The taxation of the digital economy 1 provides an excellent area for observing the attempt to govern globalization by way of legal tools 2 as well as the difficulties in making such a perspective concrete.
On one hand, the incomprehension caused by the low level of taxation of large companies in the digital economy in jurisdictions where they generate important revenues, has contributed powerfully to the emergence of a previously unknown cooperation. During the last decade, almost 140 countries have been working, under the aegis of the OECD, on ways to adapt their corporate taxation rules to a number of new phenomena, closely related to the digitalisation of the economy, that challenge the traditional premises on which international tax law has relied for a century. This movement has already led to significant changes in rules and practices worldwide, particularly in the fight against tax evasion and tax avoidance.
On the other hand, when the attention is focused on the more specific question of the taxation of the profits of large companies in the digital economy, the difficulty of getting governments to converge towards a common legal framework remains evident, even though the business models privileged by these companies raise identical problems all over the world. Thus, despite an almost global consensus on the limits of the current tax system, initiatives that are purely national – and highly political – are multiplying in the greatest disorder, giving rise to outcries and threats of reprisals from States that consider themselves victims of discriminatory practices. The prospect of a uniform legal response to a clearly global issue thus remains, for the time being, very hypothetical.
Before analysing the fiscal challenges raised by the digitalisation of the economy and the way in which States are trying to respond to them, it is useful to briefly recall the origins of the current international tax system to understand its current inadequacy.
A global tax system designed in the 1920s
The way in which States share the right to tax the income of multinational companies is still based on a model conceived by the League of Nations in 1928.
Paradoxically, if taxation has always constituted an obvious attribute of sovereignty, it appears to be one of the oldest areas for the development of a form of legal globalisation. Since the 1920s, the League of Nations has embarked on an ambitious programme of reflection on how States could coordinate their fiscal jurisdictions in order to avoid certain “frictions” detrimental to international commerce. Mostly, the League of Nations wanted to avoid cases of “double taxation”, linked to situations in which two States may, on the basis of their respective domestic rules, claim the same right to tax the same profit or the same estate. For example, in the case of amounts earned abroad by an individual or a company, the “source” State can legitimately wish to tax this income generated on its territory, while the “residence” State usually counts on taxing all the income of taxpayers domiciled on its territory, including foreign revenues.
To reduce these risks, the League of Nations published in 1928 a first model convention whose provisions still guide, in substance, bilateral tax treaties that most countries in the world sign with their economic partners. These conventions aim at allocating the various sources of revenue, or even different estates, between the signatory States for the purpose of avoiding situations of double taxation. Thus, for example, the signatories generally agree that the source State should have the right to tax so-called “passive” income (income from the estate in particular) generated on its territory – including if perceived by residents of the other signatory State – while, conversely, the State of residence is alone to tax earnings, including where the economic activities are carried out in the other State (unless the company has a “permanent establishment” there, i.e. a fixed and permanent establishment with its own activity).
New ways of creating value
Recent evolutions in the models of value creation favoured by many companies tend to challenge this traditional model of repartition of taxable income between source States and seat States. To put it briefly, these models are making it more difficult every day to identify the place or places of value creation of an activity and, consequently, the distribution of the rights to tax this value between the relevant States.
In this respect, large multinational companies certainly offer the most striking examples of such struggles – which explains why they have become for many the symbol of practices, even abuses, which are not really their own. We are thinking here of companies such as Facebook or Google, which generate significant revenues from services provided to users and customers residing on the territory of States where the companies might have no physical presence. However, the 4,500 bilateral conventions which, today, are all inspired to a greater or lesser extent by an OECD model (which took over from the League of Nations in this matter), grant in principle to the country where the company is domiciled – or at least, has a fixed place of business (branch, office, factory…) – the right to tax the relevant profits. Thus, for example, only Ireland is entitled to tax the profits generated by Google in France when its lucrative advertising space placement activities are carried out by employees residing at group’s European subsidiary in Dublin. This disconnection between, on the one hand, the territory in which a large part of value is generated– or at least yielded – and, on the other hand, its place of taxation appears today, to many governments to be incomprehensible for the public opinion.
