm. EC, decision 2001/354/EC of 20 March 2001, Deutsche Post AG, case COMP/35.141, OJ L 125, 5.5.2001, pp. 27–44.
20: The price p that maximize profit (p) is the same price that maximizes (1-CIT rate) profit (p). Things are obviously more complicated than that, but this is the general idea.
21: A presentation of the productive efficiency consequences of different tax systems can be found here: M. P. Devereux and S. Loretz (2010), Evaluating Neutrality Properties of Corporate Tax Reforms, Oxford University Center for Business Taxation, Working Paper 1007.
22: Of course, it is always possible for an MNE to ask for a favorable APA by suggesting a possible investment on the side.
23: Irish Times, 17 January 2015
24: “Il est temps de garantir équité et transparence dans la fiscalité des entreprises en Europe” (free translation). P. Moscovici (2015), Fiscalité: l’heure est à la transparence l’Opinion, 18 March.
25: Mason (2022), supra note 8.
26: CJEU, 15 November 2011, joined cases C-106/09 P and C-107/09 P, EU:C:2011:732.
27: OECD (2011), supra note 11, at 22.
28: Ibid., at 28–38.
29: S. Jones (2021), Switzerland plans subsidies to offset G7 corporate tax plan, Financial Times, 10 June.