But this is not the only difficulty faced by States. Indeed, most of the value produced by the activity of many companies, including in “traditional” sectors (hotels and restaurants, for example) now relies on so-called intangible assets – an algorithm for Google and, more often, a trademark, patents, know-how, etc. – which may themselves be located almost anywhere, lodged with dedicated subsidiaries. The Starbucks case, which arose in the UK in 2012, brought this problem to light. Despite a turnover exceeding £2 billion generated in the UK, the company managed to pay no income tax there, by relocating all its profits via a subsidiary located in the Netherlands and in charge of controlling the group’s intangible assets for the whole Europe – assets whose for which the right of use was charged at a high price to local subsidiaries, particularly the ones in the UK 3 . At the same time, all of the large American companies in the digital economy were benefiting from sophisticated schemes that made it possible – in all legality – to minimise their overall tax rate by locating the bulk of their profits with subsidiaries established in accommodating jurisdictions (Cayman, Bermuda…), after having wiring them through various tunnel jurisdictions whose legislation and bilateral conventions authorised this type of practice 4 .
Finally, there is an even more serious difficulty: that of identifying the value generated by certain services – a value that must be understood in order to be located and then taxed. For example, while it is clear that Google or Facebook’s advertising revenues are linked to the information collected from users (through their search history, in particular), is it necessary to and, if appropriate, how should this user participation – this “free labour”, as the Collin-Colin Report 5 described it – be valued in the creation of the overall value of the service sold by Google or Facebook to its customers? And, beyond that, how can the “fair value” of certain intangible assets – Google’s algorithm or Starbucks’ Frappuccino recipe, for example – be assessed, in order to determine the acceptable level of remuneration? This question of the amount of “transfer prices”, i.e., the prices at which companies in the same group charge each other (and therefore “off-market”) for goods or services, is obviously crucial for governments: the rate of taxation depends directly on the amount of profits, which in turn depends on the amount of charges deducted by the company as a result of the transfer prices charged by its subsidiaries and parent company.
The BEPS project facing the “challenges of the digital economy”
States reacted rather late to these new facts. They initially preferred to focus their attention on the most obviously questionable behaviours when, in the aftermath of the 2007-2008 crisis, the need to regain budgetary leeway became apparent and, at the same time, the press revealed massive tax evasion practices sheltered by a few “tax heavens” (Liechtenstein, in particular 6 ). Annoyed at having had to provide huge financial aid to save banks which, at the same time (at least for some of them) facilitated tax evasion that reduced public revenues, governments quickly implemented various vigorous measures, both at the national level (as with the FATCA law in the United States in 2010 7 ) and internationally (through initiatives targeting bank secrecy and aiming at the development of information exchange between tax administrations led by the OECD at the invitation of the G20 8 ).
It was only in a second phase that attention focused on the different processes by which many large transnational corporations reduce their tax rates in a way that is legal but problematic from the point of view of public finances and, in many respects, of fairness. Thus, in 2012, the OECD convinced the G20 9 to support it in a project called “BEPS” (Base erosion and profit shifting) which, as its name suggests, aimed to uncover some potentially questionable tax optimisation practices and, ideally, to propose ways to overcome them.
From the very beginning, companies of the digital economy – at the forefront of which are the “GAFAs” (Google, Amazon, Facebook, Apple) – were clearly in the scope of OECD’s efforts. The first of the fifteen “BEPS actions” thus aimed to “address the challenges raised by the digital economy” 10 . Not that these companies were then accused of adopting especially “aggressive” tax behaviour compared to other multinationals. However, in light of several features of their economic models and their location policies – those mentioned above – they raise, in the view of governments, risks of “erosion of the tax base” of a particular severity 11 .
However, during the publication in 2015 of the different BEPS Reports and in the following years, the issue of the digital economy was relegated to the background. The OECD preferred to concentrate on more fertile grounds, so as to develop consensual solutions: the creation of “anti-abuse” standards (in order to annihilate the advantage of resorting to certain purely artificial tax arrangements), the development of cooperation mechanisms between tax administrations, the multiplication of “reporting” obligations by companies, the drafting of an innovative “multilateral instrument” allowing interested States to simultaneously modify their bilateral tax treaties in order to integrate the new anti-fraud provisions proposed by the OECD, etc 12 .
Similarly, at the European level, several guidelines inspired by the OECD required national legislation to incorporate an arsenal of anti-fraud and anti-evasion provisions 13 . Some States that have long been accused of facilitating tax optimisation or even tax evasion (Luxembourg and the Netherlands, in particular) have thus amended their legislation and are trying to adopt best practices in terms of transparency and the fight against abusive corporate behaviour 14 . These trends reflect the emergence of legal standards that are likely to be imposed on a global scale as the subject of tax fraud has become so sensitive: the label of “tax haven” is becoming very difficult to bear, including by certain small States long associated with this expression 15 .
Towards new principles for taxing the profits of multinational companies?
Concerning the challenges raised specifically by the digitalisation of the economy, the observations put forth in the Collin-Colin Report and largely taken up by the BEPS Action 1 Report have never been denied. However, States have so far been unable to agree on their implications.
At the global level, the OECD has made significant efforts to try to overcome the relative failure of Action 1, which, due to a lack of consensus – notably due to the Obama Administration’s reluctance to proposals that could affect some large US companies – had not resulted in anything tangible. Contrary to all expectations, the election of Donald Trump changed the situation. Most of the concerns raised by the practices of large digital companies overlapped with those taken into account, more generally, by the tax reform initiated by the Republicans in 2017 and aimed in particular at encouraging the relocation of American company profits to the United States while introducing a mechanism for minimum taxation of their profits made abroad 16 .
New “BEPS 2.0” negotiations, led by the United States, began in January 2019 in an “inclusive framework” bringing together 140 countries. A work plan was then adopted by the G20 summit in Fukuoka on 8 and 9 June 2019. These discussions led to two sets of proposals endorsed by the members of the inclusive framework in January 2020 17 .
Firstly, a “Pillar 1” aims to take better into consideration the point of sale of various goods and services (even beyond the case of digital companies) in locating the profits generated by these sales. Concretely, the idea is to propose a profit allocation formula that takes into account not only the country where the company is headquartered but also the consumer States, in order to allocate to the latter a fraction of the profits made on their territory by companies not located there. Secondly, a “Pillar 2” intends to introduce a minimum effective tax rate for multinationals, allowing States to tax their international groups for profits made abroad and little or no tax thanks to bilateral conventions. The aim is to discourage the relocation of profits for purely tax purposes and thus limit tax competition between States.
However, due in particular to the United States’ volte-face, which at the end of 2020 distanced itself from the negotiations 18 , the OECD proposals remain for the time being in the boxes. Indeed, there is no indication that other States that have actively participated in the discussions could not be disillusioned when examining in more detail the practical consequences that the practical implementation of the two pillars might imply. In particular, there is nothing to affirm that France would be a winner if “Pillar 1” were applied: for a few billion taxes potentially collected on the profits of Google and Facebook, how many would have to be waived on the profits made abroad by LVMH or Sanofi?
Pending a comprehensive response, the European Commission’s proposals
For its part, in March 2018 19 , the European Commission presented several proposals which, pending a hypothetical global solution from the OECD, could limit the risks of erosion of the tax base and, at the very least, respond to the “sense of injustice” 20 arising, according to the Commission, from the current arrangements for taxing digital companies. The Commission thus assumes the political as well as the legal character of its approach: it is indeed a question of proving Europe’s capacity to take action on a global subject without waiting for the approval of the United States. These proposals also enable it to suggest, very subtly, that there is indeed a homogeneous European public opinion (sharing in this case a feeling of injustice) whose aspirations could be effectively supported by the European institutions.
According to the Commission, a first proposal would be to establish new rules of taxation for companies with a “significant digital presence” on the territory of a State: the latter would gain the right to tax part of the profits linked to this “virtual” presence, even in the absence of a physical establishment. This is a proposal that is as bold as it is difficult to implement: it involves revising each of the bilateral tax treaties signed by each Member State, since all of them currently determine the place of taxation of company profits on the basis of the universally accepted notion of “permanent establishment”, which itself is based on criteria of an essentially physical nature. A second proposal would be a common system of tax on the turnover generated by companies from providing certain digital services.
The latter proposal received a more than mixed reception. Only ten out of twenty-seven States considered it relevant 21 and several expressed an open hostility. This was the case, first of all, of Ireland. Even though such a tax would cause only a very limited loss of revenue for Ireland, its government argued that it was essential to keep these discussions at the OECD level in order to find solutions on a global scale 22 . While sharing this sentiment, Sweden is opposed to the very principle of a turnover tax, believing that “it would hamper innovation, investment and growth in the Union and harm its competitiveness against other regions” 23 . Denmark and Luxembourg broadly share these points of view 24 .
National tax initiatives on digital services
In light of the halting of Commission’s proposal and of a yet-to-come global solution, several European States introduced –or at least announced their intention to do so –a taxation scheme specifically targeting certain digital services. For their promoters, these measures have the advantage of maintaining pressure on international organisations and, even more so, on their more cautious members: the prospect of a multiplication of these taxes may indeed lead them to the conclusion that a global response would ultimately be the “lesser harm”.
Thus, in France, a law of 2019 25 introduced a 3% tax on revenues from marketplace services (Amazon, Blablacar, Airbnb, etc.) and on revenues from the placement of advertising messages targeted according to user data. Only companies with a worldwide turnover of more than 750 million euros and 25 million euros for services provided in France are in the scope.
Several similar measures have been enacted or are in the process of being enacted, by several other European countries: Austria, the Czech Republic, Hungary, Italy, Spain and the United Kingdom 26 . Likewise, beyond European borders, several large countries – Brazil, India, Indonesia and Turkey – have already implemented (or are considering implementing) a “digital services tax” comparable to the French “GAFA tax”, as recently underlined by the US Department of Commerce 27 .
Legally relevant answers?
The French tax (like the equivalent foreign schemes) is subject to several criticisms. First, it misses its objective in two ways: first by taxing turnover without taking into account profits (and in particular profits supposedly relocated by some of the companies subject to the new tax), and second because this consumption tax can easily be passed on to customers. Thus, at the end, rather than reducing the profits of foreign companies (Google and others), the economic burden of the tax falls essentially or even exclusively on their French customers.
Second, these unilaterally implemented taxes raise particularly heavy risks of retaliation compared to their rather modest budgetary yield (of the order of 350 million per year) 28 . The American authorities thus announced in 2019 that the French tax was, in their view, a discriminatory practice on the part of a trading partner that justified countermeasures, in application of section 301 of the 1974 Trade Act 29 . A list of 63 French products that in 2018 amounted to approximately 2.4 billion euros of imports was thus drawn up, with the intention of subsequently applying to them customs duties of up to 100% 30 . While the US Trade Representative decided in January 2021 to suspend the application of these duties, it has only done so in view of the ongoing investigation into other “digital services tax” initiatives launched by several countries (Austria, Brazil, the Czech Republic, the European Union, India, Indonesia, Italy, Spain, Turkey and the United Kingdom) 31 . The case is therefore far from being closed.
However, supporters of these domestic taxes can legitimately argue that without their threat, the negotiations conducted by the OECD since 2019 would probably never have gotten off the ground. As for the structural deficiencies of these taxes, they can be explained first and foremost by the fact that any other form of tax, and in particular a tax based on the profits of foreign companies, would inevitably run up against the provisions of the bilateral conventions signed by the relevant States.
In other words, pending amendments to the latter – if necessary, to incorporate the OECD proposals – those States most token on acting to reflect their domestic public opinion have little choice of means of action.
Moreover, the pressure exerted by some governments to move forward with international negotiations has undoubtedly had a significant impact on the very practices of many companies, which are worried about their public image. Google’s decision to sign a transaction with the French tax authorities in 2019 and a judicial public interest agreement (Convention Judiciaire d’Intérêt Public) 32 undoubtedly illustrates this concern: the US company preferred to give up one billion euros to the French State to put a definitive end to its disputes with the tax authorities rather than wait for their jurisdictional solution, even though the first and second instance courts had ruled in its favour 33 .
On the other hand, on the legal side, evolutions are still hypothetical. While not completely satisfied with the current system, many States are mainly concerned about the still unclear OECD proposals and the risks of legal certainty they would create by demanding a complete overhaul of the principles that have underpinned international taxation for the past century. In the end, the only certitude that emerges at the beginning of 2021 is that in digital taxation, as in other areas, it is indeed the international balance of powers and, in particular, the willingness of the United States to push in one direction rather than the other that will lead – or not – to the evolution of the law, in order for the latter, perhaps one day, to govern globalization.
- As the Collin-Colin Report pointed out, in 2013, the digital economy refers to companies immediately associated with this notion (companies in the advertising, information or entertainment sectors whose activity is essentially based on the use of digital technologies, and software publishing companies), but also to companies of all sizes and of all sectors that are increasingly using digital technologies and, in particular, are making ever more intensive use of the data resulting from their users’ activities. Economists nowadays prefer to speak of the «digitalisation of the economy» rather than of the digital economy, since this phenomenon encompasses most of the activities involved in the production of goods and services, including the most traditional ones (P. Collin, N. Colin, Mission d’expertise sur la fiscalité de l’économie numérique, Rapport au Ministre de l’économie et des finances, au Ministre du redressement productif, au Ministre délégué chargé du budget et à la Ministre délégué chargé des petites et moyennes entreprises, de l’innovation et de l’économie numérique, 2013, pp. 5 et seq.).
- See in particular: A. Garapon, “Une Cour peut réguler la mondialisation”, La Revue européenne du droit, September 2020, no. 1, pp. 89 et seq.
- For a presentation of this scheme, see in particular: T. Bergin, “How Starbucks avoids UK taxes”, Reuters, October 15 2012.
- See in particular, P. Collin, N. Colin, ibid., 2013, p. 21.
- Ibid, pp. 52 et seq.
- See in particular: E. Vincent “Lichtenstein, la vallée des milliards cachés”, Le Monde, February 19 2008.
- The provisions of the Foreign Account Tax Compliance Act (FATCA) require banks around the world to report all accounts held by US citizens to the US tax authorities. For more information, see: https://www.irs.gov/businesses/corporations/foreign-account-tax-compliance-act-fatca.
- In 2020, the OECD welcomes both the dramatic increase in the number of exchanges of information since the launch of its “Global Forum on Transparency and Exchange of Information for Tax Purposes” in 2009, and the significant reduction (by almost 25% between 2008 and 2019) in the volume of bank deposits in international financial centres (Singapore, Hong Kong, etc.) by non-residents (see: https://www.oecd.org/fr/fiscalite/la-communaute-internationale-a-obtenu-un-succes-sans-precedent-dans-la-lutte-contre-la-fraude-fiscale-internationale.htm).
- In a micro-formula lost in the middle of the 14 pages of the final declaration of the G20 meeting in Los Cabos (Mexico) in June 2012: “We reiterate the need to prevent base erosion and profit shifting and we will follow with attention the ongoing work of the OECD in this area”, see: https://www.oecd.org/g20/summits/los-cabos/2012-0619-loscabos.pdf).
- OECD, Meeting the Fiscal Challenges of the Digital Economy, Action 1 – Final Report 2015.
- However, the statistical data appear to be as incomplete as they are contradictory. The European Commission estimated in 2018 that, on average, digital businesses were seemingly taxed at an effective average rate of 9.5% compared with 23.2% for traditional business models (see the European Commission’s services’ impact study on the two proposals for directives: SWD (2018) 81 final/2). For its part, the US Department of Commerce notes in its investigation report on the “GAFA tax” adopted by France that several studies consider, on the contrary, that the tax rates of these different categories of companies remain globally equivalent (Office of the United States Trade Representative, Report on France’s Digital Services Tax prepared in the investigation under Section 301 of the Trade Act of 1974, December 2 2019, p. 5.).
- For an overview, see: https://www.oecd.org/fr/fiscalite/beps/actions-beps.htm.
- See the ATAD (Anti-Tax Avoidance Directive) of July 12 2016 (Dir. 2016/1164) and “DAC 6” of May 25 2018 (Dir. 2018/822).
- See in particular: X. Paluszkiewicz, F. Dumas, “L’espace fiscal européen”, Information report submitted by the Committee on European Affairs, Registered with the Presidency of the National Assembly on 9 July 2020, p. 34.
- More than the prospect of being ostracised by the international community, the fear of scaring off foreign investment – with companies themselves becoming very worried about their image – probably explains this phenomenon.
- For a presentation of the 2017 US tax reform, see e.g., S. Humbert, “Les frontières des impôts de production”, CPO, Special Report No. 2, 2020, pp. 55 et seq.
- See in particular, Statement by the OECD/G20 inclusive framework on BEPS on the two-pillar approach to address the tax challenges arising from the digitalisation of the economy, as approved by the OECD/G20 inclusive framework on BEPS on 29/30 January 2020.
- The US then expressed its «strong reluctance» to the project (CPO, Adapting Corporate Taxation to a Digitised Global Economy, 2020, p. 129).
- See in particular: documents COM (2018) 147 final and COM (2018) 148 final.
- Proposal for a Council Directive on the common system of taxation on digital services applicable to products derived from the supply of certain digital services, COM (2018) 148 final, p. 3.
- See in particular: X. Paluszkiewicz, F. Dumas, Rapp. Ibid., p. 50.
- Idem. p. 51.
- French law n° 2019-759 of July 24 2019 creating a tax on digital services and modifying the trajectory of the reduction in corporate income tax.
- The characteristics of the different taxes are presented by: X. Paluszkiewicz, F. Dumas, Ibid. p. 53.
- See in particular: Office of the US Trade Representative, Docket Number USTR-2019-009, 7 January 2021.
- According to the finance bill for 2021, this tax should bring in 358 million euros in 2021, after bringing in 405 million euros in 2020 (PLF 2021, t. 1, p. 33).
- See the presentation of the latest actions taken on this basis – against China (intellectual property rights), the European Union (subsidies to Airbus) and France (tax on digital services) – on the website of the US Secretary of Commerce: https://ustr.gov/about-us/policy-offices/press-office/fact-sheets/2018/june/section-301-investigation-fact-sheet.
- Office of the US Trade Representative, Notice of Determination and Request for Comments Concerning Action Pursuant to Section 301: France’s Digital Services Tax, Federal register/vol. 84, n° 235, December 6 2019, p. 66957.
- Office of the US Trade Representative, Docket Number USTR-2019-009, Jan. 7 2021.
- The agreement is available on: https://www.agence-francaise-anticorruption.gouv.fr/files/files/190903_CJIP.pdf.
- See in particular at last: CAA Paris, April 25 2019, min. c/ Google Ireland Ltd, Dr. fisc. 2019, n° 25, comm. 305, concl. A. Mielnik-Meddah, note F. Deboissy and G. Wicker.
Martin Collet, Taxation of the digital economy: global challenge, local responses?, Aug 2021, 140-144.
à lire dans cette issuevoir toute la revue
A plural global governance
At a first glance, it might seem anachronistic to write about global governance, since the era of grand universalist declarations, globalizations of trade and transnational agreements is on the brink of being replaced by that of rediscovered national interests, isolationisms and the selfishness of the “Me First” politics. The crisis – or perhaps the polycrisis … Continuedlire l'article
In the spiral of humanisms
The instability of our societies multiplies the crises (socio-economic, migratory, climatic, sanitary…) which are intertwined in a single poly-crisis, piling up states of emergency, from the terrorist attacks of 2001 to the pandemic of 2020, while a kind of normative madness takes hold of our societies. We must abandon the usual metaphors of legal systems … Continuedlire l'article
Geopolitics of the Energy Transformation
1. The internationalisation of energy transactions The history of energy can be written from myriad perspectives, depending on the object emphasised in each account. A household, a river, an activity, an event, a specific resource, a given technology, a country, a region, a global process or combinations thereof are some of the objects around which … Continuedlire l'article